From Teotihuacan to the Aztecs – Mesoamerican Urban Landscape

Cities and urbanism in central Mexico during the Classic and Late Postclassic periods are discussed here with little reference to Teotihuacan or Tenochtitlan. This may strike most people as an unusual approach, since these two metropoli are often seen as virtually synonymous with the concepts of “cities” or “urban” in central Mexico. These were the largest cities in the Prehispanic New World, their influence spread over thousands of square kilometers of Mesoamerica, and we know quite a bit about them as a result of archaeological and ethnohistoric research. No discussion of Mesoamerican urbanism can avoid discussion of Teotihuacan or Teotihuacan, and in fact these cities typically dominate discussions of the nature of cities in Mesoamerica. Nevertheless, it may be observed that a reconsideration of central Mexican urbanism from a “non-metropolitan” perspective illuminates many aspects of sociopolitical organization that are difficult to comprehend when we give too much weight to the big cities.
I approach the analysis of central Mexican urbanism through the concept of the “urban landscape.” This describes the spatial distribution of social complexity across the landscape of an urban society. It derives from two anthropological approaches: Anthony Leeds’s (1980) functional definition of rural and urban activities, and the regional analysis approach of Carol Smith (1976) and others. In Leeds’s words:
“any society which has in it what we commonly call “towns” or “cities” is in all aspects an “urban” society, including its agricultural and extractive domains … the terms “urban” and “rural” come to stand to each other not as opposites and equivalents. Rather the inclusive term describing the whole society is “urban” while the term “rural” refers only to a set of specialties of an urban society characterized by being inherently linked (under any technology known) to specific geographical spaces.” (Leeds 1980:6-7)
For Leeds, cities and towns are nodal settlements, places where “space-intensive” activities such as commerce, industry, administration, and education take place, while rural areas are the settings for “space-extensive” activities such as agriculture, herding, hunting, mining, and lumbering. In the regional analysis approach, cities are approached as parts of functionally-defined regions. Most regions in urban societies have settlement hierarchies, and explanations of social processes on the regional scale must consider the full urban hierarchy plus rural areas.
The urban landscape concept has several advantages over previous approaches to Mesoamerican urbanism (e.g., Hardoy 1968, 1973; Marcus 1983; Sanders and Santley 1983; Sanders and Webster 1988). First, attention is drawn to smaller cities and towns in addition to the large cities that have traditionally dominated archaeological discourse. In many ways, these smaller cities had a greater impact on people’s lives than the distant and atypical imperial capitals. I have suggested elsewhere that “the majority of the population of central Mexico carried out most of their urban activities at settlements other than the Aztec imperial capital” (Smith 1989:456). The regional approach to urbanism of Kowalewski (1990), and Blanton et al. (1993) also acknowledges the importance of smaller cities and towns.
A second advantage of the urban landscape concept is that it frees us from the common assumption that social complexity was found only in cities. There is no doubt that cities are typically the loci of social complexity within their societies. Modern studies of urbanism in geography (Garner 1967), archaeology (Trigger 1972), and economic anthropology (Blanton 1976; Smith 1976) all rely upon basic socioeconomic principles that operate to concentrate the institutions and activities of social complexity in urban places. As outlined by Trigger:
— There is a tendency for human activities to be hierarchical in character and for this to be reflected in spatial organizations.
— With increasing complexity there is a tendency for activities and social institutions to be more clearly defined and for their personnel to be more highly specialized.
— Human activities tend to be focal in character in order to take advantage of scale economies.
— The size of communities tends to vary with the number of functions they perform. (Trigger 1972)
When this natural association between cities and complexity is coupled with a sharp rural:urban dichotomy, the result is an assumption that rural areas were socially homogeneous, with simple peasants toiling away to support the cities and elites (this assumption has a long intellectual history in Mesoamerican studies; see Smith 1994a:145). Archaeologists working in rural areas of the early states of both Mesopotamia and Mesoamerica, however, have found considerable evidence for social complexity in the countryside (Schwartz and Falconer 1994), and our conceptual models need to account for these findings.
Turning to ancient central Mexico, one of the major differences between the urban landscapes of the Classic and the Late Postclassic periods is the presence of rural complexity in the latter but not the former period. This relates to major differences in the nature of regional political and economic organization in the two time periods. The central Mexican highlands in both periods was a functional region in that: (1) numerous exchanges of goods, people, and information took place throughout the area, and (2) economic activities were structured by a large central place whose influence touched all parts of central Mexico. The centripetal forces that integrated this region in the two periods were quite distinct, however. In the Classic period, Teotihuacan dominated the region politically and economically, and centralization and hierarchy were the dominant centripetal forces. In the Aztec period, however, the role of Tenochtitlan was far less prominent. Although centralization and hierarchy were significant at the city-state level, on a regional scale non-hierarchical forces such as market exchange, city-state competition, and cultural unity were paramount.
These contrasts between regional political economy in the Classic and Aztec periods are not at all clear from the nature of the imperial capitals themselves. They only show up when we look at rural areas of central Mexico and at the smaller non-imperial cities that filled the urban landscape in each period (by “non-imperial cities” I mean central Mexican cities other than Teotihuacan or Tenochtitlan).
The massive urban center of Teotihuacan is well known and has been described in many archaeological reports. Here I will only list some of the notable features of urbanization and urban planning at Teotihuacan. Teotihuacan was one of the two largest urban centers of the Prehispanic New World (Tenochtitlan was the other), with a large population and high population density. The city exhibits a high degree of centralized planning, with a common grid orientation, ceremonial avenues, and standardized temple complexes. The dominant residential form was the apartment compound. I refer the reader to the many reports on Teotihuacan for more information and descriptions (e.g., Millon et al. 1973; Millon 1981). The Feathered Serpent Web Page is another useful reference source.

Rural Areas
Most of our information about settlement outside of Teotihuacan comes from regional survey. The initial growth of the capital in Tzacualli times was accompanied by rural depopulation as people were moved into the city. Then in the Classic period, the rural Basin of Mexico was repopulated with small and relatively uniform villages; Sanders et al. (1979:116) assign Teotihuacan a large part in directing and organizing this expansion and reorganization of rural settlement (see below). Although this rural population growth was significant, regional densities were not very high and there is little evidence for agricultural intensification in rural areas. Irrigation canals were built at Teotihuacan itself, however.
The only “rural” site excavated with more than a test pit or two is Maquixco Bajo, a village only a few km from Teotihuacan itself (Sanders 1967; Sanders et al. 1979:334-355). This was a simple village under the strong control of Teotihuacan. Social complexity at small sites can be discussed under the dimensions of heterogeneity, inequality, and connectivity (see Smith 1994a; the first two dimensions are from McGuire 1983). In contrast to the Aztec period, when rural villages exhibited high levels of complexity along all three dimensions, Maquixco Bajo has few elements of complexity. Small-scale obsidian blade production is one of the few types of heterogeneity present. Houses were smaller versions of the well-known Teotihuacan apartment compounds and exhibited little variability. One structure, mound 4, stood out from the others in size and quality, suggesting the presence of some inequality. In terms of connectivity, there were few imported goods at the site apart from obsidian; Sanders et al. note the lack of jewelry or other luxury goods, and no imported ceramics are mentioned. The only external connections evident from the artifacts and architecture are with Teotihuacan.
Other rural villages in the Classic period Basin of Mexico were also under the heavy hand of Teotihuacan. Sanders et al. (1979:344) find “a much greater penetration of urban institutions into rural life” compared with the Aztec period. They suggest that rural settlement “was deliberately structured by Teotihuacan” (p.116) in order to maximize agricultural production and to consolidate control by drawing settlement away from older political centers. Analyses of systematic surface collections from rural Classic sites in the Yautepec Valley of Morelos by Lisa Cascio will evaluate the applicability of these patterns to an area of the Teotihuacan empire outside of the Basin of Mexico.
Non-Imperial Cities
Only ten Classic sites in the Basin of Mexico, apart from Teotihuacan, can be considered cities or towns. These are “provincial centers” in the classification of Sanders et al. (1979), with populations between 1,000 and 10,000; none of these has been studied beyond the initial regional surveys. There is better information on Classic cities in nearby areas, including Chingu, Tepeapulco Calpulalpan, San Ignacio, and several sites in the Yautepec Valley to the south (Dí¡º 1980; Linné ±942; Matos et al. 1981; Cascio et al. 1995; Hirth 1980).
It has long been recognized that these sites bear considerable resemblance to Teotihuacan (e.g., Sanders et al. 1979), and Charlton’s (1991) systematic comparison clarifies the situation. Although none of the sites is a copy of Teotihuacan, Charlton shows that nearly all exhibit several of the following five traits:
— Grid plans.
— Ceremonial avenues.
— Apartment compounds.
— Pyramid-plaza complexes.
— Compact and nucleated settlement.
The first four of these are distinctive traits associated with urban planning and state control at Teotihuacan, and their appearance at hinterland cities suggests that these smaller centers were controlled, or at least very strongly influenced, by the imperial capital. The final trait, compact nucleated settlement, is difficult to asses in the absence of quantified data from the smaller centers. The presence of apartment compounds does suggest a high population density; the figure for Maquixco Bajo, the village with apartment compounds, is 75 persons/hectare (although this is only slightly higher than the density of Aztec non-imperial cities, which I argue below were dispersed settlements with considerable open space).
In the absence of published work on the non-imperial cities, it is difficult to asses the nature of urban functions at these sites. The city-planning traits listed above suggest that territorial administration was probably an important activity, and the presence of large temple-pyramids points to the importance of religion, perhaps stressing the state-sponsored cults of Teotihuacan. We have virtually no economic data on any of these sites, although Cascio’s current work in the Yautepec Valley will address this issue for several Classic cities.
As was the case with Teotihuacan, the Aztec capital Tenochtitlan is well described in the standard literature. I suggest the following as good introductions to the ethnohistory and archaeology of the city: Calnek (1976), Rojas (1986), Matos (1988), Smith (1996), and Umberger (1996).
Rural Areas
A striking feature of the Aztec countryside was the large rural population. The Basin of Mexico (and probably other areas) witnessed a population explosion between the Early Aztec (AD 1150-1350) and Late Aztec (1350-1550) periods, resulting in very dense rural settlement. In many areas, rural settlement was continuous across the landscape with few breaks (Sanders et al. 1979:163-171). Agriculture was heavily intensified in response to this demographic growth, with major transformations to the hills (stone terrace construction), valleys (irrigation systems), swamps (the construction of chinampas), and settlements (house gardens). Cities and towns multiplied, most of which supported busy marketplaces that were tied together into an integrated regional marketing system (Blanton 1996).
Our knowledge of rural society in the Basin of Mexico is limited by a lack of excavation at rural sites. The only example is Susan Evans’s (1988) excavations of houses at the large village or town site of Siguatecpan, where she uncovered several commoner houses, a possible elite residence, an obsidian workshop, and a household textile industry. Brumfiel (1987) suggests that most utilitarian crafts (such as ceramics and obsidian) were produced by part-time peasant artisans, whereas luxury goods were made by full-time attached artisans in Tenochtitlan. In this model, there was little craft production in non-imperial cities. Although scattered ethnohistoric and archaeological evidence fits this model, there is not enough evidence to confirm or reject it. In fact, the strongest evidence for Aztec craft industries shows the intensive production of both utilitarian and luxury goods at Otumba, a city-state capital (Charlton et al. 1991).
My own excavations of houses at the rural sites of Capilco and Cuexcomate in western Morelos provide more extensive information on Aztec rural communities. One of the interesting findings of that project was the existence of a high level of social complexity at these sites (Smith 1992, 1994a; Smith and Heath-Smith 1994). This discussion is a summary of information presented in Smith (1994a), where I discuss the concept of rural social complexity in terms of heterogeneity, inequality, and connectivity, Indicators of social heterogeneity include the presence of a clear settlement hierarchy, spatially differentiated ritual practices, small-scale craft production, and intensive farming methods. The craft industries included intensive domestic cotton textile production, low-level manufacture of tools from local chert and basalt, and the production of bark paper. An abundance of pigments and bark beaters in one patio group near the elite residential compound at Cuexcomate may indicate the production of painted manuscripts, a luxury item in Aztec society.
These sites also exhibited several types of inequality in wealth and status. First, an elite residential compound stood out clearly from the bulk of the residences in both architectural and artifactual measures of inequality. Second, there was considerable variation in wealth among commoner households at both sites (as measured by artifactual indices). Another dimension of complexity was what I called “connectivity,” the extent to which households were tied into wider networks of commerce and stylistic interaction. There were many stylistic similarities with Aztec peoples throughout central Mexico, particularly in ritual items and the elite residences, and all households had ready access to large quantities of imported goods, including obsidian, foreign pottery from many areas, and rarer exotics such as bronze tools and jade beads.
As in the Basin of Mexico, rural population densities were high and intensive agricultural practices widespread in Aztec-period Morelos (Smith 1994b). The rural communities of Capilco and Cuexcomate were socially complex settlements that were well tied into larger networks of exchange and interaction. There is little evidence to suggest that the peasants who inhabited these sites were heavily dominated by external elites in either their city-state or the Aztec empire. It is clear that in the Aztec urban landscape, the institutions and practices of social complexity were not limited to cities, and the role of peasants went far beyond that of food producers for urban elites and specialists. Just as there was something of the “urban” in these rural sites, we will see that there was much of the “rural” in Aztec cities.
Non-Imperial Cities
Most Aztec cities have been destroyed or buried under modern settlements, and those few examples that are wholly or partially available to archaeologists e.g., Chalco, Coatlan, Huexotla, Otumba, Xaltocan, and Yautepec have barely been touched by excavation (Hodge and Smith 1994; Smith 1996). Nevertheless, we do have sufficient archaeological data, augmented by ethnohistory, to identify some patterns in the size, layout and organization of these urban settlements.
On the basis of political status and population size, three levels of urban hierarchy can be identified below Tenochtitlan in both the Basin of Mexico and Morelos (the data are assembled in Smith et al. 1994:Table 3). Powerful capitals in both regions, such as Texcoco, Cuauhnahuac or Yautepec, were large centers with populations from 10,000 to 40,000 inhabitants. City-state capitals were larger in the Basin of Mexico (ca. 9,000 population) than in Morelos (ca. 3,000 population), and non-capital towns were the smallest urban settlements in each region.
Non-imperial cities, in Morelos at least, exhibited common principles of city planning and spatial organization. Each city was built around a central civic core consisting of an open public plaza. A large temple-pyramid occupied the east side the plaza, and other structures, such as palaces, ballcourts, and other civic buildings, occupied the other sides of the plaza. These buildings had a common orientation, within a few degrees of the cardinal directions. Outside of the civic core, residences were scattered around with divergent orientations and little or no evidence of central planning (Smith n.d., 1996). This general schemea well-planned civic core with large religious and administrative buildings surrounded by scattered unplanned residential districtswas an ancient and widespread principle of city planning in Mesoamerica. Most cultures followed this pattern for their cities, from the Classic Maya to the Postclassic Zapotec and Aztec.
Tenochtitlan was one of the few Aztec cities that did not use these principles of city planning. It was founded after most Aztec cities, and unfortunately we know almost nothing about its early layout. I suspect that it was initially laid out like other Aztec cities, though. After the formation of the Triple Alliance empire in 1428, the Mexica kings set out deliberately to rebuild their city in the image of Teotihuacan and Tula (Umberger 1996). These three cities used different principles of city planning, what I have called the central Mexican imperial capital pattern (Smith n.d.). This included very densely-packed residences laid out using a common grid orientation with major avenues and highly-visible monumental temple-pyramids. In contrast to the normal Aztec civic center pattern, the rulers of Tenochtitlan walled off their temples and shrines in an inner sacred city, eliminated the large public plaza, and built numerous sumptuous palaces adjacent to the inner city, one for each ruler.
Returning to the smaller Aztec cities, one interesting feature is the dispersed nature of settlement outside of the civic core. Population densities for these cities average around 60 persons/hectare (Smith et al. 1994), which leaves considerable empty space within the city. Cuexcomate and Capilco are the only Aztec sites where residences have been mapped, thus permitting population estimates from house counts (Smith 1992). Capilco, a village of 20 houses, had a density of 100 persons/ha, whereas Cuexcomate, a town with 140 houses, had a density of 55 persons/ha. For the latter site, residential areas (houses, patios, and adjacent yard areas) cover only about one-third of the site area, leaving the rest open for cultivation or other extensive land-use activities. Although these sites are more densely settled than most Classic Maya cities (Culbert and Rice 1990), their openness does suggest that the “garden city” model of Maya urbanism may also apply to Aztec cities (Williams 1994). Little is known of Aztec urban housing, but the residences I excavated at the city of Yautepec were quite similar in size and construction to the rural houses of Capilco and Cuexcomate. In sum, residential patterns in Aztec cities may not have been very different from rural patterns, with small scattered commoner houses and house groups separated by gardens and fields.
The large size and central locations of temples and palaces at Aztec cities suggest that religion and adminisration were important urban functions, and this notion is supported by information on Aztec native views of cities (see Marcus 1983). These settlements, like most Mesoamerican cities, conform to Fox’s “regal-ritual city” model (Fox 1977), although this classification in itself does not advance our knowledge very far (Smith 1989). Unlike the Classic period, the scale of administration in the non-imperial cities is that of the city-state, not the empire. Ethnohistoric sources note the prevalence and importance of marketplaces at all Aztec cities, and excavations reveal a high level of imported goods at all of these settlements.
What is less certain is the importance of craft production as an urban activity. As mentioned above, Otumba was a large-scale producer of numerous craft items, including blades, bifaces, and jewelry of obsidian; ceramic figurines, incense burners, and other small items; basalt grinding tools; and cotton and maguey textiles (Charlton et al. 1991). Huexotla, on the other hand, exhibits little evidence of craft production beyond the ubiquitous Aztec domestic textile industry (Brumfiel 1980). At Yautepec, evidence for craft industries is stronger than at rural sites, but not nearly as abundant as at Otumba. Obsidian blade-production was a common activity, but no heavy deposits suggesting workshops or specialized dumps were found. Likewise, other production markers were recovered, but in modest quantities and from most excavations; these include ceramic molds for figurines and spindle whorls, bark beaters for paper production, and possible evidence for the working or even manufacture of copper/bronze tools. (Information about Yautepec can be found at:

Yautepec Excavations Page
Centripetal Forces
The Classic-period urban landscape was integrated by forces of centralization and hierarchy. Teotihuacan’s strong control of its hinterland was first manifest in the Tzacualli-period forced resettlement of peasants into the capital. When the countryside was later repopulated, villages and towns were established with a high degree of uniformity in size and layout. Non-imperial cities were laid out as imitations of Teotihuacan with grids, avenues, and apartment compounds (as Charlton 1991 points out, this similarity should not be pushed too far; few cities have all five traits, and the grid orientations do not always match that of the capital). The extreme primacy of the urban site-size hierarchy is another factor suggesting strong centralized political control by Teotihuacan. Blanton (1981) interprets the site-size distribution and other data as indicative of a politically-controlled solar marketing system centered on Teotihuacan.
After the fall of Teotihuacan in the seventh or eighth century AD, central Mexico went through a time of major social upheaval. During the Epiclassic period (AD 700-900) many rural sites and non-imperial cities were abandoned, and newly-established cities such as Xochicalco, Teotenango, and Cacaxtla grew into powerful capitals. Much of the population moved into defensible hilltop settlements, and themes of warfare dominated royal iconography at the capitals. The fall of the Epiclassic capitals brought about a “dark age” of low population levels and highly ruralized settlement (the Early Postclassic period, AD 900-1150) in all areas except the vicinity of Tula, where a large capital city headed up a modest local city-state. The supposed size, power, and grandeur of Tula and the Toltec civilization rest more upon Aztec propaganda than archaeological reality. The succeeding Middle Postclassic period (AD 1150-1350) saw the arrival of immigrant Nahuatl-speaking Aztec groups from north Mexico who established most of the major Aztec cities, city-states, and dynasties. These institutions rapidly evolved into the dynamic Late Postclassic Aztec civilization encountered by Hernan Corté³ in 1519.
Aztec civilization was far removed in time from the Classic period, and its development owed little to Teotihuacan or other Classic sites. The Aztec urban landscape was shaped by the less hierarchical forces of market exchange, city-state competition, and cultural uniformity. The Late Postclassic period saw a much more open and dynamic economy with little direct political control from Tenochtitlan. Population levels were much higher than in the Classic period, producing a much larger economic base in both rural and urban contexts. The numerous city-states were linked together through market exchange, elite interactions, and competition (a good example of Renfrew’s peer-polity interaction model). By the time Tenochtitlan rose to political and economic dominance in the Basin of Mexico, city-states, dynasties, and market systems were strong regional institutions in most areas.
Whereas Classic non-imperial cities were built to imitate Teotihuacan, in the Aztec period the rulers of Tenochtitlan deliberately redesigned the city to differentiate it from other cities by borrowing ideas from the ruins of Teotihuacan and Tula (Umberger 1996). The low level of administrative control of hinterland areas was partly a function of the indirect nature of imperial rule by the Triple Alliance empire, an example of a hegemonic empire (Berdan et al. 1996). Subject areas were only weakly integrated into the empire, often through the co-option of local elites, and this obviated the need for a strong imperial presence in the provinces.
The forces of centralization and hierarchy were not by any means absent from Aztec central Mexico. They played important roles at the level of the city-state, the dominant political form (Hodge 1984). Tenochtitlan was gaining in power and influence during its final century, and had begun to reorganize the political landscape of the Basin of Mexico for its own benefit (Hodge in Berdan et al. 1996). Nevertheless it was never able to reach the degree of political or economic domination achieved by Teotihuacan a millennium earlier. (It is sometimes asserted that Tenochtitlan or the Aztec empire had begun to decline in power in their final years, but there is no evidence for this idea.)
Centrifugal Forces
Centripetal forces of regional integration are typically balanced by regional centrifugal forces (Stein 1994). Unfortunately, these are difficult to analyze for the Classic period. The available archaeological evidence sheds little light on processes of autonomy and conflict prior to the destruction and collapse of Teotihuacan itself. How common was warfare? Did non-imperial cities ever try to break away from imperial control? What were the patterns of social stratification in hinterland areas?
For the Aztec period, the dynamic and highly volatile regional systems of city-state interactions mentioned above involved both centripetal and centrifugal forces. Market trade and elite interactions were friendly interactions, whereas competition, including both warfare and competitive emulation, was more of a centrifugal force. After areas were conquered by the Aztec empire, centrifugal forces manifest themselves in the many rebellions reported in the written record.
The most important centrifugal force of the Aztec period was social stratification. I have previously argued that Aztec native historical sources stressed political distinctions between states in order to mask the nature of social class divisions which were more fundamental social cleavages (Smith 1986). Individual city-state nobilities linked themselves into a single self-conscious class (in the sense of Mann 1986), and the rulers of the Aztec empire employed a number of mechanisms to exploit this elite class network for their own benefit (Berdan et al. 1996). Although we are starting to learn something of the conditions and activities of nobles and commoners in various parts of the urban landscape (Carrasco and Broda 1976; Lockhart 1982; Smith 1992), we still have little understanding of the operation of class and class antagonisms as dynamic social forces. Much of the ethnohistory is seriously biased due to the role of the postconquest nobility in providing information to the chroniclers, and archaeologists are only beginning to work out methods for analyzing social stratification and its implications.
I have tried to show that ancient central Mexican society can only be understood by looking at urban phenomena outside of the huge imperial capital cities of Teotihuacan and Tenochtitlan. These two cities were similar in many ways, partly because aspects of the Aztec city were deliberately copied from the Classic-period city. Nevertheless, these similarities can easily mask the fundamentally different patterns of political and economic organization in the two time periods. When we look at rural areas and at non-imperial cities, we see the contrast between the strongly dominant centralized power of Teotihuacan and the open complex market-based economy of the Aztec period. These were fundamentally different urban landscapes, and they had very different effects on peoples’ lives, whether in cities or villages. What is needed now is additional fieldwork at rural sites and non-imperial cities. I would guess that 95% or more of the archaeological fieldwork at Classic sites in central Mexico has taken place at Teotihuacan (the situation is better for the Aztec period, if only because most of Tenochtitlan is destroyed). Nevertheless, to really understand the capital city and its role, we need to move out of town and concentrate more research in the more distant parts of the urban landscape.

Dr. Michael E. Smith

Professor of Anthropology

School of Human Evolution & Social Change

Arizona State University