TWELVE
Relationships among Households in the Prehispanic Community of Mesitas in San Agustín, Colombia
INTRODUCTION
In the Alto Magdalena region, in southwestern Colombia, the development of communities at the core of small polities back to around 1000 BC have been traced in regional settlement-pattern surveys. Since that time, groups of households began to cluster together around places that were to become the central mounded funerary sites of the San Agustín chiefdoms during the regional Classic period (AD 1–900). What were the interrelationships among households within such central communities? What kinds of forces shaped and held together these communities while they became the central places of Classic period chiefdoms?
This chapter describes the reconstruction of the development of Mesitas, one of the biggest mounded prehispanic communities in the region. Various probable factors in the shaping of the community are evaluated. Resource control, population growth, and craft specialization do not seem to have been important for bringing about change in the sequence at Mesitas. However, the evaluation of these aspects suggests that the development of social differences among households is related to the very early clustering of some households around agricultural activities during a period when these activities were not crucial for subsistence. A traditional ritual role that some households held in the community since early times seems to explain in part the shape of the community and the greater differences among households later in the sequence. The term “household” is used in this chapter to denote remains of households as evidenced by clusters of artifacts, features, and architecture.
THE ARCHAEOLOGICAL STUDY OF THE MESITAS COMMUNITY
This chapter discusses internal dynamics in the long-term development of the prehispanic community of Mesitas from around 1000 BC to AD 900. During this long period the community—located in what is today the rural county of San Agustín, Department of Huila, in the Alto Magdalena region (Figure 12.1) of southwest Colombia—suffered transformations in size and shape but also in the importance of monumental burial mounds and the ritual activities they reflect. Regional settlement-pattern surveys, carried out since 1984 and totaling about 900 square kilometers, have documented the development of a number of similar communities in the Valle de la Plata and in the Isnos and San Agustín areas (Drennan et al. 2000) starting in the Formative 1 period (1000–600 BC).
These communities are the first societies to appear in the archaeological record of the Alto Magdalena region and are characterized by the widespread use of ceramic and lithic artifacts, the cultivation of a variety of plants, and a disperse settlement-pattern system featuring small residential households that reflect nuclear families (Blick 1993; Drennan et al. 2000; Jaramillo 1996; Llanos 1988; Quattrin 2001; Sánchez 1991, 2000). A chronology based on numerous radiocarbon dates for a number of ceramic types (Drennan 1993) permits us to divide the prehispanic sequence into five periods of different durations (Figure 12.2).
During the regional Classic period, beginning around 1 AD, the communities appear in the survey maps as concentrations of households that cluster together around groups of monumental burial mounds of the San Agustín culture (Drennan 1985; Drennan and Quattrin 1995). The analysis of settlement survey and excavation data from a large number of sites strongly suggests that during Formative times (1000 BC–AD 1) and the regional Classic period (AD 1–900), such communities emerged as the central places of individual chiefdoms or small polities (Drennan and Quattrin 1995), and that the more likely basis of political power of such chiefdoms, at least during the Classic period, was a religious ideology that linked the deceased leaders to supernatural beings in the funerary rituals (Drennan 1995). For understanding various important aspects of the long-term sequence in the region, the reconstruction of the chiefdoms’ development needed to be expanded to cover the social interaction at other levels, including the household and the community.
Figure 12.1. Map of Colombia, showing the Alto Magdalena region and the location of Mesitas in the region of San Agustín-Isnos (Reproduced, with changes, from Drennan 1995)
As a way to produce information at the community level and to complement the existing regional- and household-level information on the prehispanic trajectory of the Alto Magdalena, a number of studies1 since 1997 were focused on the Mesitas community, a settlement concentration around the Mesita A, Mesita B, and Mesita C monumental funerary sites (Drennan et al. 2000:93–117; González 2007). The studies included intensive surveys and programs of numerous small shovel probes designed to recover detailed information on household spatial distribution as well as samples of artifacts from each household located in an area of roughly three square kilometers.
Figure 12.2. Ceramic chronology for the Alto Magdalena Region (Drennan 1993) of Colombia
In this chapter, I discuss the relative importance that resource control, population growth, and craft specialization had in the development of chiefdom’s central places in the region, based on information recovered from the increasing number of households that made up the Mesitas community. These aspects had received some attention in earlier studies on chiefdom organization in the Alto Magdalena, but this work is the first attempt to study them at the level of the interrelationships among households within a central community. Using this perspective, I consider the Formative 1, Formative 2, Formative 3, and regional Classic periods at Mesitas, focusing on the relationships among households within this central community and on the forces that may have shaped and held it together while it became the central place of a regional Classic chiefdom.
Previous studies did not find strong evidence for the importance of economic and demographic factors. However, the importance of these factors at a more local level for shaping the development of chiefdoms has not been ruled out. Thus, the case of Mesitas provides a community-level perspective for evaluating the importance of possible factors in the development of chiefdoms in the Alto Magdalena. This perspective should help us understand better what households did inside developing chiefdoms and how important their activities were for the regional long-term trajectories.
CHIEFDOMS AND THE SAN AGUSTÍN CULTURE
The name “San Agustín” has been used by archaeologists working since 1914 to refer to a style of statues sculpted from igneous rock and placed inside burial mounds in dozens of sites in the Alto Magdalena region as well to a group of people who supposedly shared the use of this sculpting style and other cultural traits (Duque 1964; Duque and Cubillos 1988; Llanos 1988; Preuss 1931). A total of about 300 San Agustín statues, varying in height from 0.5 to 4.0 meters, have been found in some sixty elaborated tombs with stone-slab structures inside earthen mounds of around thirty meters in diameter (Sotomayor and Uribe 1987). Some of the statues still have the original, colorful painted surfaces (Cubillos 1980) that suggest the performance of elaborate funerary rituals during the construction of mounded tombs. The geographical distribution of the mounds with statues spans a vast region of about 100 kilometers in diameter (Sotomayor and Uribe 1987), but the megalithic tombs are much more common in a smaller, 300-square-kilometer area near the San Agustín and Isnos municipalities, where a few funerary complexes feature groups of several mounds per site (Duque 1964; Duque and Cubillos 1979). The largest funerary site is Mesitas, featuring a tight cluster of four groups of burial mounds where many of the known San Agustín stone statues are located.
Each San Agustín burial mound covers or is associated with a number of burials, but the mound itself is focused clearly on a particular individual who was buried in a stone corridor at its center, often inside a monolithic sarcophagus (Duque and Cubillos 1983). The statues, representing supernatural beings mixing human and animal features, were buried in front or on either side of the rectangular stone-slab corridor that contained the tomb. The layout of this kind of monumental tomb, associated with regional Classic ceramics (AD 1–900), seems to express a relationship between the person buried in the grave and the religious or mythical beings depicted in the sculptures, suggesting an individually oriented political organization based on a strongly religious ideology (Drennan 1995).
No monumental burial mounds from Formative times have been systematically excavated yet, but a wooden sarcophagus found in a looted tomb of the Alto Lavapatas in the 1930s was dated to around 500 BC* (Duque and Cubillos 1988:107). Continuity in the general patterns of settlement distribution through time (Drennan and Quattrin 1995) suggests that the specific places with megalithic ceremonial complexes built during the regional Classic period were already centers of activities attracting population well before 1 AD.
What were the bases of power during Formative times, how the political organization of regional Classic chiefdoms came about, and how the ceremonial activities in the central communities were related to the development of complex societies are questions that still remain to be fully addressed. Such questions will need new kinds of archaeological studies, focused on analytical levels where the information is still scant, such as the community level.
Some studies of chiefdom development (Blake 1991; Rogers 1995) previously linked community-level processes around civil and ceremonial centers with the regional development of chiefdoms. This level of analysis allows investigating internal dynamics of chiefly developments (Drennan et al. 2000; Price and Feinman 1995) and offers the possibility of linking household-level dynamics with region-wide processes (Bermann 1994:10–18; Henderson and Ostler 2005; Peterson 2006).
FACTORS OF CHANGE: THE CONTROL OF CRITICAL RESOURCES
The literature on the development of chiefdoms places much emphasis on the role of resource control (Earle 1987a; Gilman 1981), contradicting a more traditional view that sees chiefdoms as characterized by authority without real power and lacking the legitimized force that lies at the base of the economic integration in states (Fried 1967; Service 1962; Steponaitis 1978:420). Thus, special attention has been given to the idea that chiefdom development rests on economic control of basic resources by an elite group (Earle 1991a; Johnson and Earle 2000:253). Direct control of access to productive land, for example, has been proposed as a basic factor for chiefdom development (Earle 1991b:71–72; see Douglass and Gonlin, this volume). However, in the Alto Magdalena chiefdoms, at the regional level, population seems rather low and dispersed, even if some settlement concentrations do exist. Spatial analysis of such concentrations has shown no correlation with agricultural land or natural resources. This suggests that social and political reasons prevailed over economic factors in determining the location of small polities during Formative and regional Classic periods (Drennan and Quattrin 1995). Although burial architecture and variation show clear social differentiation by the regional Classic period, the number of offerings does not reflect clear wealth differences (Drennan 1995). The horizontal excavation of domestic deposits has shown only minimal economic differentiation (Blick 1993; Jaramillo 1996).
Control of resources in the region, then, was probably not strong. However, small differences in the access to basic resources at the interior of individual communities could have been important in shaping the development of social hierarchy in the Alto Magdalena, and this information likely would not come from more horizontal household excavations but from the systematic study of the artifact variation across whole communities.
FACTORS OF CHANGE: POPULATION GROWTH
Another common factor considered to be important in chiefdom development is population growth, along with the pressure it can create on resources. Demographic indicators do correlate with some of the variables that are often used to identify social complexity (Feinman and Neitzel 1984:75–78). Circumscription and warfare (Carneiro 1981) and agricultural intensification (Boserup 1965) have been proposed as ways in which population pressure triggers sociopolitical transformations. In some areas, chiefdoms seem to have developed in a context of population growth and circumscription to harsh environments (Kirch 1984:216–23; Sanders, Parsons, and Santley 1979:369–85). In the Alto Magdalena, regional population did grow through the sequence, especially from the Formative 3 to regional Classic periods (Drennan et al. 1989:151), providing some correlation of population growth with changes in social and political organization, but the relations still remain unclear. Even at demographic peaks not all potential agricultural land appears to have been utilized (Drennan and Quattrin 1995; Drennan et al 2000).
An aspect of demographic dynamics that may be important at the community level is household size. Elite households might have been structured differently from non-elite households. Larger group sizes have been related to resource mobilization by emerging chiefs (Muse 1991) since larger households seem to have economic advantages that come from having larger labor pools (Henderson 1998; Webster 1990). Household size might also be related to part-time specialization since it would require specific kinds of labor organization. A deeper study of local patterns at the interior of evolving chiefdoms in the Alto Magdalena is then an adequate base for evaluating the importance of demographic change in the development of social differentiation and political centrality.
FACTORS OF CHANGE: SPECIALIZATION OF CRAFT PRODUCTION
The idea that resource diversity can result in some kind of specialization that creates new managerial requirements has been proposed to explain how chiefdoms evolved (Service 1962). Specialization of production could also create new opportunities for the accumulation of wealth and redistributive economies (see Douglass and Gonlin, this volume). No evidence of specialized agricultural production has been found yet in the Alto Magdalena at the regional level (Drennan and Quattrin 1995; Drennan et al. 2000; see also chapters by Neff and Ciolek-Torrello, this volume) or at the level of the household during Early Formative times (Quattrin 2001). The subsistence base was broad in both cultivars and wild species, and a wide array of plants, including maize, were produced and consumed at all elevations (Quattrin 2001:81).
Craft specialization and the involvement of local leaders in prestige-goods exchange networks have been proposed as ways in which political and economic control developed in chiefdoms (D’Altroy and Earle 1985; Earle 1987b:67–69; Helms 1987; Spencer 1993; Steponaitis 1991). Craft production is a clear possibility for specialization in the Alto Magdalena. Specialization in the production of statues, gold objects, and lithic and ceramic artifacts could have been one factor for shaping these societies, as some relationship of San Agustín elites to production of esoteric knowledge has been linked to the homogeneous character of styles of elite goods in southwestern Colombia (Gnecco 1996).
However, regional analysis in the Alto Magdalena region indicates that ceramic production was not subject to regional control in any prehispanic period (Taft 1993:165–171) and household archaeology in the Valle de la Plata (Blick 1993; Jaramillo 1996) did not produce direct evidence of craft specialization during Formative period times. Data on community organization then can provide a better basis to evaluate how significant craft specialization was for the internal organization of societies in the Alto Magdalena.
THE RECONSTRUCTION OF THE MESITAS COMMUNITY
The reconstruction of the archaeological community of Mesitas took advantage of different kinds of information and various levels of analysis. The monumental burial mounds of Mesitas and some surrounding areas have been horizontally excavated for several decades, producing detailed information mainly on funerary architecture—but also on artifact diversity, stylistic change, and preservation of materials—and numerous radiocarbon dates and some information on residential architecture (Chávez and Puerta 1988; Duque 1964; Duque and Cubillos 1981, 1983, 1988; Llanos 1988; Llanos and Durán 1983).
Archaeological communities appear in the survey maps of the Valle de la Plata region as defined clusters of remains of households spaced at distances of approximately ten kilometers from each other. Since those clusters do not correspond to environmental characteristics, they have been interpreted as the relatively dense communities that were centers of small polities (chiefdoms) whose territories spanned some 100 square kilometers (Drennan and Quattrin 1995) and had populations of 4,000 to 8,000 (Drennan et al. 2000). Each of the central communities in the Valle de la Plata region was located during the regional Classic period around a complex of monumental burial mounds with stone statues of the San Agustín style, suggesting a strong relationship between religious ritual and regional political organization.
Located directly to the south of the Valle de la Plata region are the Isnos and San Agustín municipalities, where a 300-square-kilometer region including most San Agustín burial complexes was surveyed from 1993 to 2007 by the Programa de Arqueología Regional en el Alto Magdalena (PARAM),3 completing settlement-pattern studies for a total surveyed area of approximately 900 square kilometers as a base for reconstructing demographic dynamics, settlement distribution, and regional political and economic organization (Drennan et al. 2000:41–92). An intensive survey of the Mesitas area was conducted in 1997 as part of the PARAM program. In this survey, 300 1 by 1 meter stratigraphic excavations delineated the general shape of the household clustering around the monumental burial complex and also produced basic chronological information (Drennan et al. 2000:98–99).
Reconstructing the community itself required us to identify and study all the individual household remains that formed the community at the center of the cluster identified in the regional survey (González 2007). The fieldwork methodology for such a reconstruction took advantage of earlier projects in the region in which systematically placed excavations of 40 by 40 centimeters had proven successful in reflecting the actual subsurface distributions of archaeological remains (Blick 1993; Drennan 1985:137–143; Jaramillo 1996; Quattrin 2001). An area of 275 hectares (Figure 12.3) was selected following the settlement concentration discovered in regional surveys around the concentration of mounded funerary sites and also adjusting the boundaries to the specific shape of the local terrain. Inside the selected area, linear transects spaced thirty meters apart and with one shovel probe each five meters were systematically placed so as to cover all relatively flat terrain of the study area. Each of the artifact samples obtained in 2,000 probes was analyzed to produce maps with the distribution of sherds of ceramic types for each specific period (Figure 12.4). For each period, the specific clusters of sherds and other artifact classes that were at least twenty-five meters wide and were located on and around relatively flat areas were considered to be household locations. These transects permitted the identification of seventy-six specific household locations with 150 separate occupations (González 2007).
Figure 12.3. Map of the Alto Magdalena region showing mounded sites and the Mesitas area
Grids of probes formed by placing twenty-five to thirty additional shovel probes on and around each household location were used to obtain larger samples of each household’s artifact assemblage and also to get information on the approximate sizes of each household cluster. Every artifact collection obtained in the shovel-probe program at the selected areas represented, in this way, a separate household. Data analysis focused then on the differences and similarities among contemporaneous households.
Figure 12.4. Four maps of the Mesitas area showing the location of all the probes with each type of ceramic sherds; symbol size represents sherd frequency
The relative status of a household was indicated, in part, by spatial location within the community. A denser core directly associated to the concentration of mounds contrasted with the less dense periphery (Figure 12.5). In general, some households clustered near the Mesita B mounded site seem to have had a special status in the community. Indication of status for some households in the denser core during the regional Classic period includes the spatial association with the most impressive burial mounds in Mesita B and Mesita A sites and the almost exclusive use of certain types of ceramic decoration. Additional indications of higher status for the core came from finding evidence of more serving than cooking vessels and concentrations of ornaments such as hexagonal stone beads. In Formative times, before the burial mounds were built, status differences can be seen in some aspects of the ceramic assemblages. During Formative 1, for example, Household 68, near what later became the mounds of the Mesita B site, had more serving vessels than cooking vessels. Some indication of household wealth was seen in higher proportions of sherds that featured relatively costly decorations and in the greater proportions of fine materials that seem harder to obtain, such as obsidian, which probably reached the region by way of an exchange network (Jaramillo 1996:125–126) and from separate sources (Hurliman 1993). Households in the core also had, in general, more ornaments and a higher frequency of some rare types of decoration on ceramic vessels (González 2007). Thus, when comparing household characteristics, special attention was given to their location within the community in terms of an apparently high-status core and a lower-status periphery.
Figure 12.5. Map of the Mesitas area showing the location of the 76 specific household locations, the core of the community and the location of mounded funerary sites
Figure 12.6. Four maps of the Mesitas area showing the location of households for each period
The number of households in the community, as reconstructed in the systematic probe program, grew from six in the Formative 1 to thirty-one in the Formative 2, thirty-eight in the Formative 3, and finally seventy-five in the regional Classic period (Figure 12.6). The obvious demographic growth of the community and the association of the Mesita B site with a relatively wide area of flat, arable land suggested at first that demographic dynamics or the control of land could have been important in the development of the regional Classic central community. Alternatively, such a center could have grown around the centralized production of certain crafts. The spatial distribution of households and the specific artifact assemblages of each household were then used to evaluate these possible factors in the consolidation of Mesitas as a chiefdom’s central community.
EVALUATING THE CONTROL OF LAND
The importance of the first factor, control of critical resources, was evaluated by looking at the relationship of household locations to land for each of four periods. Investigating the distribution of land among households consisted of a GIS-based catchment analysis that measured and compared associated amounts of productive land. Catchment-area analysis has proven useful in assessing productivity differentials in non-market societies (Brumfiel 1976; Steponaitis 1981; Vita-Finzi and Higgs 1970). In this study, information on local soil productivity (Neira 1996) was combined with precise household location to estimate the area of prime agricultural land that could be reached within catchment circles of a 0.3-kilometer radius around each household. Assuming an association of households to the very close agricultural plots around them is supported by the reconstructed general regional settlement pattern that shows dispersed and continuous occupation of the landscape with no areas devoted exclusively to agriculture (Drennan et al. 2000). During Formative 1, the community was formed by only six households. Three households grouped together on the best piece of available land, while the other three households were isolated. One of the households at the core had direct access to 40 percent of the agricultural land in the study area inside catchment areas (González 2007). It is unlikely, however, that control of this fertile land gave households a critical advantage because the community was so small and the agricultural resources readily available and probably abundant in the local vicinity.
During the Formative 2 period, the community had thirty-one households, and the core was much denser than before. However, there were still a number of locations that offered abundant productive land and remained unoccupied. Some households did have more productive land than the average within their catchment areas, but land was so abundant inside the core that the nucleation there did not result in measurable productivity differentials between zones. This suggests that each household—whether on the periphery or in the denser core—had a sufficient land to fulfill its needs.
The Formative 3 community, with thirty-eight households, was not much larger than the previous period’s community, but nucleation increased at the core, resulting for the first time in significant—but not very big—differences between the core and the periphery in terms of household agricultural productivity. At this time, households at the core might have wanted some of the surplus that households at the periphery could more easily produce. The clustering of households in these conditions suggests that high-status households were successful in attracting followers and perhaps also surplus in the form of tribute from the periphery and beyond. Households in the periphery are associated with greater agricultural productivity and might have been “wealthier” than households at the core (González 2007). However, the amounts of unoccupied productive land that still existed during the Formative 3 period outside the core suggest that land was not a critical resource.
Table 12.1. Comparison of agricultural land inside household catchment areas between the core and the periphery of Mesitas for four periods
Core | Periphery | |||||
Period | N | Mean (ha) | Std. Error | N | Mean (ha) | Std. Error |
Formative 1 | 3 | 4.00 | 1.79 | 3 | 2.15 | 0.73 |
Formative 2 | 18 | 1.34 | 0.30 | 13 | 1.41 | 0.32 |
Formative 3 | 27 | 1.09 | 0.22 | 11 | 1.54 | 0.25 |
Regional Classic | 41 | 0.80 | 0.12 | 34 | 0.95 | 0.13 |
Source: González 1998. | ||||||
During the regional Classic period, the Mesitas community had seventy-five households. Several areas at the periphery were occupied for the first time and the core became somewhat denser, but the differences in agricultural productivity between the core and the periphery were less contrasting than they were during the Formative 3 (Table 12.1). This pattern suggests that during the regional Classic period all households had more equal access to agricultural land. In terms of direct access to land, it can be shown that throughout the sequence that one-sixth of the Mesitas households had easier access to almost 40 percent of the land inside catchments of 0.4 kilometer in diameter. However, the unevenness of this distribution did not change from Formative 1 to regional Classic, so any attempt on the part of elites to restrict access to land seems unlikely.
Interestingly, the average amount of land per household did decline during the regional Classic period for the whole community to about one hectare of productive land inside each household’s catchment area. Productive land might have become by the regional Classic a much more critical resource than it was before. Agricultural activities could have been more important during the regional Classic period, especially for households in the denser core. During this period, households of apparently high status were concentrated in the core near the monumental tombs of obviously important individuals. These households were apparently larger than households on the periphery. Thus, land during the regional Classic period was more limited than before and agriculture might have become a much more critical activity, especially for households in the core. But up to this time, no evidence of restricted access to land was seen in the land distribution and the productivity differentials did not intensify over time.
The data on land distribution suggest that direct control of land is not likely to have been a foundation for elite authority at Mesitas. These data do not support the idea that emergent elites directly controlled subsistence resources. The findings are consistent with the results of regional surveys in the Valle de la Plata that found no relationship between site distribution and the distribution of productive soils (Drennan and Quattrin 1995:219–220; see also Douglass and Gonlin, this volume). What is interesting about the data is that households at the core seem to have attracted households to nucleate near them to share some specific tracts of agricultural land despite the opportunity for any household to locate its residence on isolated but also productive lands.
The distribution of Mesitas households suggests that other, untested aspects of agricultural production may have had a greater significance in the trajectory of change than control of land. These aspects include the formation of a cooperating group. Beginning in the Formative 1, denser settlement at the core may reflect the practice of intensive agricultural practices by elite groups (González 2007). Communal activities are suggested by the presence of unusual sherd concentrations in some specific areas in the core that are different from the refuse produced by normal household activities in Mesitas (see Figure 12.4). The identification of such groups in Mesitas in a context of low population levels supports the idea that in chiefdoms labor, not land, was a main factor limiting production (Earle 1987a:293; Friedman and Rowlands 1977; Webster 1990). Any agricultural intensification present during the regional Classic period or earlier was probably not needed for subsistence but may have been instrumental in expressing differences in rank or prestige.
EVALUATING DEMOGRAPHIC CHANGES
The impact of a second factor, population growth, was evaluated by analyzing the timing and extent of demographic changes in the community. To do this, the number of households and their relative sizes from the Formative 1 period to the regional Classic period were reconstructed. The increase in number of households shows a general relationship between local population growth and the consolidation of chiefly authority. Moreover, the tremendous surge in sherd densities for the regional Classic and the greater nucleation at the core opened the possibility that the number of identified households underrepresented the total number of people living in the community and that population growth could then be a considerable force in shaping the community (González 2007).
To reach better estimates than those based on only the number of identified households, the demographic structure of the community was reconstructed for each of the four periods using not only this number but also two different indices reflecting group size: the size of the sherd concentration representing each household and the density of sherds in each group of probes representing each household. The estimates were then adjusted to the different lengths of each period.
Table 12.2. Mean household deposition rates by period at San Agustín
Period name | Period length (years) | Number of sherds | Number of households | Mean deposition ratea |
Formative 1 | 400 | 814 | 6 | 0.339 |
Formative 2 | 300 | 3,593 | 31 | 0.386 |
Formative 3 | 300 | 3,348 | 38 | 0.295 |
Regional Classic | 900 | 32,838 | 75 | 0.487 |
Note: Data obtained from a total of 5,057 shovel probes in the Mesitas area (González 1998). | ||||
a Number of sherds per household per year. | ||||
The analyses of household size reflected clear but small changes in the size of the households in the core during the Formative and regional Classic periods, suggesting that elite households had only slightly larger labor pools in the former period and that these households were located exclusively inside the community’s core. Household size did not show any significant change at the periphery. Incorporating the small changes in household sizes for the regional Classic period reflected by sherd densities and the sizes of artifact distributions, our estimates for the size of the community are 30–36 members for the Formative 1, 200–230 persons for both the Formative 2 and Formative 3 periods, and a final surge in size to 700–900 people in the regional Classic period. This growth, however, happened over a very long period and could have been the result of extremely low annual rates. The recovery of a higher number of sherds for the regional Classic period does not need to reflect a rapid surge in population. The comparison of mean household deposition rates by period (Table 12.2), adjusted for the length of each period, shows similar growth rates throughout the sequence. The regional Classic period does show, however, a higher sherd-deposition rate than for earlier periods. This slightly higher mean deposition rate supports the idea that some households in this period were larger, but the increase in this rate does not reflect drastic demographic changes.
There is at least a correlation between the clearer indications of social hierarchy of the regional Classic and population growth. A good part of this growth at the community level however, has to do with an increasing tendency for nucleation near the core of the community and that tendency was present from before, even between Formative 2 and Formative 3, when there was no observable population growth for the whole study area.
Given these varying rates of growth, any apparent population pressure appearing in the regional Classic at Mesitas should be seen as a result, rather than a cause, of the developing social hierarchy. Moreover, the moderate population densities found in what was probably one of the densest settlements in the region and in the southwest of Colombia give strong support for the conclusion of regional-level studies that carrying capacities were far from being reached (Drennan and Quattrin 1995). Thus, the information on the demographic changes at Mesitas does not support the idea that population pressure (Boserup 1965; Carneiro 1970; Cohen 1977) was the central force shaping chiefdom development in Mesitas.
The information from the Mesitas area, one of the most densely populated areas in the region, even when it shows some increase in household size at the core of the community, is entirely consistent with the results of earlier regional studies that show demographic growth in the Valle de la Plata during the regional Classic period but without any clear evidence of population pressure creating subsistence problems (Drennan et al. 2000), supporting the general idea that demographic shifts should be seen more a consequence than a factor in the development of complex societies (Cowgill 1975). Regional studies in other areas of northern South America have also reported chiefdom development in a context of population growth with very low population densities (Langebaek 1995; Spencer 1990).
Instead of creating subsistence problems, it is possible that greater nucleation at the Mesitas regional Classic center was a result of a successful system of tribute relationships that could have served in Mesitas to balance any pressure on the local resources at the core. Given that the average amount of land per household, even during the regional Classic period, was not terribly low, any tribute flowing from the periphery to support activities at the core did not need to be especially high (González 2007) to support the chiefly activities at the center.
EVALUATING CRAFT SPECIALIZATION
The significance of our third factor, craft specialization, for the changes in Mesitas was evaluated by analyzing changes through time in the ranges of activities performed by different households in the community. For each period, the distribution of different categories of lithic and ceramic assemblages was analyzed and comparisons of the household assemblages were made across the community. The spatial concentration of small numbers of households with particularly high densities of specific and rare tools or materials was considered an indication of household craft specialization. If the basis of elite distinction at Mesitas was the effective control over the production and acquisition of social valuables, we would have expected a direct association of centralized craft production in the households of greatest status in Mesitas.
Analyses of the Mesitas data did show that evidence for craft production is more common at the core than at the periphery (González 2007). The analyses, however, also show that they are not absent from the periphery, and in some cases, they are related to households of apparently lower status. This pattern suggests that elites did not have exclusive control over craft production.
Some kind of concentration of craft production, for example, is reflected in the lithic assemblages for the southwest corner of the study area. Households in this sector have the greatest concentrations of lithic artifacts made of slate. During the Formative 1, these households also show proportions of decorated vessels that might indicate wealth, but they did not show clear indications of high status in the assemblage of vessel forms. The existence of some polishing-stone and lithic artifacts in the assemblages also shows evidence for craft specialization in the core. Greater densities and greater proportions of some types of polished lithic artifacts suggest that the grouping of households near Household 68 was involved in a specific kind of craft production, perhaps the elaboration of gold ornaments, as suggested by the discovery, in this same location, of gold-plate fragments and gold drops in the second level (40 cm) of an excavation by Luis Duque in 1960 (Duque 1964:242).
During the Formative 1, Formative 2, and Formative 3 periods, artifact assemblages indicate one single special location also near the Mesita B with high sherd densities (González 2007). These deposits were interpreted as evidence for supra-household activity, perhaps communal parties or ceremonies. But there are also two other distinct groupings of households within the core, not associated directly with a monumental mound, that had the greatest concentrations of chert, obsidian, polishing stones, axes, and other tools or materials. This suggests that some craft activities were undertaken by other households and not only by the ones near the Mesita B or by any other specific group of households. In fact, a northeast/southwest distinction seemed to separate the core into two areas with different sets of craft activities. In each of these groupings, some households show some signs of different kinds of craft specialization.
The pattern in the core suggests that specialization could have been centralized in Mesitas along kinship lines. Three lineages represented by three separate groupings of several households might have associated themselves with different aspects of craft production. Household 68, near the Mesita B site, always shows some sign of high status in the community and seems related to some kind of supra-household activities. This particular household also shows evidence of craft production in the form of stone cylinders and other types of polished stone that were found only there and that could be used for a specific kind of craft production. Household 68, however, did not exclusively control craft production since other households also participated in other types of craft production. For example, Household 37 may have more intensively manufactured pottery since many more polishing stones were found there than in any other location. Households 37, 39, and 60 had many more chert and obsidian tools than any other household and were probably involved in craft production. Households 60 and 66 had higher numbers of manos and metates than expected, also suggestive of specialized activities. Thus, no elite group at Mesitas seems to have had an exclusive control over the production or acquisition of crafts or valued objects, which does not support the idea of craft production as a basis for chiefdom organization (Schortman et al. 1992:4). Other activities, such as the involvement of Household 68 in communal agriculture and redistribution, seem to have been more central to the organization at Mesitas than craft production.
To conclude, none of the factors selected for the study of the community of Mesitas (control of land, population growth, and craft specialization) seems to have been instrumental in bringing about long-term changes. Alternative factors yet to be evaluated but suggested by the information collected in Mesitas thus far include the formation of early cooperative groups during the Formative 1 period (1000–600 BC), which might have given households near Mesita B a preeminence that made them central to the community. Their traditional role as coordinators of agricultural and ceremonial activities, for example, might have made them increasingly prestigious. During Formative 2 (600–300 BC) and Formative 3 (300 BC–AD 1) times, when the community numbered approximately 250 persons, the increased importance of agriculture and the coordination of supra-household activities might have given the traditional leaders increased authority but not much economic power. The first clear indications of economic differences appear during the regional Classic period, when households in the core are larger than on the periphery, and when the size of the community would have necessitated intensifying some aspects of agriculture. Some households in the core seem to have involved themselves in craft production and obsidian exchange. The centrality of the households of greatest status during the regional Classic, however, does not seem to have come from those activities. Instead, it might be that whatever importance they had came from the traditional and ritual roles that some households played in the community as early as the Formative 1. Perhaps these roles were first related to the organization of cooperative groups and other communal activities, and later to the religious ideology reflected in the regional Classic monumental tombs.
SUMMARY AND CONCLUSIONS
Based on the studies of the Mesitas community, it is concluded that none of the three selected factors seems to have been important for bringing about change in the sequence. Agricultural land slowly became important for the community but was always abundant and never associated exclusively with elite households. During the trajectory, the community clearly grew in size, but such growth could not put any serious pressure on basic resources. Some households seem to have been involved in intensive craft production, but this production was not exclusively associated with clear markers of social status. Elite and non-elite households alike were involved in craft production that does not seem to have been centralized.
What then were the forces or the factors involved in the development of the small Formative community of Mesitas as a regional Classic chiefdom center and monumental burial site? An alternative force yet to be evaluated in Mesitas is the control of ideology and religion. This force does appear to have been prominent during the regional Classic and is clearly expressed in the single-burial monumental tombs featuring stone statues of fantastic beings. The lack of evidence for economic differentiation, specialization in production, or demographic problems in conjunction with the association of clusters of households to monumental sites of high-status individuals points to a centralized political system with a clear social hierarchy based on a belief system that associated individual leaders with supernatural beings (Drennan 1995).
The evaluation of the construction or consolidation of such a belief system might be approached in several different ways. I propose to approach it from the point of view of the individual households in a community and their interrelationships. Reconstructing what the individual households did in the community during the Formative and Classic times will help us to understand how such a social hierarchy developed or became consolidated. In the process, this reconstruction will give us more examples of the activities the households perform as social units and the utility of this concept for the archaeological research of social interaction.
The archaeological study of Mesitas has not yet found specific activities explaining the separation of elite households from the rest. However, the studies have located and characterized all households in the community for the Formative and regional Classic periods and have detected status differences. More intensive study through horizontal excavation of a sample of these households in the future could show in more detail any contrasts between them in terms of activities that may have permitted elites to gain influence and authority.
The study of the community of Mesitas has reconstructed thus far the distribution of domestic refuse in relationship to topographical features and the spatial distribution of monumental tombs. However, this information has not revealed direct evidence of the relations between specific households and the activities, including ceremonial ones, proposed to be the focus of chiefly organization at the core of the community. Obtaining such information for a settlement that spans close to three square kilometers cannot depend on only future stratigraphic excavations. Current research in Mesitas is applying remote sensing (magnetometry and GPR) to map the shape of the household structures located earlier by reconstructing refuse distributions around them and also buried features that may link specific households to their roles in the community, such as roads, burial concentrations, and floors and structures without the domestic refuse that would represent ceremonial houses. By analyzing this additional information about Mesitas and comparing the sequence to the development of similar chiefly centers in other regions, we expect to understand better how elite households constructed and institutionalized social hierarchy in developing chiefdoms.
NOTES
1. The dissertation research project on which this article is based was financed mainly by the National Science Foundation (grant SBR-9632597), the Wenner-Gren Foundation (grant Gr. 6097), COLCIENCIAS, and the Instituto Colombiano de Antropología e Historia.
2. Scientific exploration began in 1857 when geographer Agustín Codazzi described some of the mounded sites but the earliest archaeological excavations were carried out by Konrad T. Preuss in 1914. The most impressive burial sites were excavated from 1940 to 1977 by Luis Duque Gómez and Julio Cesar Cubillos.
3. The Programa de Arqueología Regional en el Alto Magdalena is a collaborative research program of the University of Pittsburgh, the Universidad de Los Andes, and the Instituto Colombiano de Antropología e Historia in Colombia, financed in part by the National Science Foundation.