THIRTEEN

Interhousehold versus Intracommunity Comparisons

Incipient Socioeconomic Complexity at Jachakala, Bolivia

CHRISTINE BEAULE

INTRODUCTION

Most prehistoric agro-pastoral households farmed, herded domesticated animals, witnessed or perhaps hosted rituals, produced and consumed trade goods, and made a variety of utilitarian implements. This range of activities made up the domestic economy, which can be divided into production (control of food and craft goods, subsuming farming, herding, storage, processing, and manufacturing), service (payments rendered to elites for their support or for community-level functions within the context of a fully developed political economy), and distribution activities (control of resources circulating through interhousehold, intersite, or interregional exchange networks) (Hirth 1996:209).

Access to a range of resources, such as land, kin-based or communal labor, raw materials, and individual talents and personalities, in addition to factors of a particular ecological setting, shape to greater or lesser degrees optimal strategic options for fulfilling each household’s demands. However, the domestic unit’s primary goal in non-complex societies is the basic accumulation of sufficient provisions to at least minimally support its residents, although subsistence needs are variably defined cross-culturally (Hirth 1993a:22). The archaeological record presents scholars of ancient household studies with abundant opportunities to reconstruct the organization of domestic life, changes in its organization, and the diverse ways in which households go about fulfilling their needs. How we achieve these goals depends on our conceptualization of the domestic economy as well as methodological approaches suitable for our particular research objectives.

Artifact assemblages recovered from domestic structures, features, and deep stratigraphic excavations at the Bolivian highland site of Jachakala (ca. AD 170–1100) document the gradual development of interhousehold wealth differentiation. Jachakala is a small site (6.72 ha) in the La Joya region of the Department of Oruro in the central altiplano (the high, cold, flat plain between the two chains of the southern Andean mountains). Deep stratigraphic excavations there in 1997–1998 provided abundant household architectural remains and artifacts employed in the reconstruction of the long-term local history of the community. The three-period chronology employed throughout the project relied on local economic changes through time rather than on ceramic evidence dating the community’s first contact with the Tiwanaku state. Specifically, site-wide changes in the relative proportions of camelid faunal packets and basalt hoe debitage were used to define the transitions between the Niñalupita (ca. AD 170–500), Isahuara (ca. AD 500–800), and Jachakala (ca. AD 800–1100) periods. In this way, changes in the production portion of the domestic economy were used to chronologically define those artifact assemblages used to test project objectives. Jachakala’s chronology corresponds roughly with the Late Formative and Middle Horizon periods in the south-central Andes.

The model employed in this project attempts to articulate the origins of complexity with shifts in the domestic economy of these prehispanic Andean households. Artifact patterns related to these changes are explored in this chapter in two ways. A comparison of those two methods of exploring changes at Jachakala, rather than the details of the model itself, are the focus of my argument.

CONSTRAINTS ON ECONOMIC DIVERSIFICATION

Household composition directly reflects subsistence practices and the range of available economic opportunities. Variation among contemporaneous households will be predictably small, according to Kenneth Hirth (1993a:23), where few opportunities for economic expansion exist, “except under circumstances of dramatic sociopolitical reorganization or environmental change (draught, famine, plague, etc.), resulting in long-term stability in form and composition.” Incorporation into a larger political system would also provide opportunities for economic reorganization on the household and/or community level.

In the absence of such circumstances, the stability of the domestic economy stems from households’ inability to procure sufficient resources to create or take advantage of opportunities to fundamentally change their economic organization. While elite households in well-established political economies could generally finance alternative adaptations, subsistence-oriented households typically cannot (Hirth 1993a:23; Douglass and Gonlin, this volume). A few of many possible factors responsible for the long-term stability of non-elite households include their preference for traditional production strategies, deep-seated beliefs about the social composition of households and the various roles of their members, and the limited range of economic opportunities in a given environmental setting.

Constraints on economic growth at the level of the household, however, do not preclude a degree of interhousehold variation. In this sense, Chayanov’s picture of peasant farmers who require external stimuli to intensify their subsistence production is far from the rule (cf. Netting 1993). Intensified domestic production in non-complex societies may include increased agricultural production, part-time craft specialization, and increased participation in or reliance on interregional exchange systems. Furthermore, many archaeological and ethnographic case studies have documented such domestic intensification in the absence of external stimuli such as environmental or population pressure (Brumfiel 1994; Clark and Blake 1994; Hayden 1990).

This is not to say, of course, that agro-pastoral households are incapable of economic diversification. Except under conditions of extreme population pressure or environmental stress, agrarian households have the ability to produce a bit more (and consume a bit less) food than they need. Similarly, the ability to store extra subsistence products long enough to accumulate a fund that can be spent in support of non-domestic pursuits (potential wealth-building activities such as cash crops, craft production, strengthened trade ties, etc.) is not necessarily a hindrance. In the Andean highlands, for instance, tubers are freeze-dried to make chuño, and meat may be stored long-term as charqui (jerky). Although preservation of stored subsistence goods elsewhere might be a bigger challenge than in the arid altiplano environment, nothing much more than time is needed to accumulate extra food to allow for experimentation with non-domestic activities, at least in theory. However, the unpredictable nature of agro-pastoral resource bases, particularly in such marginally productive environments as the altiplano (because of the high risks associated with overnight frosts and highly variable precipitation levels), is well known to anthropologists and modern Bolivian Aymara and Quechua peasants alike. This factor alone could make agrarian households unwilling to count on stores of extra food while they devote more time to non-domestic pursuits.

Given their subsistence orientation, the size and internal organization of households will vary mostly with the amount of land, animals, and labor available to each. One can frame the issue in terms of cultural conservatism, labor productivity, and time allocations (Netting 1993:105–109) or domestic (subsistence-oriented) economy. Regardless, agro-pastoral households are generally unlikely to expand or radically restructure their domestic economy to the extent that archaeologically recognizable wealth differences emerge in the process without some compelling reason to do so. Of course they do so in many cases, and these are some of the archaeological moments of interest to us. Investigating how and why this process of economic differentiation occurs presents a number of methodological opportunities to household archaeologists in particular.

THE HIRTH MODEL OF THE ORIGINS OF COMPLEXITY

The model of the emergence of socioeconomic complexity from changes in a domestic economy that was tested against the La Joya region’s archaeological record is adapted from Hirth’s (1993a, 1993b, 1996) work in the central Mexican highlands. Briefly, the “Hirth model” of political economy states that differential participation in activities such as trade and craft production can lead to increased variation in the domestic economy, organization, and size of households. This interhousehold variability forms the economic underpinnings of the political economy. The consequent development of rank or status differences among households follows from differential access to key resources such as raw materials and long-distance trade goods.

The potential for interhousehold differentiation rises then with the introduction of economic opportunities outside the traditional agricultural sector, such as an increasing focus on craft production for exchange and wage labor pursuits. This approach, also advocated by numerous other researchers (Blanton et al. 1982; Flannery and Winter 1976; Sanders, Parsons, Santley 1979; Wilk 1984, 1990), complements models equating political development and elite means of controlling certain facets of the production process. Although many have drawn strong connections on a household level between household composition and subsistence strategies (see Netting, Wilk, and Arnould 1984), Hirth’s model suggests that variability on this level is more likely because of changes in non-subsistence pursuits. These changes are in turn reflective of changes in intracommunity or intercommunity relationships, such as differential involvement in craft production and exchange.

The manner in which socioeconomic complexity emerges and the relative merits of these models of the domestic economy and incipient wealth differentiation are not the subject of this chapter. However, research on changes in the domestic economy related to incipient stratification in societies ranging in size from villages, such as Jachakala, to large urban populations presents opportunities to employ a wide range of methodological approaches. This chapter explores the relative merits of two of these analytical approaches to household data, namely comparisons of areas of a community (an interzonal analysis) and direct comparisons of household units. As the analytical summaries aim to demonstrate, patterns of differentiation among zones can reveal a level of socioeconomic organization between the household and community.

Figure 13.1. Map of the south central Andes showing the location of the La Joya research area of Bolivia (Bermann and Castillo 1993:312)

INTRODUCTION TO THE JACHAKALA PROJECT

Fieldwork conducted at the site of Jachakala documented domestic economy patterns at a community in the canton of La Joya, Department of Oruro, of the Bolivian altiplano (Figure 13.1). Changes through time in patterns of staple and craft production, wealth inequalities, supra-regional exchange, and material styles were explored extensively at this multicomponent site (Beaule 2002). In sum, the Jachakala project aimed to test whether variability in the economic organization of households underwrote the emergence of social, political, and wealth differences within the village.

The “household unit” (definition below) was the primary unit of excavation in the field, so direct interhousehold comparisons were made possible by combining artifacts from each domestic structure and its associated features. Nine household units were exposed in large, contiguous horizontal excavations averaging 24 square meters; a total of 340 square meters of surface area was opened up in this manner. Deep excavations of seventeen 2-by-2-meter units were taken down to sterile soil within each of those horizontal exposures and in a few randomly chosen locations.

The theoretical aim of the project was to test Hirth’s predictions that differential diversification of the domestic economy of some households (specifically to incorporate or focus more on craft production and exchange activities) would lead to early socioeconomic differentiation. Given that goal, comparisons of zones (areas of the site from which excavations to sterile soil extracted samples of domestic refuse that could be grouped in turn) provided a broader picture of the community’s long-term local history. These interzonal comparisons had the analytical advantage of negating the skewing of broader socioeconomic patterns provided by outlier households. Since Jachakala was physically divided into three zones of activity during the final period, these zones were simple to define and utilize. As shown on the site map (Figure 13.2), two large east-west-oriented adobe mudbrick walls physically (and emically) divide the community into three clusters. These physical boundaries were used to divide the site into three groups of units or areas of occupation: the southern, central, and northern zones. The results of my interhousehold and interzonal comparisons provide different views of the historical processes documented in the project. Each of these methodological approaches is described in turn, and their relative value in evaluating the Hirth model is discussed in the final section.

During the Niñalupita period (ca. AD 170–500), the site covered three to four hectares. A single calibrated radiocarbon date from a well-preserved hearth in the deepest cultural level dates the site’s initial occupation to 1720±60 years BP, or between AD 170 and 290. Patterns of artifacts recovered from the deepest strata (120 to 200 cm below the surface) of pits taken down to sterile soil reveal continuity in distributions of ceramic wares, black basalt lithic debitage, and camelid faunal remains, with very small quantities of semiprecious stones or other non-utilitarian imports. While no complete residential stone foundations were identified deeper than a meter below the surface, typical domestic features such as small unlined hearths, ash-filled pits, storage pits, and a small number of possible activity areas are scattered throughout the compact, silty clay matrix. This combination of artifacts and small features extends from the southernmost border of the site to the middle of the central zone.

The lack of Niñalupita phase house foundations poses little problem for this study of the domestic economy if one employs Kent Flannery’s (1976) theoretical approach to the “household unit.” This concept incorporates features such as middens, storage pits, and activity areas as well as residential structures associated with households (Bermann 1994:29). Material residues from the household unit’s range of domestic activities will be physically spread over a larger area than that enclosed by a foundation (see chapters by Ciolek-Torrello and Neff, this volume). This is a methodological dilemma acknowledged by archaeologists studying household processes and patterns, but one that is notoriously difficult, if not impossible, to address. In comparing remains from house floors and features immediately around the foundations, activity areas used by those residents are certainly missed. As Bermann (1994:30) argues for his Lukurmata household excavations, only part of each household unit is typically exposed (see also Bermann and Estevez Castillo 1993). The area actively used by modern Aymara peasants, he writes, covers more than 500 square meters on average when one includes outbuildings and associates activity areas along with the residential compound. Therefore, household excavations such as mine, which focus largely on the central domestic structure(s) and activity areas immediately adjacent to them, inevitably miss some proportion of the space actually employed from day to day by its residents. However, household units from Jachakala’s two later periods included large, spatially contiguous middens next to each structure. Within this and other related household features such as storage pits and hearths, we have a reasonable picture of the range of activities practiced by their residents, even if the activities, such as agriculture, associated with artifacts took place at some distance from the house itself. And those smaller domestic features—storage pits, hearths, fire pits, offering caches, and others—are well represented in the Niñalupita period excavations.

Figure 13.2. Map of the site of Jachakala, showing foundations of domestic and public structures and dividing walls; contour lines in the southwest corner mark ten-meter intervals in altitude

For the Niñalupita phase in particular, groups of domestic features associated, presumably, with the residents of various areas of the site (rather than particular domestic structures) must serve to represent household units in general. However, the majority of the artifacts analyzed to reconstruct the domestic economy from the Isahuara and Jachakala periods, Jachakala’s other two periods, came from the middens next to each house’s foundation. Middens are ideal contexts for reconstructing domestic processes because refuse from a range of domestic activities is more likely to be deposited there than on house floors.

Although, as mentioned above, no distinguishable household units could be identified, Niñalupita period artifact-distribution patterns, together with the small sample of recovered burials, indicate that the site was founded as a small, egalitarian village of some three to four hectares. These patterns reveal relatively little variation in the quantity or quality of materials to which groups of households in the southern and central zones had access, including both utilitarian as well as imported goods. However, early evidence for interzonal wealth differentiation dates to the Niñalupita occupation.

It is primarily in the centrally located strata (30 to 120 cm below the surface) of pits to sterile soil that the initial differences among households in different areas of the site are evident. The community began to slowly expand to the north, growing in size to cover approximately four to five hectares during this transitional period. The southern area of the site’s Isahuara period yielded the same approximate proportions of ceramic, lithic, and faunal remains, while domestic refuse from the central zone yielded slightly higher quantities of some categories of goods.

These interzonal differences (i.e., between samples of domestic refuse grouped into two areas of the site) during the Isahuara period (ca. AD 500–800) mark the expansion of wealth differentiation between areas of Jachakala. Only two identifiable household foundations were ever uncovered during this phase; site-size estimates are strictly based on the depth at which randomly placed pits throughout the site reached sterile soil. Evidence suggests the gradual development of differential access to both utilitarian (basalt bifacial hoes, camelid faunal remains) and luxury goods (imported semiprecious stones, marine shell fragments, some classes of ceramic wares) among residents of the southern and central areas of the site.

Structures from the Jachakala period date to approximately AD 800–1100, based on the recovery of Tiwanaku IV and V phase ceramic wares from within and around their foundations. By the end of this period, the site had grown to its final size of 6.72 hectares, again expanding primarily to the north. The dense scatter of remains visible on the surface of the site also dates to this occupation. These include the stone foundations of numerous architectural units, which are spatially grouped into three zones or sectors of the site. The two large dividing walls were probably constructed during this period. Many smaller, straight walls also appear within each of the zones, perhaps once serving as windbreaks, for privacy, or as spatial segregation of outdoor activities.

The southern and central zones both contain the remains of numerous household units that were the focus of extensive horizontal and deep stratigraphic excavations. Jachakala’s northern zone, on the other hand, yields evidence for numerous non-domestic functions introduced during the Jachakala period. As the smallest of the three zones, it covers less than two hectares, physically separated from the residential areas of the site by a large, double-rowed wall broken by two narrow doorways. Three small extramural depositories, both of very large circular temples, and a sizeable structure composed of two conjoined circular foundations (possibly a double llama corral) are located in this area of the community. Limited excavations in the cemetery located here revealed five male individuals; two female-infant pairs were buried elsewhere. No grave goods were found with any of the buried individuals. Also, a single circular household foundation was identified and excavated next to the large dividing wall but still within the northern zone.

This brief summary of Jachakala’s local history is intended to serve as a backdrop for the discussion of methodological approaches I used to explore interhousehold and interzonal patterns in greater depth. What is most important here is their relative utility with reference to how patterns of incipient wealth differentiation are reflected in comparisons of domestic architecture and artifact inventories.

ARCHITECTURAL DIMENSIONS OF RANKING

According to Richard Wilk, the house should be viewed not as an architectural expression of the classically cited public/private boundary but rather as an important means of manipulating boundaries between the domestic sphere and larger social groupings such as the clan or community (Wilk 1990:40). Thus, certain aspects of a society’s social and economic organization can be discerned from the intrasite spatial patterning of architectural remains (see Ciolek-Torrello, this volume). Uniformity of housing is enforced in ethnographic cases like the Kekchi Maya, especially within the context of differential participation in external economic pursuits such as cash crops. Such opportunities tend to lead to more open or accessible means of expressing newfound wealth in a public (i.e., architectural) fashion (Wilk 1990:38; see also Gonlin, this volume). Among comparable villages located in more remote settings where wage-earning opportunities fail to penetrate, egalitarian interhousehold relations are reinforced by communal landholding systems, rotating leadership, obligatory generosity, and various methods of social control such as envy and witchcraft accusations (Wilk 1990:38). Wilk also notes that housing tends to be dispersed and, for the most part, evenly spaced in agrarian subsistence economies (this distribution is also the case at Jachakala, where household units are fairly evenly spaced in the southern and central zones). Competition for the choicest land parcels and for access to pooled labor among extended families increases with the introduction of extraneous economic opportunities (1990:39). Wilk’s observations provide an architectural dimension to changes associated with incipient complexity and also a measure of resistance to economic differentiation in the form of housing homogeneity.

In addition to exploring architectural differentiation as a response to changing sociopolitical circumstances, domestic architecture can also be used to measure a community’s or region’s internal social organization. House size, location relative to some desirable point on the environmental or social landscape, and quality of construction are just three features of household architecture frequently cited as reflective of residents’ socioeconomic status. Among many others (Blanton 1994; Smith 1987), Robert Santley (1993:80) has argued that house size is a typical indicator of status differences, with elites in larger and better-made dwellings (see Douglass and Gonlin, this volume). Such architectural differentiation among residents of different sociopolitical statuses should, he writes, be more pronounced as “articulations between households become increasingly vertical” (Santley 1993:80).

These observations fit well with Hirth’s model in providing additional architectural correlates for both domestic and political economies. If houses differ in terms of size, quality of construction, associated features, and so forth, one can infer by analogy similar changes in the community’s social organization related to the composition of cooperative economic units. Since some degree of economic differentiation among households (or sampled areas of the site, in the Jachakala case) is apparent in all three chronological phases, models of changes in household organization can be tested for correlation between activity diversification and architectural differentiation.

HOUSEHOLD REMAINS AT JACHAKALA

“Household units” are defined as the typical architecture and features associated with individual coresidential groups. Coresidential groups include the group of individuals, not necessarily composing a nuclear or extended family, who resided together in a single structure or associated group of structures and who presumably made or used the features and artifacts associated with that residential structure(s). Each excavated household unit at Jachakala was uncovered with a set of 1-by-4-meter trenches judgmentally placed to expose full foundations as well as the features immediately surrounding them. After the first three houses were uncovered in this fashion, it became apparent that large middens were usually located immediately adjacent to the southwestern or southeastern portions of the house foundations. Consequently, trenches were placed to overlap larger exterior areas in these directions. When time permitted, the adjacent middens were sampled with single 2-by-2-meter units placed more or less in the center of these features, and taken down in ten-centimeter levels to sterile soil. All artifacts and small features recovered from this set of excavation units inside and next to the house foundations are grouped together as the remains of that household unit. A sample of nine household units (numbered sequentially by zone in Table 13.1) is included in this interhousehold analysis.

Figure 13.3. Illustration of House 4 (N511 E509), a fairly typical example of domestic architecture from Jachakala; the midden (N511 E509) lies to the southeast of the foundation

Variability in the size and general construction of domestic structures at Jachakala is fairly limited. Single or, occasionally, double rows of small fieldstones were placed in a roughly circular shape and topped by adobe walls. Sometimes the two courses of fieldstones were separated, as is visible in the southwestern corner of House 4 (illustrated in Figure 13.3) to form what may have been some sort of storage alcove. A few of the house foundations in the central zone of Jachakala also included a short, straight extension off of the southwest corner. This may indicate architectural mimicry of the northern dividing wall, which runs into one of the temple’s foundations.

Table 13.1. Categories of artifacts from craft production and interregional exchange activities for households 1 to 9, index of assemblage diversity analysis, and scores

As I briefly mentioned earlier, two of the household foundations uncovered during fieldwork are rectangular. One is located in the southern zone and the other one, which is technically square in shape, is in the central zone. Their interior areas (8.8 m² and 10.0 m²) are well within the range of those of the circular foundations (4.0 to 10.9 m²). There was no indication in the rectangular or square households’ artifact assemblages or activity areas as to why they were not constructed in the more common circular shape.

Interior features of households were usually limited to a small, unlined hearth excavated into the floor surface. This was sometimes accompanied by one or more small storage or refuse pits around the interior edge of the foundation. Exterior features included a large midden directly adjacent to the southwestern or southeastern corner of the house, as mentioned above. Also, a number of storage pits, with or without large jars embedded in them, might be dug into the ground surface surrounding a residence. Caches or household ritual features were usually made of some combination of the following: basalt tools, Tiwanaku-style vessel sherds, burned faunal and ceramic vessel fragments, and ash and carbon deposits. These offerings were often placed in small, undecorated bowls and placed in the bottom of a midden, the interior fire pit, or another conspicuous location next to the foundation.

Beyond this basic configuration of features, domestic architecture at Jachakala varies somewhat in terms of size, shape, and the presence or absence of specific features. For example, some but not all included storage pits adjacent to the foundation. Overall, house foundations uncovered at Jachakala include a somewhat wide range of estimated interior floor areas (from 4.0 to 10.9 m²), morphological characteristics such as single- and double-rowed circular foundations, rectangular and circular structures, interior and exterior features, and associated midden contents. Implications for the architectural expressions of rank differences, both within and between zones, can be briefly explored using these dimensions of variability. This is accomplished through comparing the individual household index of assemblage diversity scores, described in the last section of this chapter.

With only two exceptions, all excavated household remains date to the Jachakala period, based strictly on the depth of identified floors at ten to thirty centimeters below the surface. The two partial foundations excavated in Isahuara period levels in the southern zone lay almost on top of one another. Yet one of these two is clearly rectangular in shape, demonstrating that architectural variability in shape dates back at least that far. Though the majority of the nine illustrated house foundations were small and circular, they differed in terms of their interior floor area, number of rows of foundation stones, and location of features.

Two domestic structures from the southern zone of the site include the southern half of a single-row rectangular foundation (House 2). Since the northern half of the stones were removed or mixed in the fill of a large household midden that postdates this structure, the original interior square area is estimated at approximately twice the preserved length (2.2 m) and width (2.0 m). Underneath and slightly to the north of the midden overlapping the preserved half of this Isahuara period rectangular structure is another partial circular foundation of a single row of small stones (House 1).

House foundations uncovered in the central zone include one square and five circular structures. All of the houses at the site are freestanding structures, with several meters of space separating each house-midden combination from its neighbors. The single domestic structure identified in the northern zone is against the wall dividing the center from the north. This household’s lack of extraordinary features, interior area, or concentration of prestige goods makes it essentially indistinguishable from its central zone counterparts in all but location, just like the rectangular and square structures differ from the circular ones in little but shape.

We see then that variability in house foundation shape characterizes both of the site’s residential zones. The single domestic structure uncovered in the north is also not special apart from its location in the non-residential area.

ARCHITECTURAL VARIABILITY AND SOCIAL STRATIFICATION

The index of assemblage diversity (IAD) presents an ideal approach for direct household comparisons at Jachakala. By choosing lines of evidence that each represent a single aspect of the household’s range of activities, those household units that score highest are the ones with the most diverse domestic economies. This is an approach that incorporates testing the Hirth model and that can reveal patterns within and among zones. Consequently, the following analysis is intended to provide a more systematic way of comparing the diversity of the domestic economies of the sample of nine excavated household units. These results can then be compared to the range of architecture described above to test for correlations among the IAD scores and architectural features like size and shape.

Though archaeologists’ recommendations for using this statistic vary as widely as the range of formulas available in the literature, the one employed in this analysis has several advantages. Two dimensions of diversity relevant to this study include heterogeneity and richness. Heterogeneity depends on the frequencies of artifacts in the different categories. Richness is essentially the number of categories represented in a collection. Both are crucial aspects of diversity because an assemblage with more artifacts in a variety of categories is inherently more diverse than one with 100 artifacts of one type and one each of several other types. Many archaeological applications of diversity studies in fact ignore the heterogeneity component, thereby producing IAD scores that inaccurately reflect a collection’s diversity. Consequently, the formula known as Simpson’s Index (from Peet 1974) given below is used, since it addresses both dimensions of diversity.

Simpson’s Index (L):

where n_j = the number of items in category j, and N = the total number of artifacts in all categories

Categories included in the analysis were limited to those artifact types that represented craft production or exchange activities as a means to identify specific household units with evidence for increased diversification of the domestic economy as predicted by the Hirth model. Including ubiquitous basalt hoe production and consumption, camelid faunal, and utilitarian ceramic remains severely skewed initial calculations because of the extremely large numbers of artifacts in each category. These large quantities of basic goods downplay the heterogeneity of collections. Exclusion of ubiquitous subsistence remains also makes intuitive sense for the purposes of this investigation, since the expansion of the domestic economy and development of incipient socioeconomic complexity necessarily involves the incorporation of or emphasis on other kinds of activities. In other words, all households will continue their basic subsistence practices under political leadership regardless of their social strata. Therefore, wealth differences (and sometimes, the social rank differences associated with them) should be reflected in increasing interhousehold gaps in the range of activities represented in various domestic assemblages.

The categories employed here include obsidian and ópalo (a deep red volcanic glass from the mountains around La Paz) flakes and cores, projectile points (of basalt, obsidian, ópalo, quartzite, dacite, and so on), groundstone (grinding stones and manos), marine shell fragments, and weaving and spinning tools (ceramic spindle-whorl discs, bone awls, and needles). Two additional categories consist of decorated ceramic sherds (the vast majority of which are rim and body sherds from Tiwanaku-style ritual vessels) and unbaked ceramic figurine fragments. Also counted are bone scoops, antler digging implements, camelid-mandible tools, and incised or painted bone fragments (beads, flute or snuff tube fragments, and decorated pieces). Camelid-mandible tools are shaped bone tools of unknown function; these and the incised or painted bone fragments are found exclusively in the Tiwanaku state heartland and in Tiwanaku-contemporary strata of some south-central Andean sites.

Some of these artifact categories represent certain activities, such as the ritual vessels. Others are most likely markers of status; archaeologists often use differential distributions of long-distance imports to study social status. Projectile points made of obsidian are a good example of an item that might have increased the status of their owners at Jachakala and elsewhere. However, all artifact categories used in the IAD analysis reflect a kind of craft production or exchange activity that could have provided an opportunity for one or more households to accumulate wealth, social prestige, or both. Moreover, since all of these artifact types are found rather infrequently, the analysis should be unaffected by small sample sizes. The artifact counts in each of these eleven categories for the nine households are given in Table 13.1.

The IAD scores calculated for each household are provided in the last line of the table. Scores have been converted so that they range from zero (a collection with no diversity whatsoever) to one (an assemblage with the maximum degree of diversity, or equal numbers of artifacts in every category). As these numbers demonstrate, the household with the most diverse artifact collection is the central zone square structure (House 9), with a score of 0.7874. The second most diverse collection comes from the partial rectangular Isahuara period household excavated in the southern zone (House 2). The household units (foundation and associated midden) achieving the third through sixth highest IAD scores (Houses 4, 6, 5, and 8) all come from Jachakala Period central zone levels. The partial circular foundation underlying the southern zone rectangular foundation (House 1), the unique northern zone domestic structure (House 7), and the first (House 3) of the two houses atop slightly raised ashy surfaces achieved the lowest scores.

Tentatively speaking, emerging economic and social differences within the community seem to have an additional architectural dimension of standardized house shape. Though status differences were less well established than wealth differences at Jachakala (e.g., mortuary remains give no evidence for social ranking), the domestic remains at the site nevertheless present a cautionary tale to those assuming rigid links between ranking and architecture. Although differences of structural shape do seem to correspond with ranking, a quick perusal of the IAD scores, the range of interior areas of foundations, and the frequencies and relative ratios of most artifact categories reveal a continuum of variability rather than clusters of domestic groups (i.e., bimodal or trimodal distributions are expected when IAD scores are calculated and plotted for houses in very complex societies). These scores are graphed in Figure 13.4, which makes their continual distribution more visibly apparent. Had the relationship between status or wealth and architecture been more standardized or well established at Jachakala, one might expect subsistence differences between the zones, for instance, to be greater, with some household units able to maintain exclusive access to the most valuable packets of camelid meat or imported Tiwanaku-style pottery. Similarly, the largest houses (the two with the short extension walls like those off the corner of each temple) should have been those with the most diverse assemblages of non-subsistence artifact types if, in fact, those activities required (or justified) more internal space. Instead, household units with the three highest IAD scores are located in both zones, as Figure 13.5 shows.

The Jachakala case study does provide some support for the notion that shifts in the shape of domestic architecture are related to the richness and heterogeneity of households’ artifact assemblages. Yet these results should be considered in context, since other aspects of architectural expressions of ranking (size, quality of construction, and so forth) do not correspond well with the rectangular-circular dichotomy at the site.

Figure 13.4. Distribution of IAD scores

What we also see in this continuum of economic variability is that no one household in the central zone is much more diverse than its neighbors. Were one or a few household units significantly more diverse than the rest in their area, as in the case of an emerging political elite, we could expect that to be revealed in direct interhousehold comparisons such as this one. That simply is not what Figure 13.5 shows.

THE INTERZONAL APPROACH

If no one household in the central zone is much wealthier or more economically diverse than its neighbors, then the significance of the walls dividing the site into two residential and one non-residential area becomes even more difficult to discern. One possibility that emerged early in the research was to group household remains within each zone to compare them to each other as groups. This interzonal approach has some theoretical merit in the Andes, where the ethnohistoric record describes a level of social, political, and economic organization between the household and the community. The ayllu was like a large, extended kinship group, often divided into two complementary moieties. It might be a stretch to push this intermediate level of socioeconomic organization back 500 years or so before the Inka to Jachakala, but something between the household and community could explain why a small village built large walls between groups of its residents.

The objective of the interzonal approach to Jachakala’s assemblages then is to investigate domestic processes and patterns rather than to compare static architectural remains or individual household units. In particular, a comparison of areas of the community to one another will highlight spatial differences in the domestic economy that may be antecedent to later wealth differences, and that could reveal an intermediate level of socioeconomic organization. As Hayden and Cannon argue in their hallmark article on corporate groups, “while the error involved in interpreting most individual household assemblages is very large, it is greatly reduced when dealing with groups of households. Such grouping tends to average out the effects of specific historical and idiosyncratic factors acting on individual households” (1982:140, emphasis added).

Figure 13.5. Location of household units with the three highest IAD scores

Since household refuse was likely dumped nearer to residents’ homes than to their neighbors’ houses, we can compare assemblages from the two residential areas of Jachakala to test for spatial differences in the domestic economy of groups of households. In comparing these averaged central and southern zone collections for each of the three periods, I do not mean to imply that each zone was a definitive social (ayllu) or economic (corporate group) subgroup within the community. The materials from domestic features (though not houses per se) in the central and southern zones of Jachakala are compared by artifact class (lithics, faunal remains, ceramic wares, and so on) only to test for differences in the domestic economy of sections of the community. Note that the single identified household unit in the northern zone is excluded from this analysis because it is the only such domestic structure in that area of the community.

In fact, there are analytical advantages to comparing areas or zones of a community gained through this approach. House floor assemblages are more likely to consist of materials deliberately left during the structures’ abandonment; small, easily overlooked items; or ones accumulated during post-abandonment events. Assemblages from such contexts tell us little about the domestic activities and organization of a house’s original occupants and more about abandonment and post-abandonment processes. In contrast, exterior midden deposits reflect many years of steady deposits from a range of domestic activities, thereby mitigating some of the idiosyncrasies characteristic of floor assemblages. As such, middens are ideal windows on a household unit’s range of domestic activities, because the material remains of all activities that occurred around that structure have an equal chance of being deposited in that one location, which does not certainly mean that all activities are represented in middens. Interzonal comparisons offer an advantage then in allowing archaeologists to ignore the palimpsest nature of individual house floors. The most important assumption underlying this approach is that refuse produced by households in the southern (or central) zone is more likely to be dumped around their dwellings than it is to be disposed of near the houses of people living in another area of the site.

JACHAKALA’S ZONES AND WEALTH DIFFERENTIATION

Artifacts from the seven units selected for analysis were judgmentally chosen from the seventeen southern and central zone units excavated to sterile soil. Five of the seven were 2-by-2-meter pits judgmentally placed next to house foundations on the surface to locate deep middens. The remaining two units were 2-by-2-meter pits randomly located at the site and systematically excavated in ten-centimeter levels down to sterile soil (reached between 140 and 195 cm below the surface). Three of the seven are located in the central zone, three lie in the site’s southernmost zone, and a seventh unit, though technically on the border of these two zones, is classified as a central zone unit.

Instead of reciting the results of each interzonal analysis, I provide one particularly telling example here to illustrate some of the patterns revealed by this approach that were more indicative of changes in the site’s history than the interhousehold analysis. These examples concern the lithic, faunal, and ceramic remains that were deliberately left out of the index of assemblage diversity test, since they would have significantly reduced the collections’ heterogeneity if included. Both approaches were used to explore different classes of artifacts. However, this should not adversely affect their comparability, because it is their utility with regard to the correlates of the Hirth model, rather than the details of their applications, that concern here.

Faunal remains from the consumption of camelids were grouped into five meat units, called faunal “packets” below, including the cranium, vertebrae (“trunk”), forelimbs, hindlimbs, and ribs (Figure 13.6). These five packets have quite different amounts of meat attached to them in a typical adult llama. The individual elements and meat utilities assigned to each of the five are adopted from Mark Aldenderfer’s work (1998). The results of a chi-square test comparing the relative proportions of faunal remains assigned to each of the five meat packets in the Niñalupita period indicate that there is less than a 1 percent chance that differences derive from random variation (X² = 15.1307, df = 4, p < 0.01, Cramer’s V = 0.13). Faunal packet proportions in the Isahuara period are also highly significantly different (X² = 13.774, df = 1, p < 0.001, V = 0.10). Comparing Jachakala period faunal remains from the southern and central zones indicates that there is less than a 1 percent chance that differences between them derive from the vagaries of sampling (X² = 34.0111, df = 8, p f 0.001, Cramer’s V = 0.19). Furthermore, since the meat utility value of the trunk packet is so much greater than the other packets, I also conducted a chi-square test on the number of skeletal elements in the trunk packet and non-trunk packet totals (equaling the sum of the forelimb, hindlimb, ribs, and head packets) in the south and center. Results were similarly strong and highly significant in all three periods.

Figure 13.6. Niñalupita (top, left), Isahuara (top, right), and Jachakala (bottom) period faunal-packet proportions in the southern, central, and northern zones

This example of interzonal comparisons at Jachakala reveals that, overall, there was some meaningful variability in distributions of some of the meat units at the site. We can be fairly confident that residents in the south consumed proportionally more camelid forelimb packets than their neighbors. This might be expected in a domestic economy, under which a moderate degree of subsistence heterogeneity is expected within any community. However, the central zone’s greater access to parts of the trunk throughout the community’s history is more striking. Because the trunk packet has by far the highest meat utility value, this aspect of the interzonal dietary differences could indicate the center’s greater involvement in herding, or perhaps they had greater access to or first choice of the meat packets from the animals killed. Either way, some or all of the household units in the central zone from all three periods consumed significantly more of this most valuable meat packet. These faunal differences are the earliest and most consistent indicators of early wealth differentiation at Jachakala. Other highly significant and moderately strong interzonal differences include distributions of lithic debitage from the manufacture and use/refurbishing of basalt agricultural implements, a variety of bone tools, some classes of imported ceramic wares, and other exchange goods. However, none of these patterns were as consistent through time as the faunal packet differences. In fact, some reverse from period to period; for example, the lithic analyses reveal that residents of the southern zone were more involved in agriculture than their central zone neighbors, but that pattern reverses in the Isahuara period.

Figure 13.7. Proportions of trunk faunal packet elements from each household unit at Jachakala, at 80, 95, and 99 percent confidence levels, for the Niñalupita, Isahuara, and Jachakala periods

Here though we could again be looking at a case in which one or two household units disproportionally contributed to the central zone’s significantly higher proportions of camelid trunk packet meat. A glance at the bullet graph in Figure 13.7 shows that was not actually the case. For each of the three periods in Jachakala’s history, we see direct comparisons of each household unit’s trunk packet proportions; one household unit (N517 E450) has greater access than the others, but not at a significance level above 90 percent. Moreover, the three household units with the highest IAD scores were not those with the greatest access to trunk packet meat. Therefore, early wealth differences (as measured by access to better cuts of camelid meat) do not correspond to economic diversification (as measured in the IAD analysis).

It would appear that differences in pastoral subsistence practices date as far back as the Niñalupita period. I cannot know if there was just a single household in the center more involved in herding camelids, but the entire central zone had greater overall access to the best cut of meat. There should be no reason to suspect that the two areas would be different, since each assemblage comprises the pooled domestic remains of multiple households. Because of this distribution, the differential proportions of trunk packet elements in the two zones may reflect a more systematic kind of interhousehold difference than Hirth described (1993a, 1993b) for a domestic economy. As a group, the southern zone households were doing something different in their meat-consumption patterns than the central zone households, which exceeds the moderate variability expected in a domestic (subsistence-oriented) economy. This intriguing pattern, repeated again and again in other interzonal analyses (Beaule 2002), is evidence of an intermediate level of socioeconomic organization: perhaps not an ayllu with dual moieties per se, but something akin to it.

DISCUSSION

Several points relevant to testing the Hirth model emerge from the interzonal faunal comparisons at Jachakala. First, faunal differences demonstrate emergent wealth differences, which are more comprehensive than just unequal levels of participation in ritual activities, trade, or the production and consumption of luxury goods in that the household units’ access to subsistence resources varied by zone. (Other evidence suggests that the higher overall utility of the meat packets preferred in the central and northern zone structures is directly related to those residents’ increased participation in the community’s non-subsistence activities, such as the maintenance of trade caravans or ritual feasting.) However, camelid herds were one important source of wealth in the Andes, and so the center’s consistently greater access to the most valuable meat packets indicates its greater wealth. This is one case study in which early wealth differentiation is not related to differential control over or access to crafts or exchanged goods (significantly, differential distributions of both classes of artifacts emerge during the Isahuara and Jachakala periods) but rather faunal subsistence (wool, bone, hide, and sinews) resources and transportation. Consequently, Hirth’s prediction that the origins of an incipient political economy should relate to differential participation in craft production or exchange activities is not supported by Jachakala’s data. Rather, the origins of wealth differentiation in this case are best described via interzonal comparisons of faunal subsistence remains.

One question that arises is, what exactly are we comparing when we contrast areas or zones within a community? If similarities and differences among such units of comparison (zones, neighborhoods, household groups, or the like) are meaningful, does it necessarily follow that they had some function as economic or social units in our ancient case studies? The southern and central zone collections I created from the Jachakala data contained artifacts from groups of domestic features in two areas of the site. Those midden features were associated with structures on or stratigraphically close to the surface of the site, and so were part of Isahuara or Jachakala period household units. On the surface, the zones represent social divisions, which correspond with economic differences, since Jachakala’s residents divided themselves with walls. Below the surface, however, there is no evidence to suggest that these social divisions predate the construction of those dividing walls. The interzonal comparisons of Niñalupita and Isahuara period remains contrast the domestic economy of sections of the community that are averaged, in the sense that grouping households negates the effect of idiosyncratic outliers in the sample of household units or domestic features. By grouping all artifacts of each class recovered from a group of stratigraphic levels in pits whose placement was determined by later factors (including an association with a Jachakala period structure), each Isahuara period unit’s collection is akin to a random observation of remains deposited over several centuries in one 2-by-2-meter spot. In this way, grouping artifacts from the units in the south or center provides comparable samples of domestic refuse.

CONCLUSIONS

One can question why an interzonal approach to a community’s domestic economy might be advisable, especially if well-preserved household remains can be identified and explored. Indeed, interhousehold differences within zones are invisible in this approach, which is what prompted the interhousehold analysis in the first place. Of course, it is notoriously difficult to establish the contemporaneity of households in the absence of (and sometimes in spite of) written records, and so interhousehold comparisons must conceptually deal with the degree to which household studies can provide truly comparable observations of a society’s domestic activities. In projects that aim to reconstruct a community’s local diachronic history, rather than a synchronic comparison of particular house remains, the interzonal approach offers a number of analytical and interpretive advantages.

The assemblages of artifacts compiled within zones of the site include the grouped remains from several household units (floors, small features, and middens associated with domestic structures in the south and center, while the north includes artifacts from pits inside non-domestic structures). The excavation units I placed to uncover these household units were judgmentally chosen based on how well preserved they appeared to be on the surface of the site. Therefore, there is nothing is this sampling strategy that would produce a random, representative sample of Jachakala households within each zone or within the community as a whole.

Of course, what I really want to talk about in this study of the domestic economy are differences between households, regardless of whether the residents of individual structures or groups of people in spatially proximal structures were the basal socioeconomic units in Jachakalan society. The index of assemblage diversity analysis showed a continuum of scores that suggests a similar continuum in the degree to which sampled household units diversified the range of activities represented within each group of features. (In fact, the results of interhousehold faunal analyses, not reported here, show that the three household units with the highest IAD scores were not those with the greatest access to the trunk meat packet.) This interhousehold analysis suggests two things to me: first, that no single household unit in either zone was much of an outlier in terms of its diversity of activities, and second, that the interzonal analysis is a better way to paint in broad strokes community-level patterns in activities performed by and within households. Perhaps the interzonal approach fails to acknowledge that, ultimately, members of individual households were probably the ones deciding when to acquire new goods, make different tools than their neighbors, devote more or less time and energy to some tasks, consume or store different kinds of food, and so on. Zones were not necessarily units of social divisions or economic cooperation, even during the Jachakala period, but comparing them does more clearly reveal differences between their respective ranges of domestic activities. And these differences certainly could indicate a level of social and economic organization between the household unit and the community. This possibility is the best argument for comparing areas of a community to each other to test for such social divisions elsewhere. Areas of a village or town that lie on either side of a river or neighborhoods within a city are examples that could be tested for interzonal patterns in many activities.

The interzonal approach accomplishes both analytical aims: it allows one to describe broad synchronic and diachronic patterns in Jachakala’s domestic assemblages, and it provides an indirect means of comparing groups of households. In these ways, the interzonal method is better suited to study changes in the domestic economy of a community. Household remains (i.e., individual structures) are necessary, however, to investigate questions about how households are organized and articulate with changing regional relationships, gendered divisions of labor, and other research objectives concerned with basic coresidential units. Even when such remains are available for study, a project’s ultimate objectives (rather than the presence of particular types of features) should determine whether individual household units are the basis for comparisons. In the case of the Jachakala project, the results of both kinds of analytical strategies demonstrate the utility of comparisons among zones as an important complement to more traditional interhousehold analyses.

Acknowledgments. Thanks go to Marc Bermann, Robert Drennan, and Jim Richardson for their crucial roles in shaping and guiding the dissertation on which this chapter is based. Christian Peterson, the editors of this volume, and two peer reviewers also provided insightful comments. Funding for the fieldwork came from the University of Pittsburgh’s Center for Latin American Studies, the Fulbright IIE program, and a National Science Foundation Doctoral Dissertation Improvement Grant. As always, errors are my own.

REFERENCES CITED