Archaeologists have long been interested in the conditions that gave rise to permanent forms of social inequality. For decades, the focus of attention has been on institutions that promoted inequality in ancient state-level societies (Adams Reference Adams1966; Childe Reference Childe1951; Johnson and Earle Reference Johnson and Earle1987). The scope of research has expanded in recent years to include a consideration of institutionalized inequality in pre-state, middle-range societies (Ames Reference Ames, Bentley, Maschner and Chippindale2007; Hayden Reference Hayden, Price and Feinman1995; McIntosh Reference McIntosh and McIntosh1999; Price and Feinman, ed. Reference Price and Feinman1995, Reference Price and Feinman2010). In such cases, the seeds of inequality are often sown among households in the form of differential access to domestic labor (e.g., McIntosh Reference McIntosh and McIntosh1999:6–7; Netting Reference Netting1982; Wilk Reference Wilk1991:196–197). Households also play an important role in the transmission of property and wealth across generations (Blanton Reference Blanton, Price and Feinman1995; Goody Reference Goody1990). Consequently, they have become key analytical units in efforts to understand the development of social differentiation (Douglass and Gonlin Reference Douglass, Gonlin, Douglass and Gonlin2012:8, 13–15).
We examine the evidence for social dissimilarities at the household level among the Hohokam of southern Arizona. They built the largest irrigation works in prehispanic North America. Traditional archaeological narratives viewed Hohokam social organization as fundamentally egalitarian, a “benign primitive democracy,” in the words of Emil Haury (Reference Haury1976:353). The demands of survival in a harsh desert environment supposedly dictated that everyone who lived along the same canal worked together and shared rights to the means of production. More recent narratives recognize nonegalitarian relationships, but researchers disagree about the nature of those ties, as well as the underlying social dynamics (Abbott Reference Abbott and Abbott2003; Doyel Reference Doyel1974; Ensor Reference Ensor2013; Gregory Reference Gregory, Crown and Judge1991; McGuire Reference McGuire1992; Wilcox Reference Wilcox, Crown and Judge1991, Reference Wilcox and Neitzel1999).
We direct our attention to the physical domain of Hohokam households—their architecture. In a previous article (Abbott et al. Reference Abbott, Craig, Zanotto, Judd and Kober2019), we provided a method for reconstructing the labor costs for house construction, taking into account different architectural styles from different time periods. We examine the labor costs involved in building houses at Pueblo Grande, one of the most densely populated and politically prominent villages in the Hohokam region. We focus our analysis on testing several expectations related to wealth, status, and domestic architecture; in particular, the degree to which social inequality was expressed in variable house construction costs. In addition, we investigate the degree to which inequality, as reflected in domestic architecture, persisted across generations. Even relatively egalitarian societies can have pronounced wealth differences on a short-term or cyclical basis (e.g., Cancian Reference Cancian1976; Wolf Reference Wolf1966). What precludes them from becoming stratified are the social mechanisms (e.g., egalitarian ethos, marriage and inheritance practices) to prevent the continuance and growth of social inequality over the long term (McGuire and Netting Reference McGuire and Netting1982; Netting Reference Netting1993:200–202; Wilcox and Sternberg Reference Wilcox and Sternberg1983).
Hohokam domestic architecture has the added advantage of being easily linked to household-level social groups, because dwellings for most time periods were stand-alone structures. Also, the high degree of residential stability seen at many sites, combined with their long occupation spans (300+ years), makes it possible to track the changing fortunes of individual households over time. This individual-based, long-term perspective is rare in studies of social differentiation in premodern, middle-range societies (Bowles et al. Reference Bowles, Smith and Mulder2010:8; Drennan et al. Reference Drennan, Peterson, Fox, Price and Feinman2010:73). As such, it provides a useful frame of reference for testing different evolutionary models.
The purpose of our investigations has been threefold. First, in our previous publication (Abbott et al. Reference Abbott, Craig, Zanotto, Judd and Kober2019), we developed and implemented a methodology to calculate construction costs of various styles of Hohokam architecture, which we hope will be widely applied for other Hohokam projects and beyond. To facilitate future applications, we described in detail the variables used, the rationale and justification for their measurement, the way the measurements were made, and the way the construction costs were derived. Second, we calculated and analyzed the costs of building Hohokam houses to determine if there were architectural expressions of household inequality at Pueblo Grande and what they might mean for Hohokam society. Third, in agricultural societies—and especially in those employing technologies that supplement human labor (e.g., irrigation)—high levels of inequality can emerge as productive land becomes limited due to rises in population or changes in weather patterns (Flannery and Marcus Reference Flannery and Marcus2012; Goody Reference Goody1976:97; Kohler and Higgins Reference Kohler and Higgins2016:691; Kohler et al. Reference Kohler, Smith, Bogaard, Feinman, Peterson, Betzenhauser, Pailes, Stone, Prentiss, Dennehy, Ellyson, Nicholas, Faulseit, Styring, Whitlam, Fochesato, Foor and Bowles2017; Smith et al. Reference Smith, Kohler, Feinman, Kohler and Smith2018:10). In a Hohokam context, we examine this oft-cited model about the relationship between population increases, resource shortages, and wealth differentials. We expect to encounter marked differentiation among the prehistoric residents in the irrigated valleys of the Phoenix Basin during the Late Sedentary and Classic periods, when demographic pressures arose from rapid and large-scale episodes of immigration.
Hohokam Culture
The Hohokam people are best known for their continuous, 1,000-year occupation of the Phoenix Basin (AD 450–1450; Figure 1; Table 1Footnote 1). Their villages were anchored to the landscape by huge investments in irrigation infrastructure. Multivillage irrigation cooperatives, called “canal systems,” brought water from the lower Salt and middle Gila Rivers up onto the river terraces. The largest of the farmer collectives is referred to as “Canal System 2,” which encompassed 20 settlements and nine main canals that irrigated more than 6,070 ha (15,000 acres) of corn, beans, squash, and cotton on the north bank of the Salt River (Howard Reference Howard, Howard and Huckleberry1991:5–15).
Figure 1. The Hohokam culture area.
Table 1. Hohokam Temporal Intervals.
a Controversial dating (see Note 1).
The Phoenix Basin settlements had a typical spatial layout and shared a long-term developmental sequence. The earliest habitations were arranged around a central plaza with public facilities, such as cemeteries and communal cooking areas, situated near the plaza margins (Wallace Reference Wallace, Fish and Fish2007; Wilcox et al. Reference Wilcox, McGuire and Sternberg1981). Starting around AD 800, large earthen courts for the playing of a ritual ball game were built adjacent to the central plaza. A regional network of interaction and conjoint beliefs was apparent by the distribution of 190 ball-court sites over an area the size of South Carolina (Marshall Reference Marshall and Craig2001). The regular gatherings of people from many places and ecological settings made the ball games and possibly associated marketplaces conduits for large volumes of exchange goods (Abbott Reference Abbott, Garraty and Stark2010; Abbott et al. Reference Abbott, Smith and Gallaga2007; Doyel Reference Doyel, Crown and James Judge1991; Wilcox and Sternberg Reference Wilcox and Sternberg1983).
The ball-court complex ended abruptly around AD 1070, and demographic instability followed. Migrants from outside the irrigated lowlands streamed into the riverine villages, doubling the population at Pueblo Grande (Abbott and Foster Reference Abbott, Foster and Abbott2003) and probably at many other sites (Doelle Reference Doelle, Elson, Stark and Gregory1995). The influx likely stimulated the construction of platform mounds and walled central precincts on the edge of the central plaza where community rituals were staged or where elevated residences were built for an elite class (Downum and Bostwick Reference Downum, Bostwick and Abbott2003; Doyel Reference Doyel1981; Gregory and Nials Reference Gregory, Nials, Dittert and Dove1985; Wilcox Reference Wilcox1987).
By the beginning of the fourteenth century, Hohokam residents throughout the Phoenix Basin and beyond enclosed their dwellings behind massive compound walls, marking a fundamental social change (Haury Reference Haury and Noble1991:70). Domestic life was transformed from a relatively inclusive and equitable existence to one distinguished by greater exclusivity and inequality (McGuire Reference McGuire1992:204–207; Wilcox Reference Wilcox, Crown and Judge1991:267–269). But by AD 1375, the compound walls were reduced to rubble, and the domestic architecture reverted to earlier styles. Population sizes along the Salt River plummeted, and at Pueblo Grande and elsewhere, a few scattered residence groups lingered until the final exodus from the valley around AD 1450. Many people probably moved to the middle Gila River valley, where their descendants still proudly reside today.
Hohokam Households
Households are task-oriented, co-residential family units that structure daily life in human societies throughout the world (Blanton Reference Blanton, Price and Feinman1995:108; Netting et al. Reference Netting, Wilk, Arnold, Netting, Wilk and Arnold1984:xx). They are usually defined by the range of activities they perform, particularly those related to subsistence and social reproduction (Douglass and Gonlin Reference Douglass, Gonlin, Douglass and Gonlin2012:2–5; Wilk and Rathje Reference Wilk and Rathje1982:618). Hohokam households of the Sedentary and Early Classic periods were manifest by a cluster of contemporaneous structures arranged around a common courtyard space, which equated with extended-family or multifamily units. During the Late Classic period, residence groups composed of one to several households were enclosed in walled residential compounds (Craig and Henderson Reference Craig, Henderson, Fish and Fish2007). Hohokam households did most of the work associated with building and maintaining the canals (Howard Reference Howard2014). They also provided most of the labor involved in preparing fields and growing and harvesting crops (Henderson and Clark Reference Henderson, Clark and Henderson2004).
Given the size of the canal systems, there can be little doubt that irrigation water was a common-property resource managed at the community or district level (Hunt et al. Reference Hunt, Guillet, Abbott, Bayman, Fish, Fish, Kintigh and Neely2005). In contrast, rights approximating those of private ownership likely developed to manage plots of irrigable land (Mabry Reference Mabry and Mabry1996:16–17; Netting Reference Netting1993:158). The long-term and stable usufruct also reflects a commitment to maintaining the household's corporate holdings over time. As a result, some households may have become dependent on others for access to basic resources such as water and land. Similarly, although ceramics and other craft items were probably produced by part-time specialists working at the household level, most households were dependent on multiple and often distant producers for everyday items, such as cooking and storage pots (Abbott Reference Abbott2009; Abbott et al. Reference Abbott, Smith and Gallaga2007). These relationships suggest that dependencies were a key component of the Hohokam political economy, possibly fueling the emergence and persistence of social inequality.
Pueblo Grande
The Pueblo Grande residents oversaw the intake of river water into Canal System 2, thereby placing them in a position of political prominence and economic security. Not surprisingly, the village grew to be the largest in the lower Salt River Valley, supporting a population that exceeded 1,000 citizens in the latter part of the Classic period (Abbott and Foster Reference Abbott, Foster and Abbott2003). But, prior to the multivillage irrigation cooperatives, the early settlement was simply a light scattering of small residence groups (Bostwick Reference Bostwick, Bostwick and Downum1994; Ensor Reference Ensor2013:174–176; Foster Reference Foster1995).
Around AD 1070, when the ball courts were abandoned and populations in outlying areas were on the move, habitation levels surged at Pueblo Grande with the first of three waves of migration. These waves were well documented in the eastern third of the ancient site with excavations by Soil Systems Inc. (SSI) prior to highway construction (Mitchell Reference Mitchell and Mitchell1994; Figure 2). Each wave pushed the site boundaries outward and successively farther from the central precinct of the village.
Figure 2. The SSI project area.
The SSI fieldwork exposed 14 habitation areas (HAs), each one measuring 1,200–3,000 m2, demarcating a discrete cluster of structures, extramural spaces, and a cemetery. Undoubtedly, each HA was home to a self-recognized residence group, which was typically composed of multiple households and which often persisted for multiple generations. HA-5, 6, 8, and 9 were all established when newcomers arrived during the time when the ball-court network dissolved. They built pithouses of a style defined by Haury (Reference Haury1976:53–57), called “Type S-1.”
A second wave around AD 1100 doubled the size of Pueblo Grande virtually overnight (Abbott and Foster Reference Abbott, Foster and Abbott2003). Six new HAs (HA-1, 2, 3, 7, 10, 12) were established in the eastern part of the settlement. The migrants built five distinct house styles, including Post-Reinforced surface structures and four categories of pithouses (Type S-2; Rock-Lined; Deep, Adobe-Lined; and Deep, Post-Supported). Presumably, the variability corresponded to a multiplicity of homelands from which the new arrivals originated. Just a few decades later, the wide diversity of house types gave way to architectural conformity as a single new style, Narrow-Walled Adobe pithouses, was constructed across the community. Also at this time, two platform mounds were built side by side at the village center, and one new residence group (HA-11) was founded in the SSI project area.
During the final migration between AD 1275 and 1375, the entire community adopted a new architectural form. Clusters of Massive-Walled Adobe surface structures were enclosed behind towering compound walls. Many of the HAs grew significantly, as newcomers probably exercised kinship options to crowd into long-occupied habitation areas. Three new residence groups—HA-4, 13, and 14—also took root at the northeastern edge of the site. In addition, the Pueblo Grande citizens joined their side-by-side platform mounds into a single immense edifice. Its 4 m high elevated surface became crammed with an elaborate complex of open-air courts and Massive-Walled Adobe structures accented with many unusual architectural details (e.g., corner entryways, lofty ceilings, cylindrical pillars, altars, a multicolored floor, a solstice marker).
The last phase of occupation began at about AD 1375. In the habitation areas, the Massive-Walled Adobe structures and the towering compound walls surrounding them fell into ruin, but in five of the SSI HAs (HA-2, 3, 4, 5, 7) the occupants persisted whereas the other residence areas were vacated. The later inhabitants built Narrow-Walled Adobe and Deep, Adobe-Lined pithouses, sometimes through the foundations of fallen compound walls (Abbott et al. Reference Abbott, Mitchell, Merewether and Mitchell1994). By the mid-fifteenth century and after 800 years of continuous occupation, Pueblo Grande was totally uninhabited.
Why Pueblo Grande?
If there was social inequality in Hohokam society expressed in variable construction costs for domestic architecture, then there are reasons to believe we should find evidence for it at Pueblo Grande. Platform mounds were built in the thirteenth and fourteenth centuries at Pueblo Grande and elsewhere, probably reflecting increasing social differentiation and political centralization across the Hohokam world at those times (Bayman Reference Bayman2001; McGuire Reference McGuire1992). As described above, multiple waves of immigrants impacted Pueblo Grande during the Late Sedentary and Classic periods, possibly limiting the availability of irrigable land and thereby promoting economic disparities. Under contested circumstances, the mounds probably marked social boundaries and reaffirmed property rights (Abbott Reference Abbott2000:204–206; Bayman and Sullivan Reference Bayman and Sullivan2008; Elson Reference Elson1998; Fish and Fish Reference Fish, Fish, Wills and Leonard1994, Reference Fish, Fish and Mills2000). On a political level, the people who lived in the affiliated households on or near the elevated platforms are the most obvious candidates for elite groups in Hohokam society. They likely controlled access to the mound precincts and were the primary beneficiaries of mound-related activities (Craig Reference Craig2010; Ensor Reference Ensor2013).
From an archaeological perspective, the conditions at Pueblo Grande were excellent. The SSI excavations in the eastern third of the village unearthed 14 discrete and well-defined habitation areas, most of which were associated with the intergenerational transfer of property, and, likely, the long-term maintenance of an estate (Craig Reference Craig and Beck2007, Reference Craig2010; Craig and Henderson Reference Craig, Henderson, Fish and Fish2007). Presumably, those residence groups with enduring property represented the social circumstances and sufficient time for the seeds of permanent social inequality and wealth to germinate among the households. Also, it is likely that the newcomer arrivals at different times differentially favored residence groups with unequal degrees of longevity (Plog and Heitman Reference Plog and Heitman2010:19624). At Pueblo Grande, the conditions were ripe for wealth gaps and social inequality. Hundreds of years of occupation, 14 distinct habitation groups, and a sample of approximately 350 structures built with multiple styles of construction make Pueblo Grande an ideal context for the present research.
Grewe's Role
Many of the methods utilized in this investigation were initially developed as part of a study of domestic architecture at the Grewe site, a large Hohokam village located along the middle Gila River (see Figure 1). Excavations by Northland Research Inc. uncovered a large residential district in the heart of the village with more than 250 pithouses (Craig Reference Craig and Craig2001a). A ceremonial precinct with a ball court and a communal cooking area used for ritual feasting was found directly adjacent to the residential district. Ceramic and chronometric evidence indicated that the site was occupied continuously from about AD 450 to 1100 (Henderson Reference Henderson and Craig2001). Throughout much of its occupation, Grewe was one of the largest Hohokam villages in the region, with an estimated peak population of between 800 and 1,000 residents (Craig Reference Craig and Craig2001b). Following the collapse of the ball-court network in the late eleventh century, the site was abandoned and a large segment of the population is believed to have moved a short distance away (<1 km) and established residence at Casa Grande (Craig Reference Craig and Beck2007).
Analysis of the Grewe house data focused on a sample of 132 well-preserved pithouses, all made of wattle and daub, similar to the S-1 style pithouses at Pueblo Grande. A wide variety of internal features was commonly associated with habitation structures, including hearths, floor pits, plastered floors, and benches or raised platforms (Craig Reference Craig and Craig2001a:93). This variability in domestic architecture formed the basis for an examination of potential wealth differences among households at Grewe (Craig Reference Craig and Craig2001c). Labor costs were calculated for individual houses based on a combination of field data and published experimental data. Twenty house clusters were each arranged around a courtyard that served as shared domestic space for household activity (Figure 3). The smallest courtyard group (i.e., household) contained two pithouses and covered a total area of about 100 m2; the largest contained 26 pithouses built over multiple centuries and covered a total area of more than 600 m2.
Figure 3. Courtyard groups at Grewe.
Some courtyard groups were occupied for only one or two generations, whereas others were occupied for centuries. Their longevity implies a lasting recognition of place and the emergence of household property rights transferred across generations (Craig Reference Craig and Beck2007, Reference Craig2010). Household members presumably pooled labor, shared resources, and acted as a unified body in making decisions about food production and land tenure. Presumably they were also committed to maintaining their corporate holdings over time (Hayden and Cannon Reference Hayden and Cannon1982).
The labor costs associated with building a typical structure at Grewe averaged 200–250 person-hours, with an overall range of 140–385 person-hours (Craig Reference Craig and Craig2001c). Although the most labor-intensive house also happened to be the largest (21.9 m2), it was also quite elaborate, with a 2 cm thick plastered floor and a raised wood-and-brush platform that covered the entire structure.
The two Grewe courtyard groups with the largest area (>600 m2), most houses (>20), and the longest-lived occupations (>200 years) contained the most labor-intensive houses (see Craig Reference Craig and Craig2001a). One contained five of the 10 most labor-intensive houses, and the other contained houses that were, on average, 30%–50% more labor intensive than those in other courtyard groups. Their proximity to the communal cooking area and ball court led Craig (Reference Craig and Beck2007:458, Reference Craig2010:81) to propose that wealthy households at Grewe may have sponsored feasts associated with ball court–related events.
Justification for Studying Domestic Architecture
Household inequality is generally measured in terms of differential access to valued goods, services, and status (Blanton Reference Blanton, Price and Feinman1995:108–109; Netting Reference Netting1993:189; Wilk Reference Wilk1991:197–203). The methods used by archaeologists to measure inequality are diverse, and they reflect the importance of both material and immaterial forms of wealth (Price and Feinman, ed. Reference Price and Feinman1995, Reference Price and Feinman2010). There is increasing recognition that inequality can also vary by degree and kind (Ames Reference Ames, Bentley, Maschner and Chippindale2007; Drennan et al. Reference Drennan, Peterson, Fox, Price and Feinman2010). Consequently, it can take a variety of forms and may not be apparent in all classes of archaeological evidence (Price and Feinman Reference Price, Feinman, Price and Feinman2010:6). Burial data, for example, might indicate a slight degree of differentiation, whereas domestic architecture or household artifact assemblages might indicate a much higher level (Drennan et al. Reference Drennan, Peterson, Fox, Price and Feinman2010:46–64). The extent to which wealth differences were transmitted across generations is another factor that needs to be considered given that intergenerational wealth transmission is one of the defining characteristics of persistent social inequality (Borgerhoff Mulder et al. Reference Borgerhoff Mulder, Bowels, Hertz, Bell, Beise, Clark, Fazzio, Gurven, Hill, Hooper, Irons, Kaplan, Leonetti, Low, Marlowe, McElreath, Naidu, Nolan, Piraino, Quinlan, Schniter, Sear, Shenk, Smith, von Rueden and Wiesner2009; Bowles et al. Reference Bowles, Smith and Mulder2010; Shenk et al. Reference Shenk, Mulder, Beise, Clark, Irons, Leonetti, Low, Bowles, Hertz, Bell and Piraino2010).
Our approach focuses on the variability in construction costs of domestic architecture to study status and wealth differentials across Hohokam households. In many societies around the world, households of higher status and/or more wealth live in homes that are bigger, more extravagant, and better made than those composed of poorer and low-status members (Abrams Reference Abrams and Schiffer1989; Ames Reference Ames, Bentley, Maschner and Chippindale2007; Kramer Reference Kramer and Kramer1979). As Feinman and Neitzel (Reference Feinman, Neitzel and Schiffer1984:57–59) found among 51 pre-state sedentary societies, large and ornate domestic architecture was often a marker of elite status. In contrast, less complex societies tend to build homes of similar form, composition, and energy expenditures per capita (Abrams Reference Abrams and Schiffer1989:54; Rapoport Reference Rapoport1969; Wilk Reference Wilk and Kent1990). In our study, we measure and compare construction costs to determine their variability across households.
Concerning the cost of architectural construction, there are three complementary aspects to be kept in mind: dwelling size, construction materials, and extravagance. Generally, and at Pueblo Grande specifically, house size (i.e., roofed floor area) was a major contributor to the building costs of each structure. When all else is held constant, larger structures obviously require more building materials and labor inputs than smaller houses. It is also true that structures with the same roofed floor area need not be equal in construction costs. As described below, Hohokam houses were built with multiple styles distinguished in large part by the variation in construction materials, which incurred different expenses and probably withstood the elements variably. Variability in construction costs also entailed “luxurious” extras, such as elevated floors, benches, raised platforms for sleeping, and formally designated storage space. Some Hohokam homes were better built with more expensive and longer-lasting building materials and/or appointed with more lavish amenities. This variability in quality may also be emblematic of wealth and status differentials, and it can be captured by a calculation of relative cost (cost per m2 of floor space).
Dwelling size can be related to status and wealth inequality because successful households often attract junior members, making the households of the wealthy larger than those of the poor (Wilk Reference Wilk1983:111). In turn, the labor of subordinates can contribute to the accumulation of wealth (Netting Reference Netting1982; Wilk Reference Wilk1983). The larger residence memberships of high-status households consequently require more residential space, usually in the form of larger or more numerous houses.
The association between status and house size, however, is not a simple one. Cross-culturally, house size can correspond with multiple factors, not just status and wealth (Ames Reference Ames, Bentley, Maschner and Chippindale2007:497, 501; Blanton Reference Blanton1994; Feinman and Neitzel Reference Feinman, Neitzel and Schiffer1984; Smith et al. Reference Smith, Kohler, Feinman, Kohler and Smith2018:23). For instance, the size of a dwelling in many societies covaries with the number of inhabitants who resided in it (Cook Reference Cook1972:16; Hassan Reference Hassan1981:63–77; Narroll Reference Narroll1962; Wilk Reference Wilk1983), which is an assumption commonly made by Hohokam scholars for deriving population estimates (Abbott and Foster Reference Abbott, Foster and Abbott2003; Wilcox et al. Reference Wilcox, McGuire and Sternberg1981) and determining postmarital residence (Ensor Reference Ensor2013).
House size can reflect the amount of labor devoted to construction as well as how much help the builder received from kin and other social connections (Kramer Reference Kramer and Kramer1979; Wilk Reference Wilk1983). In sedentary societies, the number and intensity of these ties tend to increase over time, thereby favoring members of the earliest-established households. A second expectation at the village scale corresponds to the order in which different residence groups were established. At Pueblo Grande for instance, the village expanded outward over time, with the longest-lived habitation areas near the village center and the youngest residence groups on the margins of the settlement. When large structures occurred, we would expect them among the longest-lived habitation areas near the center of the site.
If house size can be related to multiple, potentially interrelated factors—including the number of residents, the length of the occupation, as well as status and wealth—then house size alone is an insufficient factor for gauging the inequality in Hohokam society. It is for this reason that we consider construction style and elaborateness as well as dwelling size to be essential for this study.
Architectural Styles at Pueblo Grande
Pueblo Grande contained domestic structures that varied in size, shape, and method of construction, which led the SSI excavators to define and describe eight architectural styles at the site (Mitchell Reference Mitchell and Mitchell1994). The construction of each style was made unique by the combination of (1) the presence/absence and different forms of house pits and (2) the various means by which the builders made use of locally procured building materials (wood, brush, adobe, and rock) to form the walls and roof. Differences in construction among the various architectural styles at Pueblo Grande are apparent in Figure 4, and more detailed descriptions can be found in our previous publication (Abbott et al. Reference Abbott, Craig, Zanotto, Judd and Kober2019:322–327).
Figure 4. Architectural styles at Pueblo Grande.
Mitchell (Reference Mitchell and Mitchell1994:33) reports that SSI identified about 350 architectural features in the Hohokam Expressway corridor at Pueblo Grande. Of that total, 143 houses were sufficiently preserved to assign them to an architectural style and to calculate their construction cost. Severe damage caused by rebuilding within the long-occupied habitation areas and modern disturbance disqualified the others from further consideration. Our sample included every structure that could be assigned to an architectural type and for which its construction cost could be calculated.
Expectations
The results at Grewe illuminated clear evidence for social differentiation among the village households, manifested in the variability of construction costs for domestic architecture. The wealth parameters of entire courtyard groups were estimated by combining the construction costs for all contemporaneous pithouses in each courtyard group. The assumption was that the material wealth of each household was best reflected in its entire architectural portfolio (Craig Reference Craig2010:78). Using Grewe as a guide, we expected similar or even more pronounced differences at Pueblo Grande due to the increasing social and political complexity in Hohokam society during the Late Sedentary and Classic periods and because Pueblo Grande probably was—both literally and figuratively—at the center of those developments. At Pueblo Grande, however, treating architectural costs for entire residence groups as units of analysis was not feasible due largely to problems with preservation. It would be impossible to calculate the construction costs of all contemporaneous structures within each HA. Within the confines of these long-lived residential units, later building episodes impacted the material remains of earlier structures, and modern disturbances damaged both earlier and later features as well. A strict application of the Grewe analytical strategy made little sense for Pueblo Grande, but the Grewe findings did lead to useful expectations for the present study.
In addition to the results at the scale of courtyard groups, Craig's (Reference Craig2004) analysis at Grewe revealed patterning among individual houses. The Grewe findings indicated that the households (courtyard groups) with the highest total construction costs were also the households that typically included the largest and most elaborate dwellings in the settlement. The Grewe study implied that tracking wealth differentials was possible at the scale of individual structures, which fit with the strengths of the Pueblo Grande database.
If wealth differentials were expressed at Pueblo Grande with the construction of large (size) and extravagant (relative cost) houses, then several findings would be expected. First, we would anticipate notably large and extravagantly built dwellings to be associated with some but not all HAs, and we would expect these relationships to have persisted over time. Second, well-established households with a history of perseverance and accumulated social and economic advantages are expected to have been wealthier than relative newcomers. Consequently, because occupation length at Pueblo Grande was related to the distance from the central precinct of the village, we would expect the HAs with the largest and most elaborate houses to be closest to the platform mound (see also Wilcox Reference Wilcox, Crown and Judge1991:268). Finally, because Pueblo Grande was the most prominent village inhabited during a time of likely increased social and political differentiation, we expect the large houses to be bigger and the elaborate structures to be more extravagant at Pueblo Grande than those at Grewe.
Methods
Our work is an outgrowth of architectural energetics, which involves quantitative labor-cost reconstructions. As an assemblage of various raw materials and manufactured components, architecture has a composite cost of procuring, processing, transporting, and assembling raw materials into the finished product. Architectural energetics calculates the construction expenses into a common unit of measurement for analytical comparison—in other words, a single analytical attribute in the form of labor-time expenditures.
Architectural energetics avoids a subjective assessment of architectural outlays in comparative research by relying on three lines of evidence accessible to the archaeologist. First, the kinds and amounts of raw materials used in house construction are determined based on excavation data and field maps. Second, the key tasks associated with obtaining and assembling the building materials are identified based on ethnographic data. Third, the labor requirements associated with the various construction tasks are derived from replicative experiments (Craig Reference Craig and Craig2001c:116; Erasmus Reference Erasmus1965). We adopt this objective and comparative approach to develop a methodology to calculate construction costs of various styles of Hohokam domestic architecture.
Various elements of our methodology have been published previously as supplementary materials linked to our earlier article (Abbott et al. Reference Abbott, Craig, Zanotto, Judd and Kober2019). They comprise a narrative and other texts to usher the analyst through the construction-cost calculations. The texts include a list of defined variables and the formulas with which to calculate them. Constants, such as wall height and the rate of adobe mixing, are also listed. In addition, three examples of the costs measured for individual structures will aid future applications.
There are many calculations needed to measure the labor costs for each of the Pueblo Grande structures. We rely on base rates obtained from experimental archaeology that translate the amounts of building materials into labor costs measured in person-hours. There were three primary components for construction costs: (1) the labor to dig and mix the adobe; (2) the labor to cut and transport the wood; and (3) the labor to erect the walls, floor, roof, and internal features of the structure. By summing these various expenditures, we derive the total labor cost of the structure. Also, an estimate of the relative cost can be computed by dividing the total cost of the building by the total floor space.
Results
Our analysis was designed, in part, to test if several expectations related to wealth and status differentials were evident in the architectural details at Pueblo Grande. Among the expectations were differences in the architectural costs and extravagance across the Pueblo Grande habitation areas as well as the long-term accumulation of social and economic advantages differentially enjoyed across households and expressed in the architectural history of the village. We also took advantage of the comparable data previously amassed at Grewe to determine if the architectural indicators of wealth and status recognized there during the Sedentary period were similarly apparent during the contemporaneous part of the occupation at Pueblo Grande. In addition, we compare the Pueblo Grande and Grewe results to examine if imbalances between irrigable land and population size possibly set limits to resources, thereby stimulating social inequality. Before moving to a discussion of these topics, we first examine several findings pertaining to structure size and function.
Size and Function
We began by calculating size (square footage) of the floor and considering its distribution. The distribution of house sizes at Pueblo Grande (Figure 5) fit well with Hohokam structures in general. Wilcox and others (Reference Wilcox, McGuire and Sternberg1981:158) and Crown (Reference Crown1985:76) found that typical Hohokam dwellings had floor areas greater than 10 m2, with small structures being much rarer and likely used for a special purpose. In our database, nine cases had floor areas under 10 m2 (Table 2), and they could be divided into two groups.
Figure 5. Histogram of structure size.
Table 2. Small Structures at Pueblo Grande.
The smallest four rooms had all of the largest relative costs among all structures at Pueblo Grande, and none had hearths. Crown (Reference Crown1985) noted that small structures with insubstantial construction—often lacking roof supports and built with unlined hearths—were found on the outskirts of Hohokam settlements. They possibly served as places for the purification of warriors or as women's shelters for puberty, menstruation, or childbirth. Their peripheral location and makeshift construction, however, poorly mirrors the smallest examples in our dataset. Instead, the small structures at Pueblo Grande were (1) built inside HAs, (2) substantially built (as indicated by the high relative cost), and (3) not heated with a hearth. These characteristics fit well with expectations for storerooms (Crown Reference Crown1985:83; Doyel Reference Doyel1981:27–31; Haury Reference Haury1976:68).
The five small structures in the second group were all very near 10 m2 in size (see Table 2). All but one had a hearth—the exception was a Massive-Walled Adobe room in HA-8. It had the next highest relative cost, and we interpret it as having been a large storeroom. We also conclude that the other four examples were small but otherwise typical structures, perhaps providing housing for individuals (Crown Reference Crown1985:84; Haury Reference Haury1976:62, 68).
Pueblo Grande Differentials
As a first and general assessment of the wealth distribution at Pueblo Grande, we extracted the total cost for all of the 106 Classic period habitation structures in our database and calculated a Gini coefficient. The Gini index is a standard quantitative measure of inequality (Smith et al. Reference Smith, Kohler, Feinman, Kohler and Smith2018:25). Its values range from zero (perfect equality, all units have the same amount of wealth) to near 1 (complete concentration, one unit has nearly all of the wealth). The Gini score for the Classic period structures at Pueblo Grande was surprisingly low (0.19), signifying a general uniformity of domestic construction costs across the village and a near absence of social differentiation reflected in the house data. The corresponding score for the 22 Sedentary period Type S-1 pithouses in our sample was even smaller (0.12). It emphasized social equality at Pueblo Grande over the long term during the Sedentary and Classic periods. Admittedly taken aback by the difficulty of interpreting those results, we sought further refinement by digging more deeply into the construction data.
Our search for architectural expressions of possible wealth and status differentials at Pueblo Grande continued by plotting the total labor cost for each structure by nine categories of architecture. The categories correspond to a temporal sequence of architectural styles, including Type S-1 pithouses of the Sedentary period; Type S-2, Rock-Lined, and Deep Post-Supported pithouses of the very Early Soho phase; Deep, Adobe-Lined pithouses and Post-Reinforced surface structures of the Early Soho phase; Narrow-Walled Adobe pithouses of the Late Soho phase; Massive-Walled Adobe structures of the Civano phase; and finally, Postclassic dwellings (i.e., Narrow-Walled Adobe and Deep, Adobe-Lined pithouses built during the latest temporal interval at Pueblo Grande).
When we divided the plot of total labor cost for each structure using the median cost for all structures in our sample (309 person-hours), nearly all of the Sedentary S-1 pithouses fell below the median (Figure 6). The total cost of the examples of all of the other architectural styles also plotted below or near the median, with the exception of the Post-Reinforced and Massive-Walled Adobe cases. The total costs for these two architectural styles were notably more diverse, which is a point we return to below.
Figure 6. Total labor cost by architectural style (median = 309).
Here, we note that the variability in total labor across our entire sample was largely explained by structure size. When we plotted size against the total labor (Figure 7a), a strong positive correlation was clearly apparent (Pearson's r = 0.821). About two-thirds of the variation in total labor costs was explained by size alone (r 2 = 0.675).
Figure 7. Scatterplots: (a) total labor cost by structure size, (b) relative cost by relative adobe cost, and (c) relative cost by relative wood cost.
The high diversity in the total labor costs of the Post-Reinforced and Massive-Walled Adobe structures corresponded to several unusually large houses. All of these big Post-Reinforced and Massive-Walled Adobe dwellings were found in HA-6 (see Figure 6). This HA was also distinguished among its peers in the SSI project area as having been one of the longest-inhabited residential groups (established during the Sedentary period), and it was the closest to the village's central precinct and platform mound. It also contained Feature 2795—the only example of an especially large, enclosed, and probably unroofed court found in the SSI project area (Figure 8).Footnote 2 Such enclosures, albeit much larger in size, were common in the ritually charged space atop the platform mound (Downum and Bostwick Reference Downum, Bostwick and Abbott2003).
Figure 8. HA-6 at Pueblo Grande.
HA-6's distinction may also have been a long-term one, lasting for 250 years from the Early Soho to the Civano phase. There are no extra-large Narrow-Walled Adobe constructions of the Late Soho phase preserved in HA-6 (see Figure 6), which muddles an otherwise strong pattern of outsized structures in HA-6 over time. We can only speculate that distinctively large narrow-walled cases were built in HA-6 but did not survive in the archaeological record. If so, the HA-6 occupants continuously distinguished themselves from their neighbors via their occupancy of notably larger dwellings for more than two centuries (from the Early Soho–Post-Reinforced, to possibly the Late Soho–Narrow-Walled Adobe, to the Civano–Massive-Walled Adobe).
The next question we asked was whether the architecture in HA-6 or elsewhere was particularly extravagant in its cost per square meter. When we plotted the relative costs for the nine architectural/temporal categories, most of the categories included structures with relative costs that fell only below or near the median value (19.98 person-hours/m2). Only examples of the Post-Reinforced, Narrow-Walled Adobe (both Late Soho and Postclassic), and the Massive-Walled Adobe structures were characterized by high relative costs (Figure 9). There were two extreme outliers among the Massive-Walled Adobe buildings (Features 2010 and 669.22). Both were in HA-5, and both were extremely small (see below).
Figure 9. Relative cost by architectural style (median = 19.98).
As we dissected the variation in relative cost, there were three findings of particular note. They pertained to (1) the underlying factors that contributed to the differences in relative cost, (2) the two extreme outliers and other small rooms at Pueblo Grande, and (3) additional patterning that further distinguished the architecture in HA-6.
Underlying Factors
One factor alone, adobe costs (digging, mixing, and applying the adobe), accounted for nearly all of the variation in relative costs among the various architectural styles. In short, Post-Reinforced, Narrow-Walled Adobe, and Massive-Walled Adobe structures often had higher relative costs because they were constructed with more adobe per square meter than the other architectural styles. A clear linear relationship was exhibited between relative cost and relative adobe (RELADOBE = adobe/m2; Figure 7b), which was associated with a Pearson's r2 = 0.843. A similar relationship between relative wood costs (RELWOOD = wood/m2) and relative cost did not exist (Figure 7c).
Small Outliers
The two extreme outliers in Figure 9 were the smallest two structures in our database. Each had floor areas less than 2.5 m2. These tiny spaces were furnished without a hearth and were almost certainly used as storerooms. They were made with solid adobe walls, and their extreme relative cost values were probably explained by our analytical procedures. All Massive-Walled Adobe structures, regardless of size, were assumed to have had walls that stood 1.8 m high.Footnote 3 Based on that assumption, an extreme amount of adobe was used to build the walls relative to the square footage of those tiny rooms.
Returning to the correlation between relative cost and the relative amount of adobe, it was clear that the smallest structures were out of place relative to the larger habitation rooms (see Figure 7b). When we eliminated the nine small structures (see Table 2) from the calculations, a Pearson's r = 0.938 (r 2 = 0.880) was associated with the distribution. Consequently, when the anomalous small structures were removed, nearly all (88%) of the variation in the person-hours/m2 across all of the architecture styles was explained by the amount of adobe/m2. This result highlights the great expense incurred during the Civano phase when Massive-Walled Adobe structures were built with towering adobe walls surrounding the compound of each HA. At Pueblo Grande, a considerable part of the expense for erecting Massive-Walled Adobe rooms was reduced by the builders when they added immediately available slabs of indurated caliche to make up about 30% of the volume in their walls. Across the Hohokam territory, massive-walled compound rooms and the surrounding compound walls were common during the Civano phase (e.g., Gregory Reference Gregory and Doyel1987:208), and they were typically made with solid adobe (Wilcox and Shenk Reference Wilcox and Shenk1977). Such construction undoubtedly pushed the relative cost even higher.
Another Distinction for HA-6
The graph in Figure 9 also pointed to a further distinction for HA-6. Among the Sedentary period S-1 pithouses, two structures (Features 780 and 3110) in HA-6 stood out as having unusually large relative costs. Feature 780 was actually located among a group of pit structures situated 15–20 m south of the main area of HA-6 (see Figure 8). It may have been part of a separate Sedentary period residence cluster that did not persist into the Classic period or that became consolidated into HA-6 during the Soho phase. Feature 780 was one of the small structures at Pueblo Grande listed in Table 2. Its high relative cost corresponded, in part, to an especially thick floor plaster that drove up the adobe costs, but it was also attributed to the structure's limited floor area. Like most other S-1 pithouses, Feature 780 utilized two main roof supports and a single primary roof beam. Consequently, the wood costs to build it were similar to those for other, larger pithouses but at a higher cost per m2.
The high relative cost for constructing Feature 3110, in contrast, could not be explained by a diminutive size. Its 12.8 m2 of floor space was well within the “normal” range of Hohokam pit dwellings. What distinguished it was an unusually thick plastered floor and post-supported internal benches that covered more than a third of its floor space, mostly in the back of the structure. In this way, Feature 3110 was unique among all other structures excavated at Pueblo Grande. Internal post-supported furnishings were uncommon at Pueblo Grande. When present, they typically covered only a small percentage of the floor space (the next largest covered 19%).
Comparisons to Grewe
The Grewe analysis showed previously that pithouse size, total cost, and elaborateness of construction distinguished some Preclassic courtyard groups from others within that village, and those differences persisted over time. On that basis, the emergence of status differentiation and wealthy households within the Grewe community became apparent in the architectural details (Craig Reference Craig and Craig2001c, Reference Craig and Beck2007). We next sought to determine how the Sedentary period architecture at Grewe and Pueblo Grande were similar to or different from each other.
The boxplots in Figure 10, as well as the coefficients of variation in Table 3, show that the variability in the total cost, size, and relative cost were all greater among the Grewe pithouses than those at Pueblo Grande. These findings correspond to the Gini index calculations presented above as well as to a wide gap in Gini scores for house size reported by Pailes (Reference Pailes, Kohler and Smith2018:Figure 6.3) between the Salt River and Gila River valleys during the Sedentary period. The Gini values for the Salt River settlements were all below approximately 0.20, and above approximately 0.25 for the Gila River communities. All of the evidence together shows that some Grewe pithouses were large and others were small. Some domestic structures were expensive, and others were much less so. And some were built lavishly or made more durable when others were not, signifying a degree of social differentiation within the Grewe community. The variation at Pueblo Grande was less heterogeneous, implying less social differentiation expressed by the architectural dissimilarities. When measured against the Grewe architecture, Pueblo Grande house size and total labor costs were uniformly greater, implying the Sedentary period households at Pueblo Grande were typically wealthier than their Gila River neighbors.
Figure 10. Boxplots comparing Grewe and Pueblo Grande: (a) structure size, (b) relative cost, and (c) total labor.
Table 3. Comparison of Grewe to Pueblo Grande.
a includes Type S-1 pithouses only.
The high relative costs at Grewe (see Figure 10b) are partly associated with the “small-structure” effect described above. But they also reflect the elaborateness of many of the pithouses, including both small and large structures. Most of the small structures with high relative costs had plastered floors and benches or platforms that covered at least half of the floor. Benches and platforms were common in large structures as well, particularly habitation structures associated with the two long-lived courtyards discussed previously (Craig Reference Craig and Craig2001b:Figure 6.28). In general, raised floor features were far more common at Grewe than at Pueblo Grande. Almost a third of the houses at Grewe had benches or platforms that covered at least 35% of the floor, whereas the only pithouse at Pueblo Grande with a bench or platform covering that much area was Feature 3110 (see above).
Theoretical Fit
The architectural differences during the Sedentary period between Pueblo Grande and Grewe were clear and unequivocal, but they seem at odds with theoretical expectations. The model examined here claims that where access to productive land is made scarce by population growth, the stage is set for the emergence of social inequality (Bogaard et al. Reference Bogaard, Styring, Whitlam, Fochesato, Bowles, Kohler and Smith2018:203; Flannery and Marcus Reference Flannery and Marcus2012; Goody Reference Goody1976:97; Kohler and Higgins Reference Kohler and Higgins2016:691; Kohler et al. Reference Kohler, Smith, Bogaard, Feinman, Peterson, Betzenhauser, Pailes, Stone, Prentiss, Dennehy, Ellyson, Nicholas, Faulseit, Styring, Whitlam, Fochesato, Foor and Bowles2017; Smith et al. Reference Smith, Kohler, Feinman, Kohler and Smith2018:10). As it turns out, however, theory and Hohokam architecture fit neatly together.
The prime mover for the inequality model is increases in population levels, which at Pueblo Grande were experienced during several waves of immigration, including the first one at the end of the Sedentary period. These influxes presumably were a constant strain on social institutions and economic arrangements, which may have included access to irrigated farmland. The pressure of unequal access to a vital resource may have established the conditions for social inequality. The model, however, raises the question of why we do not see inequality more clearly in the architectural variation at Pueblo Grande. An important new study by Christopher Caseldine (Reference Caseldine2020) of the Salt River hydraulic infrastructure helps us to understand.
One of Caseldine's (Reference Caseldine2020) principal findings is that water scarcity was not a major limiting factor for agricultural production along the Salt River during the Sedentary period. The supply of water was almost never insufficient to irrigate all of the planted fields. Inequalities, therefore, were probably relatively few. Interestingly, the masterful balance between water demand and supply along the Salt River was unachievable along the Gila River, where the stream flows were just one-third of the Salt River discharges, and the narrow terraces made productive land an unevenly dispersed and insufficient resource. Hunt's (Reference Hunt2007) simulation of maize production predicted that water shortages regularly plagued the Gila farmers (see also Zhu et al. Reference Zhu, Woodson and Ertsen2018:743–744). Consequently, unlike the Salt River valley, architectural expressions of inequality were salient along the Gila River valley. Consequently, our comparative results now make sense. During the Sedentary period, wealth (large and costly houses) was common at Pueblo Grande and wealth differentials were customary at Grewe.
Discussion
Our methods were designed for application across all of the Pueblo Grande structures, allowing comparisons of the construction costs of different house styles, and, consequently, tracking the architectural labor costs over time. Moreover, we expect these methods will, with only minor adjustments (e.g., distance to obtain building materials), be applicable across the Hohokam culture area and elsewhere. As exemplified by our anlaysis of Grewe and Pueblo Grande, the Pueblo Grande dataset may serve as a baseline for comparison with future architectural databases from new excavation projects or from extant archival collections. Indeed, the abundance of high-quality evidence from Pueblo Grande will serve as a measuring stick against which even small datasets can be usefully assessed as being more or less like Pueblo Grande.
Our analysis showed that most of the overall construction cost for each of the Pueblo Grande houses was accounted for by structure size, as measured by the square footage of the floor space. To some readers, this result might imply that the analytical effort to measure house size alone would be sufficient to capture the significant cost differentials among Hohokam domestic structures. By this logic, a simple square-footage calculation of floor length by floor width is all that is needed. In addition, the concerted effort spent to make the detailed measurements used to calculate adobe, wood, and building costs would be superfluous for many future studies. We think, however, that this reasoning would be a mistake. It is because we made the measurements that we can be quite confident that the Pueblo Grande architecture was largely undifferentiated by elaborate relative costs. We also do not expect the architecture across the Hohokam culture area to necessarily conform to the Pueblo Grande cases. Indeed, at Grewe, the relative expenditures for constructing pithouses were diverse, and most structures had higher relative costs than at Pueblo Grande. It was the diversity and the high relative costs at Grewe that were partly related to status-based differentials. At Grewe, structure size alone leaves out much of what is important about construction costs.
What we found at Pueblo Grande among the domestic architecture, which had not been noticed in its entirety before, were architectural differences probably associated with wealth and status differentiation across the residence groups. One habitation area—HA-6—was exceptional in several ways. It stood out from its peers for 300 years, from its inception in the Sedentary period through the end of the Late Classic period. Two of its Sedentary pithouses were distinguished by extreme costs per square meter. Both had thickly plastered floors, and one was built with raised benches covering one-third of the internal space—an extravagance not seen in any other structure excavated at Pueblo Grande. Two exceptionally large Post-Reinforced structures were built during the Early Soho phase. Five extra-large Massive-Walled Adobe rooms dominated HA-6 in the Late Classic period along with a unique walled, unroofed, open-air court. The five largest houses in our sample, which required roughly 50%–90% more labor to build than the typical house, were all located in HA-6. These exceptionally large houses were probably associated with larger households, with more people and larger potential labor pools at their disposal (Abbott and Foster Reference Abbott, Foster and Abbott2003:41). HA-6 also contained a large, walled-in, unroofed courtyard (Feature 2795), which required almost as much labor to build as the typical house. Consequently, access to domestic labor appears to have been a key source of differentiation between households at Pueblo Grande. HA-6 was one of the longest-inhabited residence units in the SSI project area, and it was the habitation area closest to the village's central precinct and platform mound. Each of these unusual occurrences on its own might raise a passing interest, but their congregation in just one residence group is a clear sign that the households of HA-6 continuously benefited from the form of social inequality that architectural distinctions reflect (see also Foster et al. Reference Foster, Mitchell, Breternitz, Robinson, Coupland and Banning1996:32).
One of the advantages of a labor-based approach for measuring household inequality is using a common unit of comparison applicable to different generations of house builders. Consequently, we can document the extent to which households living in HA-6 were able to maintain their position at or near the top of the social hierarchy for centuries. Such persistence is a hallmark of institutionalized inequality (Price and Feinman Reference Price, Feinman, Price and Feinman2010:2). It is also evidence for the intergenerational transmission of wealth (Bowles et al. Reference Bowles, Smith and Mulder2010). In most instances, however, the wealth was not in the architecture. The architecture was not passed down from one generation to the next, with the possible exception of Massive-Walled Adobe rooms, which were more durable and likely had longer use-lives than other architectural styles (Abbott and Foster Reference Abbott, Foster and Abbott2003:41–45). The primary form of material wealth transmitted across generations was probably land that came with household membership, including both residential property and agricultural fields (Craig Reference Craig2010). Like many Hohokam villages, Pueblo Grande was internally structured, with habitation areas arranged in a semicircular plan around a central plaza and platform mound (Abbott and Foster Reference Abbott, Foster and Abbott2003:25). The larger, more labor-intensive houses in HA-6 suggest that the residence unit may have been part of a high-status residential district bordering the platform mound complex (Foster et al. Reference Foster, Mitchell, Breternitz, Robinson, Coupland and Banning1996:41). Further assuming that more people lived in these exceptionally large houses than in the typical house, the residents of HA-6 may have had access to larger and better tracts of arable land, given that studies of intensive agricultural societies have consistently shown a strong correlation between the size of a household and the size of its land holdings (Douglass and Gonlin Reference Douglass, Gonlin, Douglass and Gonlin2012:13–15; Netting Reference Netting1982, Reference Netting1993:85–87).
At both Pueblo Grande and Grewe, high-status households were distinguished architecturally by larger and, in some instances, more elaborate houses. They were also distinguished by where they lived—near the communal cooking area and ball court in the case of Grewe, and adjacent to the platform mound complex in the case of Pueblo Grande. The proximity of these households to public areas for ceremonial expression lends support to the idea that ritual participation played a key role in the emergence and persistence of inequality (Blanton Reference Blanton, Price and Feinman1995:121–123; Mills Reference Mills and Mills2000:8–9; Vésteinsson et al. Reference Vésteinsson, Hegmon, Arenborg, Rice and Russell2019). Nearby households presumably had greater admittance to the facilities and exercised greater control over their use. They were likely also among the main sponsors and beneficiaries of any ritual events that took place under their watch. In some instances, high-status households may have even been involved in planning and performing the events, as indicated by the presence of a large, walled-in courtyard in HA-6 that was well-suited for rituals and ceremonies, similar to courtyards on the platform mound (Downum and Bostwick Reference Downum, Bostwick and Abbott2003:194). Importantly, these ritual spaces represent public arenas where inequality and status are frequently negotiated and confirmed (Vésteinsson et al. Reference Vésteinsson, Hegmon, Arenborg, Rice and Russell2019:188).
Our comparison of the Sedentary period architecture at Pueblo Grande and Grewe also proved useful for examining a model for the emergence of social inequality in middle-range societies. The model stipulates that population growth can upset the human/resources ratio, leading to scarcities of vital resources. In turn, should differential access to essential assets develop, advantages could accrue to only some community members, leading to the emergence of social differentiation. Just as the model would predict, the Pueblo Grande residents, who enjoyed an abundant water supply, lived in relatively large homes without much distinction among themselves. In contrast, the people of Grewe likely felt the pressure of population increase on the supply of water and irrigable land, which led to wealth dissimilarities manifest in the highly variable sizes, total costs, and relative costs of their abodes.
Conclusion
Studies of middle-range societies, such as the Phoenix Basin Hohokam, offer the opportunity to investigate the origins and evolution of institutionalized inequality. In particular, the size and elaborateness of construction of domestic architecture can reflect differences in household status.
For this article, we had a threefold purpose. First, we recognized that exploring the development of inequality requires an ability to detect it in the archaeological record in as many ways and in as many cultural contexts as possible. Despite the social and economic complexities in Hohokam society, signs of permanent inequality are rare, and they are the subject of considerable debate. Spurred by previous success with the architectural variation in Preclassic pithouses at Grewe, we developed methods to measure the construction costs of various styles of Hohokam domestic buildings based on archival data. We believe our flexible approach, which can include various kinds of house construction, can be easily adapted and applied to case studies from across the Hohokam territory and possibly beyond.
Second, we analyzed the construction costs for domestic architecture at Pueblo Grande and Grewe. Our analysis showed evidence for institutional inequality in the architectural remains in HA-6 as compared to the other residence groups at Pueblo Grande. The data suggested that the conditions that promoted inequality may have been related to the advantages enjoyed by well-established and long-lived residential groups over newcomers. These included land tenure and the well-established social networks with which to mobilize labor. Wealth in land and its intergenerational transmission are keys to persistent inequality in societies practicing intensive agriculture. In addition, the spatial associations between high-status households and ritual facilities (i.e., the ball court at Grewe, the platform mound at Pueblo Grande) signal that control over public functions may have been a path to political advancement for Hohokam leaders and their families.
Third, we examined a model about the emergence of social inequality in premodern agricultural societies, and its fit to the Hohokam case. A comparison of the Sedentary period architecture at Pueblo Grande and Grewe proved to match theoretical expectations quite well, thereby supporting the model and illustrating the utility of our approach for measuring architectural construction costs for studying household wealth and social inequality.
Acknowledgments
Just as we were finishing our manuscript, our dear friend and coauthor, Doug Craig, passed away. It was pure pleasure working with Doug on this project and on our preceding effort (Abbott et al. Reference Abbott, Craig, Zanotto, Judd and Kober2019). We dedicate this article to Doug for his many outstanding contributions to Hohokam archaeology. His brilliance and insight will live long in our memories.
We are most grateful to Mark Elson, Suzy Fish, Kathy Henderson, Bob Hunt, Doug Mitchell, Mike Smith, and two anonymous reviewers for their advice and suggestions. Their input significantly improved the quality of our presentation. We also thank Pueblo Grande Museum for giving us access to the SSI excavation notes and records, and for giving us space to work. We are especially thankful to Holly Young at PGM for answering all of our questions and guiding us through the SSI archives. This article absolutely would not have been possible without Holly's assistance.
Data Availability Statement
The SSI Pueblo Grande materials are curated at the Pueblo Grande Museum in Phoenix, Arizona. A file of the architectural data recorded for this study will be made available on request.
