Refuse or Ritual Deposit? The Complexity of Wari Household Archaeology

Abstract

The excavation of residential areas is a growing focus of research in Andean archaeology. Studies reveal that interpreting household remains from some prehispanic societies can be complex because of the nature of abandonment ritual, which may involve burnt offerings, the placement of valuables on floors, or the purposeful destruction of ceramic vessels that are distributed in patterned ways. The goods that constitute these offering practices can be confused with post-occupation refuse, especially when excavation units are relatively small. In this paper, I discuss the importance of assessing site formation processes in residential spaces and illustrate how different modes of household abandonment can make comparative analysis a complex exercise. I describe and compare several examples from Wari-affiliated residences at the sites of Cerro Baúl and Cerro Mejía, located in the department of Moquegua, Peru, to show how ritual depositions corresponding to house abandonment might affect the interpretation of daily domestic life. In particular, I examine how ritual assemblages have been interpreted as evidence of feasting to support propositions regarding the Wari political economy. I advocate that archaeologists interested in domestic areas, lifeways, and the political economy engage in large-scale horizontal excavations to ensure they can correctly distinguish between the remains of quotidian practices, the goods associated with ritual depositions, and refuse resulting from feasting, which is best substantiated with features and facilities to host empowering events.

1. Introduction

The Wari Empire of South America spread throughout much of modern-day Peru during the first millennium CE. Wari iconography, D-shaped temples (a form unique to the empire), and ritual practices are found in many regions (Bergh, 2012; Reid, 2023). One such practice is the purposeful smashing of pottery and the placement of other goods on floor surfaces or in stratified pits. This general type of deposit created through ritual action upon the abandonment of a building was also practiced by the Tiwanaku, Wari’s neighbor to the south, which means its performance may be affiliated with a religion, cult, or broader cosmology rather than being restricted to a particular polity. Nevertheless, some Wari deposits are spectacular; the practice is seemingly more widespread within the Wari Empire, and it has drawn the attention of scholars studying the polity. Ritual abandonment deposits have been identified in Wari D-shaped temples, other sacred precincts, and residential structures (Cook, 1987, 2001; Glowacki, 2012; Isbell, 2000; Ochatoma Paravicino et al., 2015). In this paper, I consider how the components of ritual deposits associated with house abandonment have shaped interpretations of the Wari political economy and statecraft. I suggest that many residential assemblages are partially composed of ritual deposits and that determining the nature of features during excavation is essential for interpreting household assemblages, understanding social differences, and discerning the organization of the Wari Empire. I present a case study that details the material patterns of domestic assemblages. Even though such patterns vary between societies, it demonstrates the importance of defining activity areas and quotidian practices for the investigation of complex polities.

Defining the context in which artifacts are found is necessary to answer nuanced questions about the formation of institutions in early states. I contend that this process begins during excavation and includes correctly interpreting the nature of features and collecting materials in a way that facilitates the identification of activity areas. In furtherance of this strategy, I describe the types of ritual deposits, how they differ from domestic features, and the offerings selected for inclusion based on Wari-affiliated houses at Cerro Baúl and Cerro Mejía in the department of Moquegua, Peru. I emphasize the difference between refuse, items discarded and left behind or dumped by neighbors in abandoned houses, and ritual deposits, which consist of goods selected for sacrifice that held value and together are the material expression of a specific activity. The contrast between ritual goods and garbage or de facto refuse that archaeologists expect to find in domestic contexts is fundamental because ritual goods were selected as offerings and are salient elements of ritual assemblages that may be used to infer ritual practices and religious affiliations. This is very different from abandoned refuse or the debitage of daily living from which archaeologists infer the regular activities taking place in a particular residential setting. Ritual goods may be special and not a common component of quotidian domestic practice. Since inferences that consider ritual goods as elements of residential assemblages may skew interpretations of many kinds, such as the prevalence of feasting and its role in the political economy, I briefly introduce the Wari Empire and review how Wari domestic architecture and ritual offerings have been interpreted before presenting patterns observed among house remains in the southern province of Moquegua, Peru.

1.1. The Wari Polity as an Empire

In the Andes, the Moche, Tiwanaku, and Wari are all described as expansive states because they grew over time and eventually occupied areas outside of their original territory during the first millennium CE. If these polities were labeled as empires, cross-cultural research offers expectations for testing (e.g., Alcock et al., 2001; Boozer et al., 2020). Subjects by conquest, alliance, intermarriage, or economic coercion could be controlled directly or through cooperating local leaders, who might be offered access to more labor and a greater variety of resources. Colonists filling empty econiches may have come from the core of the polity, but it is just as likely that they originated from subsidiary regions (Nash, 2022a). In any case, Andean archaeologists have avoided calling these polities empires for a very long time. They are not alone. Archaeological models of political expansion discuss the activities of state-level societies, whereas empires are the province of civilizations with a decipherable, recorded history.

Archaic states, such as the earliest empires, command theoretical interest because their institutional configurations differed from their successors (Adams, 2000). Researchers exploring state formation have found that these polities are not merely small versions of later states but are experimental regimes that developed institutions, which can only be understood through archaeology (Algaze, 1993; Feinman & Marcus, 1998; Feinman, 1998). As these fledgling states expanded beyond their geographic core and incorporated foreign subjects, they would have confronted limits and made strategic decisions to meet administrative needs (Flannery, 1999; Stanish, 2003). The Wari polity is an excellent example to study these processes of institutional development, and household archaeology offers multiple types of data to test competing models and understand elite strategies of governance as early expansive polities attempted to stabilize provincial holdings and maintain relations of control.

Wari expansion, which commenced sometime between 400 and 600 CE (Nash, 2022a), changed the ecology and demography of many regions forever. Some areas were largely abandoned (Eitel & Mächtle, 2009), while others were newly populated, such as the upper Moquegua drainage. The capital swelled with people and was truly a cosmopolitan city of 50,000 people or more. Located in the department of Ayacucho, Wari (also sometimes spelled Huari) was 15 km² in area and the center of an empire, which stretched across 1300 km of the Andean sierra and established enclaves in several coastal valleys as well (Figure 1). Wari had multiple monumental temples, palaces, and multi-storied buildings (Lumbreras, 2007). Centers built in several highland regions follow Wari architectural canons, exhibit similar conventions of spatial organization, and are notably intrusive in style (McEwan & Williams, 2012). Each represents a huge construction project supported by equally large landscape transformations and infrastructural developments (e.g., lengthy canals, reservoirs, and terraced hillslopes), which required a command of labor and local resources. Wari-affiliated sites on the coast have been more difficult to detect and may represent smaller enclaves, different relations with local elites, or somewhat autonomous allies (e.g., Bracamonte, 2022; Conlee et al., 2021; Giersz, 2016).

Settlement patterns are not well defined in most regions and are complicated by the fact that many Wari-affiliated sites, especially those of modest size, do not have many, if any, Wari-style decorated ceramic vessels or only have those judged as mediocre in execution, which have often been considered local copies or “knock-offs”. Researchers working in different areas interpret these material expressions in a number of ways; the link between decorated pottery, political affiliation, religion, or other types of group identity has not been problematized (Nash, 2018, 2019; for an exception see Watanabe, 2022). The situation is complicated as decorative schemes and icons from several regions and/or polities (e.g., Cajamarca, Huancavelica, Nasca, and Tiwanaku) contributed to the multiple named styles associated with the empire (Knobloch et al., 2023; Ochatoma Paravicino et al., 2022a, 2022b). Hybrid examples are common, and some regions produced distinctive, recognizable styles to carry icons linked to Wari incursion (i.e., Nieveria). Recent typologies are helpful, but tracking these goods to the exclusion of others provides a partial picture of the polity.

Many maps of the Wari Empire illustrate a shaded area, which would suggest a continuous territorial realm, but ethnohistorically reported types of land tenure, resource movements (supply chains), and demographic distributions (e.g., Chacaltana, 2010; Murra, 1968) suggest that a complex patchwork based on network relations of interacting and cooperative elites (with attached subjects in train) is equally possible. An imperial hierarchy, which may have included some semiautonomous allies (e.g., Cajamarca), was connected by roads that facilitated the movement of goods and people (Edwards & Quiñínez-Cuzcano, 2023; Williams, 2023), who carried ideas and technologies with them. Similarities in elite practices, some of which may derive from state institutions (Nash & Williams, 2005), offer a means to connect the polity’s core with centers of power in different regions; smaller communities would be linked to the overarching empire through local leaders, and each province’s practices may differ based on pre-existing patterns and the type of relationship between decision-makers in the capital and officials charged with that area (Nash, 2021). This type of political reality is difficult to illustrate when only the major sites in each region can be included. Also, if communities continue to use goods traditional to the region, settlements without radiocarbon dates may be attributed to the previous period because of reliance on surface pottery during surveys; new settlements and changes in regional production regimes instigated by imperial intrusion would go unrecognized. In short, studying an empire without historical records is challenging and requires a multi-pronged approach. The elements of domestic activities and ritual deposits are present at the majority of sites and offer variables that can be compared across settlements, within regions, and between provinces.

1.2. Wari Administrative Architecture

Ideas about Wari residential activity were primarily based on patio group structures in urban complexes at the capital. Patio groups were described as nuclear family dwellings (Brewster-Wray, 1989), and these units in some contexts probably do correspond to some kind of salient familial grouping, but other examples correspond to specialized activities (Nash, 2011) or appear relatively empty rather than residential (Anders, 1986). A generalized patio group form is found among many groups in the Andes, but Wari patio groups are distinctive in that the patio is in the center and has rooms that enclose it completely on the four sides (Figure 2). They differ from later Inka patio groups, called kancha, because Inka buildings may have rooms on all four sides but do not abut to circumscribe the patio (Gasparini & Margolies, 1980; Hyslop, 1990; Figure 2).

Figure 2. Examples of patio group units within larger complexes. See also Figure 3. (A) The Moraduchayoq compound at the Wari capital. Note that the walls were based on wall trenching, and some are projections of incomplete walls (see Isbell et al., 1991). (B) The Qorikancha, the main Sun Temple of the Inka Empire, located in Cusco. Redrawn from Gasparini and Margolies (1980). Note that the thick black lines represent original Inka walls.

Figure 2. Examples of patio group units within larger complexes. See also Figure 3. (A) The Moraduchayoq compound at the Wari capital. Note that the walls were based on wall trenching, and some are projections of incomplete walls (see Isbell et al., 1991). (B) The Qorikancha, the main Sun Temple of the Inka Empire, located in Cusco. Redrawn from Gasparini and Margolies (1980). Note that the thick black lines represent original Inka walls.

The patio groups excavated at the Wari capital were in the Patipampa sector within the Moraduchayoq compound (Schreiber, 2012, p. 37, figure 25), which is within 300 m of an elaborate set of subterranean burial chambers (Mongachayoq) and a little more than 400 m from the site’s sacred precinct (Vegachayoq Moqo). The compound’s size, which is just under 0.3 ha, does not approach the proposed palace complexes, which range from 7 to 12 ha (Ochatoma Paravicino & Cabrera Romero, 2023), but its location is consistent with city dwellers of elevated status.

The features of the Moraduchayoq compound were elevated as diagnostic of Wari administration, carried out by mid-level authorities in the capital and by Wari-affiliated officials in the provinces, by a foundational edited volume, Huari Administrative Architecture (Isbell & McEwan, 1991). Several chapters emphasized the similarity between sites through the presence of patio groups rather than a shared plan with multiple corresponding sectors in provincial centers. Based on perceived parallels with the Inka mode of production (Godelier, 1977), patio groups were described as venues for feasting that were integral to Wari administration because hosting feasts was a means to maintain relationships between administrators and their followers (Glowacki, 2005; Isbell et al., 1991), who could leverage these relations to command labor or extract resources.

Archaeologically, Andeanists equate feasting with administration but also recognize its importance as an element of ritual (W. Isbell, 1978, 1991; Morris, 1982). Linking artifactual evidence of feasting to a specific purpose (i.e., power strategies or political control) is very difficult. Ethnographic research offers examples where community gatherings often have overlapping motives. Ritual can legitimize political power; however, the existing political capital of a ritual sponsor is often used to pull together the resources for celebratory gatherings. Ritual, economic, and political aspects of commensal gatherings are not easily delineated (Abercrombie, 1986; Dietler, 2001; B. Isbell, 1978; Rasnake, 1986). Models derived from community interactions, which can be less asymmetrical than those sponsored by states, may not be appropriate for all scales.

The Moraduchayoq compound had five or more patio groups located together in a walled complex based on wall trenches and targeted excavations. Some of the patio groups could only be accessed by passing through another or from other types of spaces in the compound. No patio group was excavated in its entirety; rooms, parts of rooms, and portions of patios were compared with regard to artifact types, ceramic vessel forms, density, and other attributes (Brewster-Wray, 1990).

Analysis of the assemblage composition and spatial organization was used to propose that domestic and administrative activities were both encompassed within patio group structures; however, because of site disturbance, analyses could not be made from an activity area approach. Residential routines were inferred from “domestic garbage” called “dumps”. Materials found in “dumps” were presumed to be in secondary context but considered as evidence of typical quotidian practices, which had been deposited near their original activity areas, the other patio groups units that continued to be occupied in the compound (Brewster-Wray, 1989; Isbell et al., 1991). Dumps were contrasted with floor deposits based on density, the size of sherds, and other post-excavation attributes. High percentages of serving wares were interpreted as the remains of feasting residue and thus administrative activity. High-value items were also among the “refuse” assemblages and supported the notion that the householders held a relatively high status (Brewster-Wray, 1990). Preciosities were present in areas interpreted to be floor deposits and those defined as dumps.

The scope of this research was laudable; however, it was hampered by a lack of other equally well excavated and studied residential units for comparison at the capital and the degree of looting throughout the city. In addition, follow-up studies were impossible because of the threat of terrorism in the 1980s and 1990s (Schreiber, 2005). Dense ritual deposits had been discovered in nonresidential spaces, such as those found in the 1940s at Pacheco and Conchopata (Cook, 1987), but the idea that thick ritual deposits might be left on house floors upon abandonment was less plausible at the time than expedient discard behavior within a densely occupied urban setting.

1.3. Ritual Deposits in the Wari Empire

Excavations of ritual depositions associated with house floors can be difficult to identify. The most notable and common feature is the presence of smashed vessels. These are often large, elaborately decorated, and can be reconstructed to a great degree. Ceramic vessels were an important form of ritual offering (Cook, 1987, 2001; Isbell, 2000; Isbell & Cook, 1987). The destruction of these vessels removes them from use and takes them out of the sphere of exchange; they can no longer be displayed as prestige goods during salient gatherings. Also, when these objects are smashed in a structure, it removes that space from quotidian and/or ritual use. That room in a house or temple is transformed into a locale of religious observance and ritual memory (Groleau, 2009; Isbell & Groleau, 2010). Deposits are recognized as offerings primarily because the artifacts, oversized decorated face-neck jars and urns, had already been associated with the Wari offering tradition. Unfortunately, the less elaborate vessel types are mentioned as present, but inventories of specific offering deposits are not typically outlined; the size of elaborate deposits, the density of pottery fragments, and the need to reconstruct the vessels pose challenges to the comparison of ritual assemblages. The other elements of ritual deposition in these spaces have been overshadowed by a focus on elaborate vessels (see Cook, 1987; Glowacki, 2005, 2012; Isbell, 2000).

Wari ritual caches have been found and recognized throughout the archaeological investigation of this complex society (Glowacki, 2012). Most of the literature focuses on large, decorated vessels, but other amazing offerings have been recovered as well (e.g., Arriola Tuni & Tesar, 2011; Cook, 1992; Cuba & Amachi, 2022). Several scholars have described caches of broken oversized jars and elaborately decorated vessels found in buried pits or stone-lined chambers (Menzel, 1964, 1969, 1971; Ravines, 1968, 1977). This phenomenon has been examined most extensively by Cook (1985, 1987, 2001, 2012), who describes them as ritual tribute and the result of single depositional events. Cook (1987) also uses design similarities between coastal Wari burial tunics and those illustrated on oversized anthropomorphic face-neck jars to suggest that these depositional events may be linked to funerary ritual or ancestor cult activity (see also Cook, 2012, pp. 104–106, figures 75a–75f). Focusing on the imagery carried by the vessels in these offerings (Conchopata 1942, Conchopata 1977, and Pacheco 1942) she suggests that the iconography communicated and legitimized the new hierarchical social order as an archaic state emerged. This interpretation does not necessarily apply to residential contexts or to vessels that lack decoration; however, political interactions would have been important at several scales (Nash, 2002; Nash & Williams, 2009). The idea that such deposits were important to “politico-religious” (Cook, 1985) relations cannot be excluded even though the houses in question range from elaborate or palatial to more modest manifestations on “commoner” house floors.

Isbell (2000) has also discussed the smashed offerings of “cerámica gigante” or oversized vessels that include large face-neck jars or large urns along with other vessel types in buried pits and across floor surfaces at Conchopata. Cook (2001), interpreting these finds, suggests that ceramic offerings vary in context and purpose but can be interpreted as ritual tribute or payments (related to the ethnographic pagapu; see Chacaltana & Nash, 2009), possibly made to ensure well-being in some contexts and an important facet of maintaining positive relations with ancestors in others. Isbell (2000, p. 51) suggests that smashed vessels found in residential contexts possibly pertain to the belongings of a deceased member of the house and were broken in association with mortuary practices or were deposited over the actual tomb (e.g., Isbell & Groleau, 2010).

In either case, the ritual depositions encountered in Wari residential contexts may represent broad cosmological concepts people had about buildings and the lived environment (e.g., Groleau, 2009) or perhaps some ritual affinity between houses, temples, tombs, and the locations of other Wari ritual caches. Important insights can be gained from examining ritual deposits from different contexts; however, in this paper, I describe examples of ritual deposits found in residential spaces and the importance of distinguishing between materials resulting from ritual versus quotidian activities.

2. Materials and Methods

This analysis is based on practice theory (Bourdieu, 1972/1977) because it focuses on the material elements of patterned dispositions possessed by a group of people conducting activities according to a shared, but negotiable, view of how ritual depositions should be performed as well as ideas about when and where these were deemed appropriate. Repeated and patterned activities, in this case, ritual events, correspond to the performances of people who shared ideas about ritual practice and its meaning (or were willing to participate and observe). In the same manner, variation in ritual practice may suggest different ideas about rituals or that alternate rituals were performed under different circumstances. The idea that ritual depositions were required during construction and when a dwelling was abandoned distinguishes a group of people who shared certain cosmological views, while variations within these kinds of rituals may help identify different groups.

Since houses are very large artifacts, dwelling form and the organization of activities in houses result from people’s dispositions toward residential space and architecture. Practice theory can be materialized in archaeology by examining dispositions and choices in tool manufacture, style, and use (Lechtman, 1977, 1993, 1999; McGaw, 1996; Wright, 1996). The knowledge behind the use of an object is just as important as its manufacture in this respect. Studies of the “chaine opertoire” can include houses and other buildings, and I would argue it can be extended to any activity, especially rituals. Practices can be broken down into steps and compared to identify common patterns or highlight variations between contexts. According to Lemmonier (1992, p. 112), the raw material choices, the form and manufacture of tools, and the organization of gestures necessary to carry out an activity are “the way members of a social group perceive and organize in their minds the phenomena around them, be they social relations or the material world”. He also suggests that “[t]echnologies in one society may also be related [to one another] because they share the same actors, the same places, the same artifacts, the same materials, the same sequence of gestures, or the same technological processes” (Lemmonier, 1992, p. 8). In the Andes, a good example may be the pouring of libations. This practice is part of many diverse events, but the tools and gestures do not vary a great deal when performed within a community by the same practitioner. This means that an examination of ritual depositions in residential contexts may shed light on many overlapping and related activities.

As Janusek (2004, p. 5) explained, “social systems and structures consist solely of human relations and activities”. Applying a practice perspective and approaching domestic ritual deposits as activity areas, we can study the material remains as choices or techniques that represent ideas. We can also expand our inquiry and examine the productive activities and preparation preceding the social relations and the deposition of objects together in the ritual cache. While Dobres (1999) asserts that choices in technological schemes can represent individual identities, the present examination focuses on the co-residential group (see Nash, 2009) and the shared experience of participating in a ritual.

My analysis approaches ritual abandonment of structures as a set of practices that materializes concepts attributed to buildings and architectural features, cosmology relating people to their world, and the meaning of objects and their appropriate deposition. These rituals were likely personal and occasional rather than cyclical or annual. This may mean they were subject to idiosyncratic variations; however, as a type of broadly practiced ritual, there should be recurrent practices, which represent widely shared beliefs that we can observe in several Wari-affiliated settlements. Highlighting these commonalities will help us understand archaeological deposits and some of the general ritual practices that spread during the first millennium (see Nash, 2017).

Excavating Residential Contexts

Because of extensive looting, poor preservation at many highland sites, and the complexity of urban deposits, it can be difficult to determine the nature of items found on house floors. Were they valued goods placed in a patterned way during ritual abandonment activity or post-occupation garbage? Excavation methodologies are important. Often the first phase of archaeological research prioritizes stratigraphy, dating, and culture change writ large. Early research at Wari and affiliated sites only opened small windows to the past. This is not a criticism but the reality of how many research programs develop. Testing before launching large-scale horizontal excavations is a wise move and promotes a nuanced research design. At the same time, small excavations are insufficient for the next phase of research. It is important to follow telephone booths with large-scale exposures, which are necessary to flesh out patterned practices and understand processes of change through time, especially in domestic contexts.

As materials found within the domestic frame, if ceramics left broken on house floors are refuse, these items are secondary deposits representing a palimpsest of activities, a combination of detritus from several productive acts, and if abandoned houses become dumpsters, materials might come from multiple neighboring households. As garbage, these materials are inventoried, counted, identified, and used for comparative statistics in socio-economic schemes of archaeological inference, such as estimates of social difference or the prevalence of long-distance exchanges.

In contrast, as ritual offerings or the material representation of a single salient ritual deposition, the items are in primary context. This shifts the interpretation dramatically. The objects are all related to each other, were purposely deposited in a specific way, and were selected by an individual or group of people participating in an event. As a ritual deposit, these items are related to each other and can be analyzed as an assemblage. They represent activity areas; careful excavation, collecting materials according to a grid, and piece plotting diagnostics may reveal organized arrangements across the space, which can add to the interpretation of the context. An analysis of several such events may reveal common patterns and broadly shared practices as well as distinctions between events and/or groups of people within the empire.

Many structures were built in conjunction with ritual depositions (Chacaltana & Nash, 2009; Cook, 2001), and the closure of many structures also required a ritual event (Nash & deFrance, 2019). Ritual depositions and the burning of offerings are reported for many Andean societies in the past and present, but I suggest that their prevalence in Wari-affiliated households requires attention because of their potential impact on interpretations of Wari residential activity. In short, if materials gathered for a special ritual deposit are instead categorized as residential garbage, activities typical to the domestic sphere may be misconstrued. The realization that these materials are not domestic trash requires that Wari “elite” domestic activity be reconsidered and/or current interpretations be bolstered by additional lines of evidence.

Research at Cerro Mejía was initially carried out to understand the essential elements of domestic activity by examining small, humble dwellings along with larger, elaborate residences. The goal was to test the hypothesis that elite houses, in a manner similar to that proposed for the patio groups in the Moraduchayoq compound, were used for Wari administration, defined in part through feasting. The inclusion of more modest houses in the sample would facilitate the identification of the material markers of political relations and administrative practices in larger, more complex houses as the specialized activities of elite officials. Yet, research on Cerro Mejía has demonstrated that most houses incorporate ritual activity and that the characteristics taken as signs of feasting are not limited to structures with the Wari patio group form. The percentages of pottery taken alone do not seem to be an accurate indicator of feasting as a regular activity.

When I initiated work at Cerro Mejía, I expected to find evidence of feasting in house compounds with Wari patio group components and feasting to be absent in the other houses at the site. Instead, I found the situation was far more complex because many houses contain what is construed as the material markers of feasting; they had relatively high percentages of bowls or other vessel forms associated with serving and consuming meals relative to vessels dedicated to cooking or storage. I also quantified the faunal remains as the remnants of festive meals. If instead these remains are offerings associated with the final closure rites performed by the family as they abandoned their dwelling, they may have exceeded quotidian norms for the household during its occupation. Access to the resources included in such offerings may have resulted from the accumulation of political capital over time rather than the standard of living maintained by the residents of the house or sponsorship by a patron to carry out essential rites. This possibility raises two important questions: (1) Can fragments of consumption vessels be used to infer political feasting in all instances, or can these same material remains be components of salient ritual offerings? (2) Can evidence of ritual feasting associated with the closure rites of a residential structure serve as the basis for interpreting Wari administrative practices in the absence of facilities to generate festive meals or host guests on a regular basis?

I suggest that evidence from final closure rites can be used to understand many things about the group that created the ritual assemblage. The remains of an important offering event may be a good indication of the relative prestige of the family group and represent the socio-political connections of the sponsoring kin group to the Wari state. At the same time, materials found in these caches may not have been a part of domestic life, nor may feasting have been a regular part of domestic settings. Some goods may have been obtained specifically for inclusion in the offering; patterns inferred from ritual deposits may not validly represent quotidian modes of resource consumption.

I describe features and deposits from houses excavated from Cerro Baúl and Cerro Mejía. These sites are part of an intrusive Wari-sponsored colony established early in the seventh century in the department of Moquegua, Peru (Figure 4). Occupations on these two adjacent hills were connected by an elaborate irrigation system, and they probably are sectors of the same low-density urban center, but construction of a modern, paved road between the hills dictates that they be treated as separate archaeological sites (see Nash, 2024). Ritual offerings of some form have been identified in all the residences analyzed by the author thus far (n = 14). The sample cross-cuts categories of social difference as manifested by the size and elaboration of the residential structures.

The largest is a palace in Sector A on the mesa summit of Cerro Baúl with an area of more than two thousand square meters. I have excavated approximately 30% of this compound (Figure 3) and uncovered an entry court (Un 25), a ceramic workshop (Un 40A), a garden (Un 40C), other craft areas (Un 41B, 41C, and 41E), and a Wari patio group (shaded, Nash, 2012a, 2017; Nash & deFrance, 2019), which had a stone-paved floor. The thirteen other houses are from the slopes and summit of Cerro Mejía, the smallest of which is less than 27 m². This study is based on comparative analysis and relies on horizontal excavations and detailed documentation. Materials were collected to examine activity areas; the methods employed were essential for identifying ritual deposits versus quotidian domestic features.

During excavations, a one-by-one-meter grid was established to control the collection of materials. Items such as diagnostic vessel fragments, decorated sherds, points, marine shell, stone tools, semi-precious stone, etc. were left on floor surfaces when possible and piece plotted on drawings, including their elevation relative to the unit datum (taken from the lowest point below the object, not its top). Overhead photographs were taken using a bipod constructed from PVC pipes with a camera attached at its apex. Mosaic images of floors were pieced together in Adobe Photoshop. Features were defined in the field and excavated separately. These were photographed, drawn (both in plan and profile), and collected by strata when relevant. Sediment samples for fine screening were collected from each meter square as well as features. Many ash features were collected in their entirety. These samples often contained micro-debitage in addition to charred organics. All sediment was passed through a window screen, which has an average aperture of 1.8 mm, in order to retrieve fish vertebrae from small species, such as sardines, and the remains of guinea pigs, recover obsidian retouch, and collect all the debitage associated with lapidary work. Smaller sediment samples were also taken from the center of each square meter for elemental analysis; however, that information does not contribute to this study.

3. Results

In general terms, but largely based on the more elaborate deposits, there is a specific stratigraphy associated with household abandonment. I describe the deposits in reverse order as they would have accumulated through practice. The floor is first. Several spaces in the palace had stone-paved floors, which made it easy to identify and note the presence of purposeful breaks that were created for offerings by removing stones. These pit offerings are described further below. In several places, remnants of white plaster adhered to the stones. White plaster was also applied to floors made of a compacted, clay-containing mixture, such as some rooms in the brewery complex (Williams et al., 2001, p. 83, figure 7). Identifying use surfaces in most residences was more challenging. There is little difference between the material used to coat the walls and the tamped substance constituting floors. One must often look for ash patches near hearths, flat-lying objects, and small lithic debitage or a “dirty” appearance to discern occupation surfaces. Carbon and bone are often moved by roots in areas with vegetation; they are not reliable indicators of a floor surface. Residential floors typically have small, embedded debris, even between paving stones, but often also have tools and other items left on their surface. These materials would not include all items once used by the household but may be sufficient to infer some quotidian activities; small debris, when present, can substantiate hypotheses based on the morphology of larger tools. Based on current data, plazas and patios with benches have relatively little de facto refuse (sensu Schiffer, 1972). These areas typically have denser ritual deposits. In contrast, kitchens and other work areas have larger tools and craft debris left in place; offerings are limited to doorways or placed in hearths.

The first phase of ritual deposition includes pit offerings and burnt offerings. These depositions may have been made before or contemporary with feasting activity, which is observed in the palace but is not apparent in all contexts. Some ceramic vessels may have been broken early in the ritual process because fragments of vessels found in some floor deposits have been recovered from burnt offerings or pit features, which often were targets for the subsequent deposition of smashed vessels. To be clear, I am not saying that pieces of the same vessel somehow made their way into deeper strata but rather that most pieces of a vessel were deposited in one room while a few pieces were placed in a pit or burnt offering in another (see Nash & deFrance, 2019). These acts of ritual distribution were also detected during the study of materials from Cerro Baúl’s brewery and may materialize salient cosmological concepts and the value of some types of pottery beyond the storage, preparation, or consumption of foods and beverages (Nash, 2019). Finally, vessels were broken, and fragments of these vessels were arranged with other objects in meaningful ways throughout the structure. Piece plotting and vessel reconstruction indicate that some pottery deposits were thick; some formed piles that were on top of or in a higher stratum than the other forms of ritual deposits. Some structures were set ablaze and exhibit extensive burning of entire rooms (i.e., the brewery, Nash, 2011); however, most of the houses in the sample have localized burnt offerings. In the following section, I describe examples of ritual deposits found in the palace on Cerro Baúl and the more modest residences on Cerro Mejía.

3.1. Pit Offerings

At the Wari capital, Conchopata, and other Wari-affiliated sites, caches of large, smashed jars have been found in basin-shaped pits. In residential contexts, the pits used to deposit offerings are smaller, relatively shallow, and created by breaking the floor surface. The examples described from Conchopata (Groleau, 2009; Isbell & Groleau, 2010) included items such as molds used in ceramic production, stone tools, blue stone, camelid remains, and ceramic faces broken off of large jars (described below). Likewise, the examples identified in the palace on Cerro Baúl sometimes included a decorated sherd that could not be associated with a reconstructable vessel from the complex’s assemblage; two included fragments of faces but of a different kind than those found at Conchopata (see Figure 5).

Twenty-one pit offerings of this type were identified in the excavated areas of the Sector A palace. They were common in the gathering spaces with benches, the entrance court (Un 25) and central patio of the palace residence (Unit 9B). Four of the five rooms surrounding the patio also had pit features (9A, 9C, 9E, and 9F), but 9G, which visitors passed through to enter the central residence, had an intact, paved floor without any stones removed. The platform, 9D, was somewhat eroded, which makes definitive identifications of these types of features questionable.

Table 1. Contents of pit offerings.

Feature	Decorated Sherd 1	Face Frag	Frag of Distributed Vessel	Tool for Pottery Production	Grindstone	Point	Chrysocolla or Sodalite	Camelid Bone	Rodent 2 Bone	Fish Bone	Seashell 3
9A, R7			x
9B, R4								x
9B, R5								x		x	x
9B, R7			x		x			x		x
9B, R8 4	x		x		x	x	x	x	x		x
9B, R14								x		x
9B, R15	x	x	x					x		x	x
9C, R8				x	x	x	x	x	x
9C, R9						x		x	x
9E, R3						x	x
9E, R4								x
9F, R3					x			x	x
9F, R4								x	x		x
9F, R5B	x		x		x		x	x	x		x
25, R3						x		x	x	x	x
25, R4								x		x	x
25, R5	x		x					x		x	x
25, R7	x	x						x		x	x
25, R9	x		x								x
25, R18	x				x						x
25, R19											x
41B, R17							x	x			x
41B, R21							x				x

¹ This is probably an undercount, as many sherds from decorated vessels are not decorated. ² This includes general rodents and elements identified as guinea pigs (cuy, Cavia sp.). ³ Whole shells, shell beads, and shell fragments, including Spondylus sp., are combined in this column. ⁴ Technically, this may not be a pit feature, as it was a deposit found where a vessel had been sunken into the bench of the central patio. It is included here due to the similarities of its assemblage with other pit features.

indicates some of the common materials present, which include camelid, guinea pig (cuy) or fish remains, seashell, stone tools, and semi-precious stones. All contained vessel fragments. At least seven of the pit features included fragments of vessels that had been ritually distributed throughout the palace. There could be more, but the quantity of fragments in the palace assemblage dictated a focus on decorated vessels, rims, and a room-by-room approach. A few pit features had unique items that are worth mentioning. For example, 9A-R7 had the remains of an unfired vessel, 9E-R4 had the polished pincer (chela) of a crustacean (probably a river shrimp), 25-R3 and 25-R4 each had a few human elements, and 25-R18 had a fragment of a Tiwanaku incense burner.

Pit features are more difficult to identify on Cerro Mejía because of the more rustic nature of the prepared floors and the prevalence of vegetation, which creates soft pockets where roots were once present at some point in the past. Floor preservation is patchy in most spaces. Since many pit features were later covered with smashed pottery and other objects clustered on the floor around them, it is likely that the contents of pit features, if present, were collected along with the elements of floor offerings instead. If that is the case, however, their contents would have been different, as no guinea pig or fish remains have been recovered from houses on Cerro Mejía, and seashell frags are exceedingly rare and small. No whole shells have been found.

3.2. Burnt Offerings

Burnt offering deposits are the result of in situ burning. Ash features resulting from ritual blazes exhibit patterned color distribution with a dark center radiating out to lighter shades of gray. Materials within the matrix correspond to this pattern and show differential charring based on their location in the ritual deposit, which means some items may exhibit partial burning or none at all. This pattern overlaps to some degree with hearths, although many burnt offerings exhibit evidence of higher temperature combustion than most hearths. Bone may be white or blue-gray. Features with light gray to nearly white ash may have little to no botanicals remaining.

Excavation is usually required to define the nature of an ash feature because domestic hearths are informal at both sites (Figure 6E,F). Formal hearths are associated with the brewing of chicha. These have been found in the brewery on Cerro Baúl, and four were located in a lateral room of the patio group in Unit 145 on the summit of Cerro Mejía (see Figure 6C,D). They have three to four stones oriented vertically to support large jars with conical or truncated conical bases (Nash, 2011). Two small rectangular hearths framed with stone slabs have been found on the summit of Cerro Baúl in different contexts and appear to be associated with craft activity rather than food preparation.

Since domestic hearths are informal, they can only be identified from their profiles after excavation. Hearths have stratified ash layers that overlie heat-altered sediment, which would appear as a darker red or orange on the clay floor. In situ burnt offerings can show slight basal heat alteration in the underlying matrix; however, it is less profound, and these features exhibit more uniformity in the color matrix. They have no strata. In most instances, hearths are basin-shaped, whereas burnt offerings overlie the relatively flat floor surface.

Ash features can also correspond to refuse, which might be encountered as a pile on the floor, presumed to be in a temporary place of discard or in a permanent locale such as abandoned rooms. On the summit of Cerro Mejía, refuse has been noted along exterior compound walls and is visible on the surface as concentrations of ceramic sherds. Patterns of garbage disposal on the slopes of Mejía remain unclear. Excavations on Cerro Baúl have uncovered two stratified sheet middens, one of which also included intrusive pits that were characterized by stratified waste deposition (see Figure 6G). Middens and garbage pits include ash mixed with burnt and unburnt materials. The matrix of refuse deposits is patchy in color because dust and fine silty sediment, probably derived from floor sweeping, is mixed with ash, hearth contents (somewhat burnt), and other debris (materials swept from the floor unburnt); decaying food waste or other kinds of organic material, such as chicha dregs (Nash et al., 2024), can appear as dark to light brown pockets among the ashy deposits. See Figure 6 for examples and

Table 2. Types of ash features encountered in residential contexts.

In situ burnt offerings	Ash exhibits some uniformity in color patterning. These features often have dark spots with radiating shades of gray. Material inclusions follow the same pattern of burning as the surrounding matrix. Some may exhibit high-temperature combustion, which is not conducive to successful cooking, and have white ash, light gray ash, white to blue bone, and no botanicals. The underlying stratum should show slight signs of heat alteration, but it was not always detectable. These deposits are typically found on flat floor or bench surfaces. Additional objects may be added after burning.
Formal hearths	Stratified ash deposits associated with vertical support stones. Associated materials exhibit differential exposure to heat. These features are basin-shaped in profile and demonstrate obvious and intense heat alteration to the underlying and surrounding floor matrix.
Informal hearths	Stratified ash deposits with significant heat alteration to the underlying and surrounding matrix. Typically, they are basin-shaped in profile. Associated materials show differential heat exposure. This may be due to ritual activity at the time of abandonment.
Sheet middens, garbage pits, and temporary discard	Ash and other sediment mixed with burnt and unburnt objects. Pockets of decaying organic material range from yellow to brown. Swept floor sediment, organics, and ash combine to create a patchy matrix. The underlying stratum does not exhibit heat alteration. These features may form stratified deposits on flat floor surfaces or be deposited in pits. Either may exhibit a “basket dump” pattern in profile. Discard from a single temporary event is rare.

for a summary of the different types of ash features.

Many burnt offerings were found in Baúl’s Sector A palace. As most ash features were collected in their entirety as float samples, the contents of those selected for analysis provide the best picture of their organic contents. Unfortunately, only one of the palace’s burnt offerings was examined. A more comprehensive analysis conducted by Victor Vasquez of float samples from Cerro Mejía showed that some samples had only a few pieces of charred wood (Nash, 2022b). Burnt offerings differed in size and the degree of burning involved; many had no identifiable faunal remains or visible charred botanicals. Only a few hearths and burnt offerings offer information about the plants processed and consumed within houses.

The float sample analyzed from a burnt offering in the palace pertains to 9A-R5 (feature 5 in room 9A). It contained both maize and molle (Biwer, 2019), which might be considered valuable as ingredients of chicha (Nash et al., 2024). Faunal remains in the feature included fish and rodents (Nash & deFrance, 2019). Based on stratigraphy, it seems that the remains of a young child, probably as a secondary bundle, were placed on the burnt offering after it was extinguished (none of the bones exhibit heat alteration), followed by broken pottery, including fragments of vessels that were divided between two piles in this room (this is described in more detail below).

In the northeast corner of the central patio, there was a large burnt offering, 9B-R9 (see Figure 6A). This ash feature was ephemeral and overlaid the stone-paved floor surface. Two smaller ash features were located just to the south and west of the larger feature. These burnt offerings seem to frame a small, protruding chunk of bedrock, which may have been a focus of ritual during the life of the palace before its abandonment (see Nash, n.d.). Faunal materials include vizcacha, guinea pig, and camelid (Nash & deFrance, 2019). Among the other items was a rustic point made from white chert and sherds from two different vessels, whose fragments were found in other parts of the palace. Interestingly, exploration of the subfloor layers in this area exposed earlier burnt offerings as well. The lowest of these, associated with Floor 3, contained molle, maize, and bottle gourd seeds (Biwer, 2019); however, there is no way to know if this lower level pertained to a domestic context or some other type of space.

Small burnt offerings on Cerro Baúl were associated with clusters of objects and may have been components of complex acts of ritual offering. In 9C, three different burnt offerings were associated with portions of decorated bowls. In rooms 9A and 9C, burnt offerings were situated near doorways and would have obscured these entrances. A small patch of blue-gray ash in the southwest corner of the patio and another in room 9C appear from their texture to be some type of burnt textile. It is possible that small burnt offerings represent the destruction by fire of valued goods. Their intermingling with broken pottery and other preciosities does not support an interpretation of temporary garbage disposal. Burnt offerings were associated with other goods on Cerro Mejía but were misinterpreted as garbage features (Nash, 2002) because they were excavated before the palace on Cerro Baúl and their contents were relatively modest in nature.

The clearest example was encountered in a very rectilinear, well-built house on Cerro Mejía’s slopes, Unit 6. The excavated spaces were incredibly clean compared to other dwellings, and the objects left behind were of a relatively high value rather than abandoned, heavy stone tools. In room A, there was a hearth and an ash feature. Divided between these two deposits were the fragments of a modeled vessel with the face of a woman on its neck, who had incised tear bands filled with red pigment (Figure 8A). The ash feature had two points made from white chert, and the hearth had an additional white point as well as three metal tubes made from rolled lamina. In this case and in several others on Cerro Mejía, domestic hearths appear to be a location for offerings. As with this example, some items were not burnt and appear to have been placed on top of hearths. Some houses had relatively large quantities of camelid bone deposited in hearths (Nash, 2012a, p. 22), which may be indicative of feasting (Nash, 2010). These bones could be the remains of the final meal prior to abandonment. They were found in the hearths with other valued goods such as points, beads, or decorated sherds. The bones and other materials have little evidence of burning, which is consistent with boiling or roasting while meat or sinew is present to protect the bone. Hearths stuffed with bones and other items would not have functioned for cooking; these materials would have been placed in or near the hearth after the meal was prepared and consumed as a ritualized deposit at the time of house abandonment. This replicates the way burnt offerings were covered with piles of pottery or associated with clustered objects on Cerro Baúl. The use of hearths for ritual depositions of pottery and camelid bone was also found by Glowacki (2005, p. 109) at Pikillacta. Items associated with burnt offerings are overshadowed by the sheer quantity of pottery fragments, especially when they are from decorated serving or drinking vessels.

3.3. Vessels and Sherds

Analysis of the ceramic assemblage determined that there were more than 300 vessels or large parts of vessels placed as offerings in the excavated areas of the palace. This minimum number of vessels is based on the presence of more than 50% of the vessel’s rim, a unique paste, or a unique form. Most of the vessels exhibit evidence of use, and some were well worn (see Figure 5D). The mix of production styles suggests that the pots were accumulated over the life of the palace; however, some may have been introduced specifically for the abandonment ritual. A few may have been placed as acts of remembrance, especially in the garden of the palace, where ancestral remains were located (Nash, n.d.).

Vessels were placed throughout the structure but in different ways. In the two areas that had benches and seemed to be designed for socio-political gatherings (Un 25 and Unit 9), vessels were found in clusters throughout the space. In other areas, such as the ceramic workshop, vessel offerings were clustered in doorways rather than throughout the room. Descriptions of ceramic vessels as offerings in the Wari literature have focused on those that were purposely fragmented; however, in the palace on Cerro Baúl, there are some indications that a few vessels were carefully placed intact rather than being shattered. This is most apparent in the patio of Unit 9, which contained the largest number of vessels (n = 56). Most of these were found broken on the patio floor. Vessels recovered from the bench appear deposited in a patterned way, with one vessel located in the center of the bench along all four sides of the patio. To the north and west, these were jar forms; to the south and east, these were relatively intact bowls. Another intact bowl was also placed near the center of the patio floor.

Spatial analysis of the distribution of ceramic fragments throughout the structure reveals some other interesting patterns (for a plot of selected vessels, see Nash & deFrance, 2019, figure 10). For instance, in Room A, there were twenty-one smashed vessels. This included a rayed head tumbler (an oversized kero), a huaco retrato (portrait vessel cup), seven bowls, and twelve jars of different forms. Pieces of a fine, everted bowl (also called an escudilla by some researchers) and the rayed head kero were found in two clusters, one north and the other south of the door. The former grouping partially covered the burnt offering 9A-R5 described above. Pieces from these vessels and four others were in both clusters and seem to have been purposely divided. This sort of deposition has also been noted at Pariti, where pieces of elaborate vessels were deposited in two different pits. The ritual deposit from Pariti pertains to Tiwanaku, but Wari vessels are among the assemblage (Villanueva & Korpisaari, 2013); we cannot rule out that some types of deposition were shared among the elite members of interacting polities (e.g., Nash, 2015). The ritual division of vessel fragments was replicated at a smaller scale on Cerro Mejía, as mentioned above, in Unit 6.

Vessel fragments were purposely placed in other ways. Sherds from several vessels were found in different rooms. The huaco retrato found in Room A was complete, but pieces of an identical vessel were found in the doorway to Room G and the partially excavated plaza 41E to the west (for a discussion of ritual sets see Williams & Nash, 2021). Another interesting example is VR 102 (see Figure 7A). Most of the face was found in the eastern room (Un 9F), whereas the headdress (the rim) and many fragments of the large globular body were found smashed throughout the patio (Un 9B). Also, pieces of two decorated, thin-walled bowls were recovered from Room A, Room F, the doorway to Room G, and the plaza (Un 41E). This set of incurving bowls has hooks that face in opposite directions but are identical in all other ways (see Figure 7B,C). They are still quite fragmentary, and it is likely that pieces were purposely placed in other rooms we have yet to excavate. The practice of placing pieces of decorated vessels in different rooms was also part of the ritual abandonment of Cerro Baúl’s brewery.

There is also evidence to suggest that vessels cannot be completely reconstructed because people purposely saved ceramic fragments, especially those with decoration, as mementos of the occasion or for some other purpose. Groleau (2009) describes evidence of this practice at Conchopata. She noticed that the face lugs, which are the only form of decoration for some types of jars, were often missing from reconstructed vessels and seem to have been purposely removed. In addition, she reports that some ritual deposits included these face lugs along with other items of value. A variation of these practices may have been played out in the palace on Cerro Baúl as well. In Room C, a rayed head tumbler is missing its face; likewise, several offerings in this and other rooms appear to incorporate decorated sherds from vessels that are not present. Since only 30% of the compound has been excavated, it is quite possible that the other parts of these vessels will be found in other rooms; however, the evidence from houses on Cerro Mejía, which were excavated in their entirety, suggests that some decorated sherds were used as offerings in the absence of the entire vessel. In most instances, less than 30% of a vessel is present. In modest houses a single sherd was found (see Figure 8). These sherds cannot be considered evidence that the complete bowl, cup, or decorated serving jar was ever present in the house or used during regular feasting events.

Larger portions or nearly complete broken vessels did appear in a few houses on Cerro Mejía. For instance, there was a rustic face-neck jar smashed in Unit 17 (Figure 8C) and another, which appears to represent a woman, mentioned above, found in Unit 6 (Figure 8A). In Unit 4, a miniature vessel had been placed on a bench in the patio (Figure 8I). The vessel was relatively intact and may have contained a liquid offering of some kind. It was found with the base of a cup, but the cup was incomplete. For the most part, other smashed vessels were bowls. Storage jars are sometimes used as suitcases or moving boxes to carry useful goods from one house to another; serviceable vessels of many kinds may have been retained rather than destroyed by houses of modest means. Baskets and textiles are not preserved at the site but are alternative types of “storage vessels”. The percentage of vessels present in a residential context may not be the best measure of feasting in the domestic sphere.

4. Discussion

The central patio in Unit 9 had bones of camelid and ocean fish on the floor and bench in a manner suggesting they may be elements of a discarded meal (Nash & deFrance, 2019). If this is the case, the abandonment of the structure would have involved one or more elaborate meals. The high percentage of bowls and serving wares supports this hypothesis. What remains uncertain is whether feasting was a regular occurrence in this setting. Other lines of evidence, such as preparation facilities, are needed to verify that feasting was part of a house’s “use-life” (Nash, 2011, 2012b). In the case of the palace on Cerro Baúl, no kitchen was uncovered that could have created a festive meal. If feasts were a regular activity, such a facility should be present. Yet, this facility may remain covered, and there are two venues in the palace well-suited for guests to gather and share a festive meal (e.g., Unit 9B and Unit 25). Likewise, the larger houses on the summit of Cerro Mejía had patio spaces to accommodate guests beyond those living in the compound. Unit 145 had a central trapezoidal patio and a room dedicated to making chicha and large meals (Nash, 2010). Ironically, few ceramic vessels were recovered from this structure. In the absence of the different hearth features, the compound would not meet the typical metric used (i.e., the number and relative percentages of serving vs. domestic vessel forms) to identify feasting and discuss related Wari administration.

In a similar vein, faunal remains like those recovered from the palace on Cerro Baúl demonstrate differential access to rare foods, such as ocean fish, river shrimp, and guinea pig, which were not found in the houses excavated on Cerro Mejía (deFrance, 2014). This distinction is also important when considering what constitutes a festive meal. It is quite possible that the quantity of meat served, the cuts of meat prepared, or its mere presence (versus boiled bones to flavor broth) in some modest households on Cerro Mejía would qualify as a feast to those in attendance. Yet, would it be appropriate to connect such feasts to the government hierarchy or acts of state administration? I do not offer an answer to this important question; more data are needed. I would, however, underscore that these meaningful comparisons cannot be made or contemplated without more complete household assemblages, careful excavation, and detailed documentation of depositional patterning. The formation processes at many Wari-affiliated sites, which are relatively clean and do not often include the practice of disposing of domestic detritus in abandoned houses, offer an opportunity to examine ritual assemblages and gather data to test current interpretations that link feasting to “administration” as an institutionalized manner of governance in this early state.

5. Conclusions

Some ritual performances may cross-cut social differences. Family groups experience many life crisis moments that require special patterns of consumption, where rituals mark important changes in the household. Examples of these are births, deaths, weddings, and perhaps public or ritual service in community affairs. How these rituals are elaborated depends on the resources available to the kingroup, but if modern practices are any indication, even the most impoverished households devote substantial resources to ritual activity (e.g., Bolin, 1998; B. Isbell, 1978).

A change of residence, the abandonment of one dwelling, and the occupation of a different location is an important transition that may coincide with a significant reorganization of the corporate group. Thus, we may expect broken libation vessels, food offerings, or the ritual deposition of whole tools or other objects to coincide with all house abandonments. In some cases, these deposits may correspond to mortuary rituals (see Toohey et al., 2016).

Regardless of the motivation of ritual deposits, this analysis shows that the components of abandonment ritual overlap with the material correlates of feasting activity. High percentages of consumption wares have been prominently linked to Wari administrative activity. In this paper, I have suggested that residential assemblages are complex in nature. In many instances, house floors may be covered with “ritual deposits” rather than heaped upon with “refuse”; in this case, the unstructured discard of feasting remains. To distinguish between these possibilities, and potentially others in the diverse, multi-regional Wari Empire, investigations of residential areas require horizontal exposures and excavation strategies geared toward recovering artifacts as components of activity areas as well as detailed contextual and comparative analyses to interpret the full gamut of residential activity and to correctly interpret the materials recovered. I would submit that analyses of artifacts from small tests (i.e., 1 × 1 m or 2 × 2 m) are usually insufficient to identify activities, understand formation processes, accurately identify features (e.g., large ritual deposits versus refuse dump), and link materials recovered to aspects of social life such as discard, ritual, or feasting.

I am not suggesting that feasting did not occur in Wari residential patio groups or Wari-affiliated houses with other forms. Rather, I would like to suggest that deposits of smashed vessels may be offerings and activities specific to the abandonment of the structure rather than refuse that can be linked to the regular, routine practices that took place in the household. This may have implications for our understanding of “administrative activities” and where they took place. I would propose that other activities played an essential role and that Wari statecraft was more complex and entailed diverse activities and multiple, formal institutions to sustain control relations over such a large area for more than three hundred years. As I have discussed elsewhere, feasting requires preparation facilities for chicha and large meals (Nash, 2011; Nash et al., 2024). Residential structures that lack these features may have the right goods upon abandonment; however, if they lack the appropriate facilities to create this type of refuse on a regular basis, we may need to re-evaluate the complexity of household deposits and our understanding of how feasting played a role in the Wari political economy.