Hearths, Grinding Stones, and Households: Rethinking Domestic Economy in the Andes



This case study of Andean house remains in central Peru adopts a gendered perspective and, on the basis of archaeological evidence, challenges the normalized domestic model of “lo Andino.” Although most Andean archaeological studies equate economically independent households with independent structures, this analysis reexamines the accuracy of this convention. The presence of a grinding stone, an essential tool of an economically independent unit, varies among house structures, which suggests that multiple houses shared labor on a daily basis.

Ethnohistoric records from the Spanish Conquest, as well as ethnographic research on traditional Andean lifeways, have generated the concept of “lo Andino,” which has often guided archaeological interpretations of the prehistoric Andes (Van Buren 1996). Part of this concept relies on an agrarian model that characterizes the coresidential household as the “basic economic unit” on the basis of the agricultural production that occurs outside of the house structure and is typically associated with men; communal work (i.e., supra-household production) within the ayllu, or community kin group, is only performed at specific times according to the agricultural cycle (yanantin; see Urton 1997). This model, however, ignores the part of the domestic economy that occurs within the house, a sphere of activity most commonly associated with women. In accepting these basic tenets, scholars have failed to question the implicit economic order, social relationships, and gender roles that this model assumes. This case study of Andean household remains in central Peru draws on a gendered perspective of archaeological evidence to challenge the normalized domestic model of “lo Andino.” In the normative “lo Andino” model, men and women living in single coresidential units perform gender-divided, noncollaborative domestic tasks on a daily basis. This analysis shifts the focus of the domestic economy and its organization from the production-associated agrarian “male” component to the multifaceted site of the house and, in doing so, redefines the social dynamics of household production.

By starting from the archaeological record and the artifacts of daily living, we can develop a more accurate understanding of everyday life and the domestic economy in the past. This chapter examines household excavations from six sites in the Andes and identifies the presence and absence of two features, hearths and grinding stones, that are essential elements of the Andean domestic tool kit. The case study sites are located in the Upper Mantaro Valley in central Peru (Figure 3.1): the sites of Hatunmarca, Tunanmarca, and Umpamalca date to the Late Intermediate period Wanka II (1350–1450 c.e.) and Marca, Chucchus, and Huancas de la Cruz date to Wanka III (1450–1533 c.e.), during the Inka Empire (D'Altroy and Hastorf 2001; Earle et al. 1987).

Figure 3.1.

Mantaro River Valley with archaeological sites: 2, Hatunmarca; 7, Tunanmarca; 41, Umpamalca; 54, Marca; 74, Chucchus; 59, Huancas de la Cruz (redrawn from Earle et al. 1987:Figure 1).

Archaeologists have often used hearths as markers of domestic units because of their nonportable nature. Unlike some domestic artifacts, such as ceramic pots and stone tools, that can be transported upon abandonment of a domestic unit, evidence of a hearth is literally burned into floors or living surfaces and remains for archaeologists to uncover. Ethnographically, however, hearths do not always represent independent domestic units (Weismantel 1988).

Despite clear ethnographic evidence that grinding stones are fundamental to cooking technology in the Andes, they have largely not been considered essential to the domestic tool kit (Aldenderfer and Stanish 1993; Nash 2002; Owen 1993; Stanish 1992; for an exception see Russell 1988). This is because although grinding stones often appear in domestic contexts, their presence is variable in the archaeological record: a single grindstone is not always associated with a single hearth and single domestic unit. Some scholars suggest that the uneven presence of grinding stones is a result of abandonment or postabandonment processes, in which fleeing populations took their grinding stones with them or new populations moving into the region removed the stones for their own use (Nash 2002). Although these possibilities cannot be ruled out, it is important to realize that these occurrences were also probably variable and depended largely on local circumstances of tradition and practice. In this study, I examine distribution patterns of the passive portion of the grinding stone, the batán, a large, flat or concave stone, often weighing over 20 kilograms. Migrating populations abandoning a site had to consider the practicality of transporting these cumbersome, heavy stones; if the grinding stones were in fact transported, despite their weight, then this may demonstrate a cultural preference, and we would not expect to find many of these implements at abandoned sites. Postabandonment processes of grinding stone removal would also depend on local factors, such as the proximity of the abandoned site to more recent settlements. Although the removal and reuse of grinding stones can affect their distribution at a site, unless all grinding stones were removed, a general pattern of distribution should still be preserved.

I suggest that this patterning is significant and can be analyzed to understand domestic economic organization at a site. The material evidence calls for a reevaluation of our a priori assumptions of economic organization and gender roles. Given the distribution of grinding stones and hearths within sites, I suggest that on an everyday basis, the single coresidential unit was not an independent economic unit; instead, multiple coresidential units engaged in communal activities of a domestic nature. The presumed male–female economic unit typically adopted in Andean archaeology largely ignores a more complex and flexible domestic economy that incorporates men and women of different life stages into a variety of everyday production and consumption activities.

A Gendered Approach

Gendered anthropology, now considered just “good anthropology,” recognizes that people and the intersection of their sex, gender, life stage, social status, et cetera compose the social relationships of the past and present and play basic roles in the construction of broader social processes (e.g., Brumfiel 1991, 1992, 1996; Conkey and Spector 1984; Gero and Scattolin 2002; Hastorf 1991; Robin 2002). An engendered approach does not just identify women and their roles: it identifies people of different ages, genders, life stages, and so on. In this chapter, I suggest an engendered approach to understanding domestic production in the Andes. The gender-divided dynamic of men working in the field and women processing the harvest, typically adopted by archaeologists in the Andes, is linked to defining the basic economic unit, the household, as a single coresidential house, where tasks are gender divided. The archaeological evidence, however, demonstrates that these coresidential units do not have the complete tool set necessary for everyday production and consumption, which challenges assumptions regarding gender roles and division of labor. A new model for domestic economy in the Andes that takes into consideration the multiple aspects of everyday production and consumption is needed. Here, I start with the houses and the everyday implements and features associated with them; by looking at how these vary, I provide an alternative to the typical domestic economic model.

Household Archaeology and Definitions

Household archaeology is a burgeoning field that has already provided numerous insights into understanding social relationships, the domestic economy, gender roles, division of labor, and other areas of study. Wilk and Rathje (1982) first defined this approach in the early 1980s and it has taken a firm hold in Mesoamerica, with archaeologists in other regions rapidly following suit (e.g., Bawden 1982; D'Altroy and Hastorf 2001; Earle et al. 1987; Hendon 1996; Kent 1987; King 2006; Robin 2002; Sheets 1999; Stanish 1989; Wilk and Ashmore 1988). Ethnographers, anthropologists, and sociologists have wrestled with definitions of “household,”“kin group,”“domestic unit,” and the like for decades. Defining a unit that can be identified and analyzed across cultures and millennia is difficult because of the diversity of the past and present.

In archaeology, however, a need to compare households has prompted their definition in terms of function, not form or kinship: households are social groups that share in a definable number of activities and they can be identified empirically in archaeological samples after extended study (Ashmore and Wilk 1988:4). Members do not have to coreside (although they often do), and individuals can belong to more than one household. Because the definition is based on activities, rather than on physical components or familial relations, Ashmore and Wilk's concept of the household is cross-culturally applicable; the basic functions of a household hold firm regardless of time and space (Ashmore and Wilk 1988:4). The four general functions of a household are classified as production; distribution or consumption; transmission and transference of property, rights, and roles; and cultural reproduction via child socialization (Aldenderfer and Stanish 1993:6; Ashmore and Wilk 1988:4; Wilk and Rathje 1982:622–631). To define a household in the archaeological record, archaeologists have to identify these functions.

In the Andes, archaeologists have largely accepted the idea of the “household” as a basic economic unit (Bawden 1993; D'Altroy and Hastorf 2001; Earle et al. 1987; Nash 2002; Stanish 1992). The household has been employed as a way of identifying daily subsistence practices and the organization of everyday activities. In his study of vertical complementarity in the Andes, Stanish (1992) identifies the household as “co-residential work units.” Following Goody (1982), he asserts that agricultural work units tend to coreside, and the dwelling unit, reproductive unit, and economic unit are closely linked. According to Stanish, this “co-residential agriculture work and habitational unit” (Stanish 1992:38) should appear in the archaeological record as remains of domestic activities such as hearths, storage, sleeping areas, kitchen middens, and nonelite residential structures. In each “unit,” these activity patterns should repeat; thus, independent household units can be identified by duplicated areas of the domestic production and consumption activities. These material correlates of household functions, however, are based on ethnographic and ethnohistoric models that assume coresidence; this assumption narrows the scope of study to physical structures and nullifies any attempt to define “household” on the basis of activities.

Recent studies have demonstrated that single coresidences are not the only economic units with evidence of daily production and consumption (Goldstein 2007). Nash (2002), in her study of Wari households in the south-central Andes, has suggested that independent domestic units (residences) are not necessarily economically independent:

Mothers, associated with cooking fires, define the simplest family group—a woman and those to whom she serves food. Nevertheless these food preparation divisions do not necessarily carry over into other productive tasks that benefit the larger group spread out in several structures. Meals may be cooked in separate rooms, but prepared food may be sent between structures to maintain ties that coordinate larger labor activities. [Nash 2002:51]

The classification of the household as a coresidential work unit is drawn from the agrarian production component of the domestic economy: men from a single house tend their own fields and provide for the family members who live with them. Although this may accurately describe agricultural production (although perhaps not), it does not address the complementary component of the domestic economy that occurs within the house, where women, as well as men, perform an essential portion of domestic labor on a daily basis.

Many studies have looked at how agrarian production is divided into everyday agricultural production and only occasional events of communal production; fewer have examined the productive activities that occur outside of the fields. Instead, the domestic economy as a whole has been characterized by a labor model that really has only been demonstrated to apply to agricultural production. By using archaeological remains as a guide and not an a priori assumption of domestic organization, we can develop culturally appropriate models of basic economic organization in past societies.

In this chapter I will use “house” to mean a coresidential structure in which at least some members of a household reside. “Household” signifies an independent economic, and presumably social, unit. I do not equate the house with the household—that is, those who live together in a single structure do not necessarily form the basic economic unit.

Deconstructing the Traditional Household and Labor Organization

In traditional descriptions of domestic economy in the Andes, the household shares daily tasks of an agrarian nature and lives in a single structure or in adjoining structures. Ethnohistoric records of the Inka Empire have contributed to the development of this house-based model of the household. The Inka Empire was largely organized as a means of extracting labor and tribute from its populace (D'Altroy and Earle 1985; Earle 1994). Drawing on Goody (1982) and Mayer (1977), Stanish (1992:27) suggests that during this period, the basic tax unit was the “traditional household,” a married couple and their children, who worked together as a unit to provide tribute. Within the household, the government categorized all individuals into different classes on the basis of their productive capacity (Guamán Poma 1956 [1613]:137). Infants, children, adults, elders, and disabled individuals were classified according to their abilities to provide labor or tribute to the empire: “[the empire] separated the Indians into ten classes to be able to count them, in order that they were employed in work according to their capacity and that there were no idle people in this reign” (Guamán Poma 1956 [1613]:137; my translation). Each age and gender was assigned a specific role, according to seasonal activities, and was required to participate in the maintenance of the household. Men engaged in seasonal agriculture, as well as lithic production and some herding; women also played a role in seasonal agricultural work, as well as spinning, weaving, and herding; children were responsible for many small tasks, such as gathering fuel wood, plants for dyes, or herbs, as well as herding (Guamán Poma 1956 [1613]).

Individuals were embedded within extended kin groups, ayllus, that provided a larger network of social and economic support on a limited basis. Even so, they were essential components of pre-Hispanic Andean life: members cooperated in the management of land and herds held in common, as well as in networks of reciprocal labor exchange. Labor in an Andean community has thus been divided into two discrete entities: household and corporate. These two units, however, are both bound by the same relationships: kinship. Ethnohistorians describe the role of the ayllu, or community kin group, in relation to punctuated events of agricultural production, such as the harvest or canal cleaning; they do not, however, specifically address the role that these kin relationships played on an everyday basis. I suggest, on the basis of a reinterpretation of Guamán Poma's descriptions, and, later, archaeological evidence, that although an ayllu probably did not form a basic domestic work unit, closer kin members that belonged to the same ayllu probably did participate on an everyday basis to provide the basic economic requirements that sustained a household. Instead of single house units constituting single households, it is more likely that individuals from multiple houses pooled labor on a daily basis to meet the requirements of daily life.

Although Guamán Poma's ethnohistoric documentation of household productive tasks highlights the roles of individuals, it also emphasizes the necessity of an integrated household for efficient production. Many different types of labor were required to maintain a household and provide tribute. Guamán Poma's description, however, does not define the members that make up a household. His accounts of different age-grade tasks describe a network of individuals who mutually support one another in their work on a daily basis: children help their parents or grandparents, as well as care for the “young ones” of the household; an infant is cared for by his or her mother, who is helped by her sisters, her grandparents, her aunts, or any other nearby relative (Guamán Poma 1956 [1613]:164). Although the Inka Empire extracted tribute based on a single household unit, it is unclear whether this limited daily economic ties among these units.

As seen in Guamán Poma's description of everyday domestic tasks, related individuals worked together to maintain the households. As recognized by both Weismantel (1988) and Lambert (1977) in their ethnographic research, household tasks are more easily accomplished by pooling the economic potential of multiple individuals and households. Instead of employing a model that isolates domestic production from corporate production, it may be more useful to recognize the potential integration of house and supra-house production. By looking at the degree of integration of house economies, we can enhance our understanding of the reality that may or may not correspond to the ideal model.

An Archaeological Approach: Grinding Stones

The archaeological record and the artifacts of daily life enable us to develop a more accurate understanding of daily life and the domestic economy than that derived solely from ethnohistoric and ethnographic sources. Once we have identified the tools essential for domestic production, we can assess the degree to which coresidentiality and economic work units correlate, which will refine our understanding of households and their economic activities. We can begin to understand how household production and consumption shaped daily life in the past.

In the Andes, essential elements that compose the domestic tool kit include a hearth, cooking pots, and a grinding stone (Bollinger 1993; Johnsson 1986; Weismantel 1988). Although hearths and cookware are typically associated with domestic contexts, the presence or absence of grinding stones has not been considered. As Bollinger, an Andean scholar, reports, “In the domestic economy, there were essential pits that served as storage. Equally important was the recipient for grinding and the grinding stone (batán) to grind maize” (Bollinger 1993:27; my translation). In his study of a Bolivian Aymara community, Johnsson also notes that “[t]he most important piece of equipment in every kitchen, after the hearth, is the grinding stone, k'iaña. The stone, which is used several times a day, can have various shapes” (Johnsson 1986:47). Grinding stones are still essential tools of households today (although they are increasingly being replaced by blenders and the convenience of preprocessed flours), and they probably served a similarly important function in the past.

Grinding stones are used to grind a variety of grains or tubers to make flour (Bollinger 1993:32). They are used to extract oils from vegetables, such as maize or peanuts (Bollinger 1993:32); they also are used to press and mince vegetables for cooking (Johnsson 1986:47). Different-shaped stones are used for different activities: round, flat, or oblong stones are used to grind or mince, while spherical stones are used to tenderize meat (Johnsson 1986:47). Some grinding stones are stationary, a household fixture, and others are portable, depending on their specific uses. In the Andes, the types of foods that were ground in the past probably varied according to available resources; however, hard grains such as quinoa, maize, or kiwicha, which grew in different ecological zones, could have all been processed on grinding stones, either to remove the chaff or to make flour.

In her study of Zumbagua life in the Ecuadorian Andes, Weismantel (1988) suggests that grinding stones often serve as markers of a household's life stage; thus, not all households have grinding stones, and others may have more than one. When the children of a household grow up and marry, they do not always establish their own home right away. Even when the couple move out of the parents' home and begin to cook their own meals, they often do not have a grinding stone but return to the parents' home to complete this daily task. Weismantel suggests that this is a conscious economic decision: large households, with various generations of inhabitants, have a greater capacity to complete daily household tasks.

In the Andes, hearths and food remains have been employed as markers of domestic activity because they are essential elements for food production: cooking requires hearths and leaves behind distinct residues. Despite clear ethnographic evidence that grinding stones are fundamental to cooking technology in the Andes, they have not been considered essential to the domestic tool kit. Grinding stones often appear in domestic contexts, but their presence is variable: a single grindstone is not always associated with a single hearth and single domestic unit. Given the everyday role grinding stones play presently in Andean peasant society and probably played in the past, their presence or absence should be taken into account to understand how the domestic economy functioned on a daily basis.

Case Study: Mantaro Valley

I examine domestic excavations from six sites in the highland Andes of central Peru and identify the presence and absence of two features in individual house compounds: hearths and grinding stones. The sites that compose the Upper Mantaro Archaeological Research Project (UMARP) date to the Late Intermediate period and Late Horizon (1000–1450 c.e. and 1450–1533 c.e., respectively) (Figure 3.1). The data for this study were compiled from the UMARP site report (Earle et al. 1987), Glenn Russell's (1988) dissertation on stone tools, and the edited volume Empire and Domestic Economy (D'Altroy and Hastorf 2001). I am not identifying a pan-Andean practice of domestic organization, but I highlight potential avenues of future research based on the presence or absence of material evidence. Although other household studies exist in the Andes, the horizontal excavations of domestic contexts of UMARP are exceptional in that excavation records provide plans identifying both grinding stones and hearths. Grinding stone counts are based on the presence of fragmentary or whole batanes, the passive components of grinding implements.

Excavations focus on elite and commoner houses, but for this study only commoner houses are examined because, as part of feasting, elite houses often process and prepare foods for nonhousehold members, which could skew the results (Russell 1988). The sample of houses is drawn from six sites in the Mantaro Valley: Hatunmarca (J2), Tunanmarca (J7), Umpamalca (J41), Marca (J54), Chucchus (J74), and Huancas de la Cruz (J59) (Earle et al. 1987). Houses consist of at least one structure and a walled patio; excavated areas range from 12 to 100 percent of the total patio group area. Twenty-nine patio groups were at least partially excavated, 17 nonelite and 12 elite; 16 of the nonelite patio groups are analyzed in this study (records for one of the nonelite patio groups were incomplete).


The patio groups of commoner houses have between one and three separate structures within a single compound (Figure 3.2a–c; Table 3.1). Drawing from Weismantel's observations in the Ecuadorian highlands, I propose that house composition varies according to its “life stage.” Thus, a more “mature” dwelling would house a larger group of people and enable it to function more efficiently. “Younger” dwellings, on the other hand, whose members are far fewer, often rely on the efficiency of the larger, parent dwellings for everyday necessities. Applying this to the Mantaro case, all else being relatively equal, a dwelling with more structures has the potential to house more individuals than a single-structure compound. If Weismantel's observation that the number of grinding stones in a house correlates with its age applies to the Mantaro houses, we should see a correlation between the number of structures and the number of grinding stones, a necessary tool required for the labor-intensive, everyday chore of grinding various foodstuffs required for household production. Houses without grinding stones should be “younger,” and they probably maintained a close relationship to other house compounds to fulfill their daily domestic economic needs.

Figure 3.2.

Figure 3.2.

Examples of compounds with one to three separate structures: (a) Patio group J7 = 4, a one-structure patio group (redrawn from Earle et al. 1987:Figure 8); (b) Patio group J41 = 5, a two-structure patio group (redrawn from Earle et al. 1987:Figure 22); (c) Patio group J7 = 8, a three-structure patio group; the northeast structures without numbers do not open onto the shared patio of J7 = 8 and were not considered part of this patio group (redrawn from Earle et al. 1987:Figure 14).

Figure 3.2.

Figure 3.2.

Examples of compounds with one to three separate structures: (a) Patio group J7 = 4, a one-structure patio group (redrawn from Earle et al. 1987:Figure 8); (b) Patio group J41 = 5, a two-structure patio group (redrawn from Earle et al. 1987:Figure 22); (c) Patio group J7 = 8, a three-structure patio group; the northeast structures without numbers do not open onto the shared patio of J7 = 8 and were not considered part of this patio group (redrawn from Earle et al. 1987:Figure 14).

Figure 3.2.

Figure 3.2.

Examples of compounds with one to three separate structures: (a) Patio group J7 = 4, a one-structure patio group (redrawn from Earle et al. 1987:Figure 8); (b) Patio group J41 = 5, a two-structure patio group (redrawn from Earle et al. 1987:Figure 22); (c) Patio group J7 = 8, a three-structure patio group; the northeast structures without numbers do not open onto the shared patio of J7 = 8 and were not considered part of this patio group (redrawn from Earle et al. 1987:Figure 14).

Table 3.1.  Area excavated, number of fragmentary or whole grinding stones, and number of hearths in nonelite patio groups
PeriodStructureNo. of StructuresArea Excavated (%)No. of Grinding StonesNo. of Grinding Stones Standardized for Area (%) ExcavatedNo. of Hearths
  1. Note: WII = Wanka II; WIII = Wanka III. See Figure 3.1 for site numbers.

WII2 = 4129270
WII7 = 41100 000
WII7 = 5127141
WII7 = 6146120
WII7 = 8334721 0
WII7 = 9135000
WII41 = 4 181111
WII41 = 5 258000
WII41 = 7 120001
WIII    2 = 6/III118.7000
WIII 54 = 2  36610 15 1
WIII 54 = 9/C133003
WIII54 = 10317529 0
WIII59 = 1 3Unknown1Unknown1
WIII74 = 1 212216 2
WIII74 = 2 218527 0

The excavation sampling strategy of UMARP focused on recovering the widest variability of features and artifacts (described in Earle et al. 1987:12–14) within each patio group. Even so, because the house compounds were not always completely excavated, we have to consider the likelihood that our sample of grinding stones and hearths may be skewed as a result of partial excavations. Attempting to standardize the data set, however, presents some problems. If we standardize the samples for the area excavated, we are assuming that the grinding stones are evenly distributed throughout the area of the patio group, which is probably not the case; this also cannot address the houses in which no grinding stones were found. However, if we look to see whether there is a correlation between the amount of area excavated and the number of grinding stones recovered, it turns out that this correlation is not significant (–.057). Thus, it does not appear that the percent of area excavated per compound significantly affects the number of grinding stones excavated (Figure 3.3).

Figure 3.3.

Graph showing the lack of correlation between the percent of each patio group excavated and the number of grinding stones recovered.

When we look for correlations between the number of structures in a patio group and the number of grinding stones, we find a correlation coefficient of .738, significant to the .01 level (Table 3.2). When this is standardized for percent excavated, the value is reduced only slightly to .652, also significant at the .01 level. Thus, regardless of the area excavated, we find a significant correlation between the number of structures in a patio group and the number of grinding stones present. However, four of the samples are based on less than 20-percent excavated area of patio groups. If we discard these samples on the basis that they are too small to be representative and then analyze the correlations between the number of structures and grinding stones, the correlation coefficient increases to .878, significant to the .01 level; standardized for percent excavated, the correlation coefficient remains significant, .847, also at the .01 level (Table 3.2).

Table 3.2.  Correlation of number of structures within patio groups to number of grinding stones and hearths
RelationshipsPearson's CorrelationSignificance
Structures–grinding stones
  Uncorrected.738.01 (two-tailed)
  Corrected.652.01 (two-tailed)
Structures–grinding stones, greater than 20% excavated
  Uncorrected.878.01 (two-tailed)
  Corrected.847.01 (two-tailed)
  Uncorrected–.075 Not significant
  Corrected–.057 Not significant
Structures–hearths,greater than 20% excavated
  Uncorrected–.156 Not significant
  Corrected–.279 Not significant

The relationship between the number of structures and the number of hearths does not follow this pattern. In nine of the 16 cases, no hearths were excavated within the patio groups; in seven, one or more hearths were present. Once again, we must consider the effect of partial excavations on the recovery of hearths; in this case, however, it is more difficult given the multiple patio groups where no hearths were found. For these cases, we cannot project with much confidence the potential number of hearths expected for each house compound. Although it does not appear that there is a correlation between the number of structures and the number of hearths (–.075), even when corrected for area excavated (–.057), the only certainty is that the distribution of hearths appears to be more variable than the distribution of grinding stones (Table 3.2). How can we make sense of these inconsistent patterns of hearths and grinding stones, two basic necessities of a functioning household?


The patterning of grinding stones and house structures is consistent with Weismantel's model: the greater the number of structures, the greater the number of grinding stones associated with the house compound. This is not altogether surprising if we consider that more people are living in these larger compounds and thus require more food processing to support their economic needs. However, taken together with the absence of grinding stones in some patio groups, a pattern of shared economic resources begins to emerge. Grinding stones are associated with a task that requires substantial time and labor investment. By coupling the economic activities of smaller households with those of larger households, the smaller household can depend on a larger labor pool that often includes many children who are capable of completing multiple tasks, such as grinding. Weismantel describes how a well-established household has more household members, as well as a more elaborate food production tool kit: “This household was well into the most expanded phase in its life cycle, with in-laws and grandchildren in residence. This fact was represented by the presence of two kutana rumis, big grinding stones, on the kitchen floor” (Weismantel 1988:176). Older houses, with multiple generations, have more available labor, which makes them economically more productive. Younger houses remain intimately connected with their parent houses, and this plays out on a daily basis when members of a younger house rely on parent houses to aid them in everyday productive tasks.

Instead of presenting a standardized image of a nuclear-family, coresidential, gender-divided productive household, this new model demonstrates a far more flexible arrangement of domestic production and consumption that can play out across various house structures. Likewise, the individual members of an independent household will shift over time, according to the life cycles of families: a single-structure house that once depended on its “parent” house compound for daily needs may become independent when it acquires enough members to become economically viable. Hearths, on the other hand, are far more variable. Although hearths, which provide warmth and can be used to cook food or boil water, are essential to daily life, they do not necessarily indicate an economically independent household.

This model of the domestic economy, focused on the houses and their components, may also prove useful as a means of assessing the age of a site. Sites with longer periods of occupation would demonstrate evidence of houses in more advanced life stages, while those with shorter occupations would have fewer “mature” houses. In our case study of the Mantaro Valley, both periods, Wanka II and Wanka III, were relatively brief (100 years and 83 years, respectively), resulting in short occupations of the associated sites. An examination of the nine house compounds from Wanka II sites demonstrates that only two have more than a single structure. This dearth of multistructure house compounds may be attributed to the short occupation of the site: fewer generations of families resulted in houses that were generally “young” (i.e., single structure). In Wanka III, however, despite an equally brief occupation, we see a shift toward house compounds with more, larger structures and correspondingly more grinding stones. As others have suggested, this may be attributed to greater demands for tribute during the Inka Empire, when more space was needed for productive tasks (DeMarrais 2001; Hastorf 1991); I would also suggest that it became more efficient than in the previous period to remain in multihouse households rather than to establish new single-house households, which resulted in a larger number of people living in a single house compound. This may be attributed to tribute requirements under the Inka Empire that quantified the “household” as the unit of taxation (D'Altroy and Earle 1985; Hastorf 2001).

Although archaeologists have increasingly paid attention to domestic spaces, we need a more refined understanding of the essential elements that constitute a single productive household. Depending on the productive needs of a household, this will vary across time and space. Likewise, although it is useful to identify repeating activities and tool kits as a way of highlighting discrete economic units, it is also necessary to recognize where certain activities or tools are missing, even though the others may be present. As seen in grinding stones, these may not be present in every house, which is significant in understanding how daily tasks are divided and organized among households. The absence of grinding stones may highlight fundamental links of economic activities between discrete architectural units that would otherwise appear economically autonomous. Taking into account both the presence and absence of domestic tools is essential to interpret economic dynamics of the house.


A gendered approach to the domestic economy liberates it from previous assumptions and historical precedents that emphasized the male sphere of agricultural production. The traditional nongendered approach, ironically, had ascribed gender to various tasks, effectively obscuring the complexities of domestic roles and relationships. In contrast, the gendered approach of this chapter attempts to “de-gender” the domestic economy, returning the focus of study to the material evidence. By turning our analysis around and centering on the house, instead of the model, and on the lived realities of productive individuals, instead of stereotyped gender roles, we emerge with a distinct portrait of the domestic economy and daily life in these ancient Andean homes.

If we continue to define a household as the fundamental economic unit, we need to identify the essential domestic tools that households require. By emphasizing the material necessities of production, we can evaluate the degree to which individual economic units correlate with coresidential units. If tools do not always correspond to coresidential units, this needs to be accounted for in our interpretation of economic organization. Instead of embracing the domestic–corporate dichotomy discussed earlier, we can employ the archaeological record to identify how activities pattern within and between households.

The uneven distribution of grinding stones among structures may suggest an alternative to largely nucleated household economies described earlier. The Mantaro Valley house studies demonstrate that small-scale household production may represent a degree of corporate production beyond that of a single coresidential unit. The absence of grinding stones in small compounds and their presence in large ones may illustrate a regular, daily pattern of supra-house production in which multiple “nuclear” families share essential household tools and labor. As seen elsewhere in the Andes (Goldstein 2007), archaeological evidence appears to point to a type of economic production that falls somewhere between the extremes of domestic and corporate units.

In the Andes, the absence of a grinding stone in a house does not necessarily point to an absence of grinding as an important daily task of this household. Instead, this absence may be best explained by expanding our scope of analysis to include several coresidential units. On a daily basis, related people may be sharing basic production and consumption activities, even though they do not inhabit the same structure. Archaeologists may be able to distinguish the age of a household and its life-cycle stage at the time of site abandonment by examining the types and quantities of productive tools in its possession. Likewise, an analysis of multiple households at a site may provide a general tool for assessing the age of a site: many households in later life-cycle stages (i.e., many structures and grinding stones and so forth) would indicate a site that had been occupied much longer (i.e., many more generations) than sites with mostly “young” households.

The variety of household tasks described in Guamán Poma's documents demonstrates that the domestic economy was characterized by far more than that which was produced in the fields. For that reason, the traditional agrarian domestic economy model cannot address the complexities of ancient household production and consumption. A gendered approach to the domestic economy has highlighted the importance individuals and lived realities play in the life cycles of houses and their economic production. The importance of kinship in the Andes cannot be ignored; likewise, the ways in which these ties crosscut households and communities have to be understood when approaching the archaeological record. Ethnohistoric records are essential to provide models of the past. However, archaeologists and historians must be careful not to interpret the documents on the basis of some aspects (e.g., nuclear families as the unit of tribute assessment) at the expense of others (e.g., division of labor on bases other than sex); likewise, a careful reading is required to avoid confusing idealized presentations of economic organization with the reality of daily life. In this way, archaeology is unique in its capacity to provide accurate representations of the repeated actions that occurred on an everyday basis.

This analysis has grouped together Andean households and domestic practices into a single entity. Although many traditions are widely shared in the Andes, food production and consumption practices can be highly localized and distinguish groups of people from each other (Weismantel 1988). As more household studies are completed, we can refine our understanding of domestic practices to be able to identify variation based on ethnicity, class, or period. Grinding stones, hearths, and food remains may not be the only important correlates of basic economic behavior in the Andes. Cooking and serving vessels may also pattern variably among households in ways that demonstrate specific economic links. With these new data, we will be required to reevaluate our current model of the domestic tool kit and make refinements that will more accurately depict past practices. Moreover, by compiling specific indices of domestic production and consumption, we may be able to identify households in their different life stages, as well as important economic links between households.