The Primacy of the Caracol Household

The archaeological research at Caracol (and elsewhere) demonstrates that the residential group constituted a basic physical and organizational feature of Classic Maya society. Characterized by both raised platforms and unraised plaza areas with one or more buildings on the various plaza sides, Caracol's residential plazas largely align with the cardinal directions, probably a distant reflection of principles still used in modern Yucatan (Watanabe 1992:65). Residential plazuela units at Caracol generally contained from four to six total buildings, although some groups contained both smaller and larger numbers. The buildings also vary in size and complexity: some are single-room perishable structures with stone foundations, while others are multi-room structures with vaulted masonry construction. Nevertheless, all these basic units seem to have functioned in similar ways.

In approximately 80% of the mapped Caracol residential groups, an eastern building functioned as a shrine (Figure 1). This building contained both honored dead and related ritual offerings. Interments were generally made in tombs (“formal constructions” or “chambers” that were “larger than necessary to hold their contents”; see A. Chase and D. Chase 1987:57 for expanded discussion of these definitions and terms), crypts (graves with “formal walls and roofs” that “are not much larger than necessary to hold their contents”), and cists (graves “with clear outlines” but without the “formal construction of either walls or roof”), although simple interments (inclusion of the body in construction fill or garbage without a formal grave) are also present. In spite of a diversity of interments, however, the limited number of burials in any one group indicates that not all the individuals who lived in the group could have been buried in these buildings. In fact, assuming that five individuals were living in any given residential building and taking into account the time depth of an excavated unit, no more than 10–20% of the population who would have lived in a given household group are reflected in burials recovered from well-excavated Classic-era residential groups (D. Chase 1997). Even if only five people lived in an entire multiple-building residential plaza group, there still are far fewer burials than former residents. At Tikal, some groups contain only males (Haviland 1988). At Santa Rita Corozal, some Late Postclassic groups are associated only with females and children (D. Chase 1982; D. Chase and A. Chase 1988). Interestingly, at Caracol individuals who were interred in these places of honor and ritual during the Late Classic period included both sexes, as well as adults and subadults.

The “Caracol Identity”: examples of typical Caracol “face caches,” of a Caracol east-focus structure group, and of tombs in an eastern building in a Caracol residential group.

More than 40% of Caracol's 259 interments contained more than one individual (D. Chase 1994, 1998; D. Chase and A. Chase 1996, 2003). The presence of multiple individuals within a single chamber in Caracol's residential groups suggests that the individuals inside tombs were not limited to the single founding individual for a particular plazuela group. In fact, the deposition of individuals in interments in eastern buildings may have been conditioned by ritual schedules rather than by individual historic death events (D. Chase and A. Chase 2002). Thus, a complicated situation can be documented archaeologically in which it is unclear what kind of ancestors were being buried in eastern buildings, especially as the deposits themselves may have been considered to be ritual offerings rather than individual-specific interments of important predecessors (see also Becker 1988, 1992, 1993; D. Chase 1988; D. Chase and A. Chase 1998; Kunen et al. 2002). Whether ancestors or offerings, interments both conditioned and were conditioned by the religious, social, and political lives of the individuals living in associated residential groups. But even if these interments related lineally to the living household, not all ancestors were interred in tombs in the eastern buildings of their residential groups: more than one ancestor could be buried in a single interment; more than one tomb could be located in the eastern shrine; and far too many mortuary structures exist in residential groups for each shrine to have housed separate lineage ancestors. If true ancestor veneration existed among the Caracoleños, these data imply that the residential group—rather than some larger unit—served as the focus for this ritual. And other community building forms are not in evidence in the Caracol landscape that could be used to suggest any particular hierarchy for venerated ancestors (as is found in Chinese society [Freedman 1966, 1970]).

Eastern buildings in plazuela groups at Caracol also yielded other ritual offerings. Caches consisted predominantly of ceramic urns with modeled faces or of lip-to-lip bowls containing the remains of human fingers (A. Chase and D. Chase 1994; D. Chase and A. Chase 1998). Limited numbers of these ritual deposits were usually placed on axis to and in front of eastern buildings in residential groups; in several documented cases, however, dozens of these containers were placed on axis to eastern buildings in front of and within stairs (Jaeger 1991). While the caches in royal or elite contexts may have been more elaborate, face caches and finger bowls were deposited throughout Caracol and in the same plazuela group locations regardless of whether the residents were of high or low status. Thus, each of the individual plazuela groups witnessed the performance of pan-Caracol ritual. Assuming that 80% of Caracol's plazuela groups participated in this kind of activity, as the archaeological data indicate, then the total number of eastern ritual buildings at the site would have numbered more than 7,000. The caching practices, like burial practices, are far too widespread to have taken place solely in households associated with a lineage head.

Besides caching practices, other indicators exist to suggest that ritual activity was uniform and widespread at Caracol. Many of these same residential groups contain the remains of modeled incensarios associated with these eastern buildings. Hourglass-shape incensarios dating from the early part of the Late Classic period have been recovered in a number of tombs from outlying residential groups. Either ceiba-tree spikes or a face representing the jaguar god of the underworld decorate these incensarios. Other plain and spiked censers come from tombs in Late Classic residential groups. During the Terminal Classic era, a number of cylindrical and flanged incense burners—again modeled with the face of a large jaguar god of the underworld—were left on the stairs of eastern buildings or were buried beneath their lower steps. Thus, throughout the Late Classic era it appears that Caracol's residential groups participated in religious rituals that were often restricted to epicentral locales at many other sites (Rice 1999).

Besides mortuary and religious ritual, other activities occurred within the Caracol residential groups. Each residential group appears to have had at least one economic specialization, and neighboring groups often had different specialties. Individual groups focused primarily on bone, shell, wood, or stone tool production. Agricultural production at Caracol also was conducted in terraced fields adjacent to plazuela groups. A consideration of settlement survey suggests that each residential group had access to approximately 2.5 ha of land. While great independence existed relative to the production of goods at each Caracol residence, we believe that the areas for the distribution of the manufactured objects were tightly controlled allocation points in the open non-residential plaza areas associated with the Caracol causeway termini, probably utilized in a system of solar markets (A. Chase and D. Chase 2001a). Classic-era Caracol, with its vast system of dendritic causeways linking all of the large termini plazas directly to the site epicenter, would have functioned as an administered economy in which distribution was tightly controlled while production continued unabated at the household level (see A. Chase 1998). As at Tikal (Haviland 1974, but see Moholy-Nagy 1997), no indication of attached specialists (or even artisans located in the immediate proximity to elite residences) exists, as is suggested for some sites (such as Copan).

It is not clear how many people lived in each residential plazuela group. If one removes the eastern ritual building from the sample and assumes five people per building in a plazuela group (provided that each building within a plazuela group was a residence for either separate families or for an extended family), the average residential group would have housed between 15 and 25 people. But, as indicated earlier, a few groups had fewer buildings, and a number had more buildings (and/or buildings with multiple rooms). Thus, some households could have been substantially larger or smaller. This estimate for residential group size compares to ethnographic and ethnohistoric data that range in number from fewer than five (Haviland 1972) to more than 40 (Hellmuth 1977) people in a single household. Matthew Restall (1997:100) indicated that generally 6–12 members of an extended family lived in sixteenth-century households in Yucatan. But testament data suggested that this family would have lived in a solar or house lot containing only two buildings (Restall 1997:105–6). Restall noted that there may have been other perishable buildings not mentioned for the purposes of inheritance; however, the smaller number of structures indicated in his data suggests that the ancient residential group, with its larger number of buildings, may have been bigger. In contrast to the archaeology and to information in central Mexican documents, Restall's data (1997:106) do not indicate separate ritual or storage buildings as existing within Maya residential groups, suggesting potential changes in the function of individual structures and/or the residential group from the Classic era to Historic times. Thus, the archaeological estimates of 15–25 people based on structure numbers may well be correct.

Skeletal data (and genetically determined traits) from burials can be used to suggest that people living in an archaeological residential group were likely related to each other. In some cases, sequent burials placed within the same building and separated by some period of time show similar skeletal anomalies, such as the presence of an extra incisor, which are presumably linked genetically (A. Chase 1983; D. Chase 1997). At Caracol, reconstruction of diet based on stable isotope analysis suggests that individuals buried within a single interment or within a single residential group ate their meals together. However, dietary analysis of individuals in neighboring groups indicates that neighbors often consumed very distinctive diets. Variation in protein and maize consumption supports the idea that neighboring groups may not have been closely related in terms of kinship (D. Chase et al. 1998; A. Chase et al. 2001). Only very rarely at Caracol do we find clustering or nesting of residential groups, as would be expected from the natural fissioning of household growth in which related kin would have sought to live near one another. Once households grew to sufficient size to require fissioning, new plazuela groups established themselves elsewhere. Given the regularly spaced dense settlement and ubiquitous terracing that occurs within Caracol's landscape, these fissioned households would have had to establish themselves on the margins of Caracol's settlement, if only because no available localized space existed within the site's established field systems. It should be noted that these archaeological data do not automatically indicate a preference for the use of any specific kinship or descent system by the ancient Maya (A. Chase et al. 2002).

Beyond the Household: The Caracol Community

A pan-Caracol identity characterized the broader Caracol community. The identity is evident in ritual caches; in the burial of the dead; and in the generalized use, construction, and layout of residential plazuela groups. This observation conforms with various descriptions of Maya communities, including Joyce Marcus's (1983) view of the city as focal and unifying and Restall's (1997:13–19) ethnohistoric description of cah as a geographical community potentially similar to the Nahua altepetl. The community identity surely extended for a distance beyond the city itself into the heart of the area over which Caracol had social and political control during the seventh and eighth centuries. For example, the apparent abundance of eastern shrines in the southeastern Peten of Guatemala may be interpreted as demonstrating this area's close linkage with Caracol (A. Chase 2004).

Of greater difficulty to discern are units smaller than the city or polity but larger than the residential group. Archaeologists have attempted to find divisions in sites—calling them barrios or sectors—and have defined them based on a variety of factors. For Late Postclassic Santa Rita Corozal in northern Belize, distinctive cache offerings and the distribution of certain building types can be used to segment the site. In particular, multiroom structures were not concentrated in one specific area of the site (as they were at coeval Mayapan farther to the north); rather, this distinctive building type appeared in a dispersed distribution scattered throughout the Late Postclassic settlement. However, these structures could also be correlated with distinct ritual (katun idols; alcohol consumption) and administrative functions (audience halls) that would have been important for town leaders. In addition, the Late Postclassic caching patterns found within the extended areas of each of the recognized multiroom structures varied stylistically from each other. Thus, the distribution of this structural type and the differences in caching patterns effectively segment Late Postclassic Santa Rita Corozal into smaller divisions (D. Chase 1982, 1986).

Other models for subdividing Maya towns and polities derive from ethnohistoric and epigraphic data. Ralph Roys (1972 [1943]:11) suggested that the Maya had both subdivisions of a town and larger recognized districts, as indicated through the use of terms such as cuchteel and cuchcabal. But other researchers have found that these terms had quite varied meanings (Restall 1997: 24–25, 27).

Epigraphic interpretation sometimes claims smaller social or political divisions than those seen on the level of the emblem glyph (Mathews 1991). Linda Schele and Peter Mathews (1998:23) have argued that political divisions of a single polity are sometimes reflected within the hieroglyphs, specifically arguing that the broader Tikal polity was composed of 13 distinct divisions. Such divisions would have been distinct from the named places also found in hieroglyphic texts (e.g. Stuart and Houston 1994).

Settlement study at Caracol clearly shows the layout of the ancient city and provides a sense of the relationships among residential groups (Figure 2). Caching patterns and the use of eastern buildings for burials are pan-Caracol patterns with no clear evidence for subset groupings such as barrios or sectors. But variation in status apparently occurs within certain settlement areas. High-status elite and royal households can be located in Caracol's epicentral palaces. The most massive architectural construction at the site, Caana, is likely to have functioned as the royal household for Caracol's ruler (A. Chase and D. Chase 2001b). High-status elite palaces also occur at some of the site's causeway termini (Figure 2). Selected plazuela groups in other locations (such as the residential group at the end of the Machete Causeway spur off the Conchita Causeway in southeast Caracol) also housed high-status elite. These high-status groups are associated with eastern buildings (or temples in a few cases) that contain tombs with painted linear decoration and texts. Dates associated with these tombs presumably represent either chamber consecration or an individual's death date (e.g., D. Chase and A. Chase 1996). The high status of these painted chambers finds confirmation in the use of the Caracol emblem glyph in one of the associated texts (A. Chase and D. Chase 1987). Other artifacts, such as jadeite earflares, also occur in the painted epicentral tombs and may be correlated with extremely high status (D. Chase 1998).

Site map of Caracol, Belize, at the end of the 2000 field season. Squares are 500 m × 500 m and represent fully mapped grids; diagonal lines are causeways.

Combining settlement data with diet leads to other interesting correlations of status variation. Stable isotope analysis indicates that individuals of highest status (those who lived in the palaces) maintained a distinctive diet that was high in both maize and protein. Closest to this diet were individuals interred in large raised residential plazas adjacent to causeway termini. Farthest from this diet, and low in maize and protein, were individuals buried in residential groups immediately surrounding the epicenter (A. Chase et al. 2001). The majority of Caracol's inhabitants, however, had an intermediate, but generally good, diet. Thus, the settlement and dietary data suggest a situation more closely representing Ernest Burgess's (1967 [1923]) concentric model describing a twentieth-century city. This prompts serious consideration of whether some of the models used to describe the ancient Maya might also not be better derived from more modern urban contexts rather than from simpler, largely kinship-based, communities.

To summarize, investigations at Caracol, Belize, provide some insight into the nature of ancient Maya social and political organization. The Late Classic Maya of Caracol maintained a basic social unit of the residence in a single plazuela group. We suggest that this group represented an extended family ranging from five to substantially more than 40 individuals, but probably most often with between 15 and 25 family members. These people lived and ate together. They also appear to have formed a corporate unit for economic purposes. The overall Caracol community maintained its own distinctive social identity during the Late Classic period (A. Chase and D. Chase 1996a). While subdivisions are likely to have occurred among the population at this time, the recovered archaeological record reveals combined status, dietary, and settlement differences. Royal individuals mostly lived in palaces in the epicenter. Support populations, probably without access to agricultural land, lived in a surrounding ring of settlement immediately adjacent to Caracol's downtown area. Beyond this small epicentral ring, a larger area of core settlement included varied status levels in immediate proximity to each other. Each of these residential plazuela units appears to have been situated contiguous to 2.5 ha of agricultural land.

Most outlying plaza groups tended to focus on a single economic activity, often related to woodworking or bone-working, but also including shell and lithic tool production; individuals in many residential groups also probably spun cloth. The finished goods produced in these residential groups were ostensibly traded for other goods, such as ceramics, at solar marketplaces presumably formed by many of Caracol's public termini groups. While production remained independent, the distribution of the finished goods would have been controlled by the site's elite at these market locales. Apparently, no specialists were attached to palaces; rather, economic production was again localized, as in the outlying plaza groups. But individuals within the palaces, presumably part of the extended royal household, are known to have worked shell and bone and probably to have spun cloth. Thus, the palace occupants alone participated in Caracol's administered economy as both producers and distributors.

The Caracol site plan provides no indication of new housing for family members in a contiguous nested housing arrangement. Rather, individual residences appear to have expanded the number of structures in the same plazuela arrangement, or, alternatively, groups fissioned, with newer residences, of necessity, being placed at the outskirts of the Caracol metropolitan zone in areas not yet filled with settlement and fields.

Lineages or chibals, while noted in the ethnohistoric materials (e.g., McAnany 1995), cannot be found in the Caracol archaeological record. The site's single-extended-family residential focus, settlement patterns, and status distributions suggest a similarity to certain contemporary or historic urban situations (Burgess 1967 [1923]). An edge-city model has in fact been suggested as providing some utility in attempting to understand the growth of ancient Caracol's economic, communication, and transportation systems (A. Chase et al. 2001; Garreau 1991).

Deranging Factors in Settlement Models

The ease in seeing Maya plazuela groups in the archaeological record has led to an emphasis on the household in settlement archaeology (Wilk and Ashmore 1988). And while a focus on the house or residence should clearly be key in investigations of the ancient Maya, this focus should not be confused with a more abstract “house model.” The sheer numbers of easily identifiable residential units at Maya sites might cause some to think that a house model or noble-house model (Gillespie 2000; Lévi-Strauss 1982) could have direct applicability for the ancient Maya and prove useful in making interpretations about the Maya past. This model, however, has been expanded to include a wide variety of social and political structures (Block 1995:71; Carsten and Hugh-Jones 1995; Waterson 1995:67), and it cannot be expected to reconstruct and differentiate ancient Maya sociopolitical structure.

The wards or barrios and the houses of the principales described by Landa (Tozzer 1941) finds potential archaeological support in Late Postclassic sites such as Santa Rita Corozal and Mayapan (as indicated by caching patterns and structure distribution of multiple room buildings and/or colonnaded halls). Some of the architectural forms found at these two sites may be similar to the Nahua calpolli, or “big house.” This organizational resemblance may be due to the closer temporal relationship between the Postclassic and Historic periods. Alternatively, what appear as historic parallels possibly may have arisen due to Spanish familiarity with Nahua practices or as outgrowths of early sixteenth-century Spanish–Maya interactions. Regardless of similarity, however, such architectural features do not go back in time to the Classic period at Caracol, as these building forms (or their analogs) do not appear in the archaeological record. The palaces recorded for most major Maya sites in the southern lowlands (Andrews 1975; Inomata and Houston 2001) appear to have combined residential and administrative functions. Only at some sites in the northern lowlands can potential Late Classic antecedents to the “big house” be found, such as Structure 424 at Edzna or Structure 44 at Dzibilchaltun (Arnauld 2001).

It is inappropriate to assume the existence of one monolithic organizational system for the Classic Maya, as great variation is found in the layouts and archaeological records of Maya sites. For example, east-focused residential groups have been examined at both Tikal and Caracol. At Tikal, east-focused groups, referred to as “Plaza Plan 2” groups, occur in barely over 14% of the mapped groups (Becker 1982:120, 2003:259); at Caracol, some 80% of the groups are east-focused. At Caracol, specialized cache containers and tombs occur within most east-focused groups; at Tikal, they do not (e.g. Becker and Jones 1999). These site-by-site differences are clearly significant. Thus, we must look at a continuum of settlements to understand best Late Classic Maya social and political organization. Within this continuum, Caracol is clearly one apex and thus not likely to be characteristic of all Maya sites. Similar variability has been noted for central Mexico, where John Chance (2000:498) and James Lockhart (1992: 104) suggested that at least two variant principles of organization existed among the Contact-period Nahua. Roys (1957, 1972 [1943]), in fact, long ago suggested that minimally three variant political organizations existed for the Contact-period Maya. Perhaps the reason that organizational terms beyond cah do not appear to be uniformly defined or employed by modern populations (Restall 1997:24–25) is that the ancient Maya used variant, as opposed to uniform, social and political forms.