Journal of Anthropological Archaeology 18, 471–506 (1999) Article ID jaar.1999.0349, available online at http://www.idealibrary.com on

The Chaco Connection: Evaluating Bonito-Style Architecture in Outlier Communities

Ruth M. Van Dyke

Department of Anthropology, California State University, Fullerton, Fullerton, California 92834-6846

E-mail: [email protected]

Received December 1, 1998; revision received February 12, 1999; accepted May 17, 1999

Because architecture shapes and is shaped by human actions and perceptions, architectural variability has the potential to provide information about relationships among prehistoric social groups. This study examines communicative and enculturative information contained in Bonito- style architecture constructed in Chaco Canyon and outlying communities during the late eleventh century A.D. Does the appearance of Bonito-style architecture at outliers constitute direct involvement on the part of a centralized, Chacoan entity or could local people have been emulating Bonito-style architecture they saw at Chaco or in neighboring communities? These questions have implications for existing models of Chacoan social organization. To investigate, a comparative architectural analysis uses data from 61 great houses in 55 outlier communities. Analysis is based on the premise that outlier similarity should reflect a unified, direct Chacoan source for Bonito-style architecture, and diversity should reflect the converse. Because highly visible, external architectural characteristics can be emulated, five internal, low-visibility great- house architectural attributes were selected for comparison. Results indicate substantial diver- sity is contained within the Chacoan world. A variety of relationships probably existed between outlier communities and Chaco Canyon, and a range of explanatory models is necessary. Bonito-style architecture is more likely to be associated with a struggle to legitimate social power than with spontaneous, cooperative communal activity. Competitive emulation may account for the appearance of Bonito-style architecture in outlier communities toward the local end of the outlier spectrum. © 1999 Academic Press

Architecture is an excellent source of (Adams 1983, 1991; Cameron 1998, 1999; information about past societies. Architec- Ferguson 1996; Hunter-Anderson 1977; ture shapes and is shaped by human ac- Kent 1990; Kintigh 1994; Lipe and Hegmon tivities and perceptions (Bourdieu 1971; 1989; McGuire and Schiffer 1983; Giddens 1984:143-158). Not only are archi- Schlanger 1986). Architectural variability tectural remains perhaps the most dura- has the potential to provide information ble and the most visible aspect of material about relationships among social groups. culture subject to the archaeologist’s gaze, Prehistoric builders made choices about but buildings provide a direct means for materials, techniques, and structural con- reconstruction of the interactive, recursive figurations that cannot be reduced to relationship between lived experience functional concerns. Low-visibility or in- and the built environment. Prehistoric ternal architectural attributes lack overt structures have been employed by ar- communicative potential and thus reflect chaeologists in the American Southwest the learning frameworks of the builders; in the construction of temporal, social, patterning among internal architectural functional, and demographic knowledge characteristics can be used to distinguish

471 0278-4165/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved. 472 RUTH M. VAN DYKE groups with shared enculturative back- relationship between Chaco Canyon and grounds (Carr 1995b). This study uses the outliers nor the function of Bonito- low-visibility architectural attributes in an style architecture within outlying commu- attempt to disentangle the social relation- nities is well understood, however. How ships among the Chaco Anasazi of north- and why did Bonito-style architecture ap- western New Mexico in the late eleventh pear in outlier communities? The broad century A.D. similarities that allow for the definition Chaco Canyon is found at the heart of and recognition of Bonito-style architec- the San Juan Basin, a topographic depres- ture at sites throughout the greater San sion in northwest New Mexico character- Juan Basin strongly suggest the form is ized by gray-green plateaus broken by not the product of independent invention. canyons, mesas, and scarps. Chaco is an What does the appearance of Bonito-style ideal place to focus an architectural inves- architecture entail for relationships be- tigation, as the imposing ruins clustered at tween outliers and Chaco Canyon and for the heart of this desolate region of stark the social structure of the outlier commu- contrasts have drawn archaeologists’ at- nities themselves? In the study presented tention for over a century. Canyon great here, data from 61 great houses in 55 out- houses such as Pueblo Bonito and Chetro lier communities are compared in order to Ketl are among the best preserved prehis- investigate these questions. The analysis toric pueblo ruins in the Southwest, with is based on the premise that outlier great standing walls in excess of8minheight. house similarity should reflect a Chacoan These planned, massive structures repre- source for Bonito-style architecture and sent a substantial investment of labor and diversity should reflect the converse. Re- design. Great house construction was ini- sults indicate that tremendous diversity tiated in Chaco Canyon during the late exists within the confines of what is con- ninth century and escalated through the sidered Chacoan. A variety of relation- early twelfth century A.D. During the ships probably existed between outlier same period, great houses and associated communities and Chaco Canyon, and a features appeared across the greater San range of explanatory models may be nec- Juan Basin, often in the midst of commu- essary. Construction of Bonito-style archi- nities of small, domestic sites. Great tecture at the local end of the spectrum is houses and associated features are re- likely related to community ritual and ferred to as Bonito-style architecture. Great power issues. Stein and Lekson’s (1992) houses outside Chaco Canyon often are concept of ritual landscape and Renfrew called outliers in reference to their spatial and Cherry’s (1986) peer–polity interac- relationship to the Canyon (Fig. 1). tion model explain how Bonito-style ar- A number of models have been devel- chitecture may have spread across a wide oped to explain the rather dramatic ap- area without necessitating direct contact pearance of these imposing structures with Chaco Canyon. across an arid, agriculturally marginal landscape. Early Chacoan explanations THE CHACO ANASAZI focused on the canyon. More recent work has recognized that relationships must The geological depression known as the have existed between the canyon and the San Juan Basin covers approximately outliers and that these relationships may 40,000 sq. km in northwest New Mexico have been an important part of the raison and adjacent parts of Colorado, Utah, and d’eˆtre of Bonito-style architecture in both Arizona. The basin’s climate is semiarid, areas. Neither the precise nature of the with high diurnal and annual temperature EVALUATING BONITO-STYLE ARCHITECTURE 473

FIG. 1. Location of Chaco Canyon and Classic Bonito phase outlier communities across the San Juan basin and adjacent areas (based on data from Fowler et al. 1987; Marshall et al. 1979; Marshall and Sofaer 1988; and Powers et al. 1983). 474 RUTH M. VAN DYKE variation and low, biseasonal precipita- sive, usually multiple-storied structures tion. Sedimentary deposits have been built in planned construction episodes carved by wind and water into landforms (Fig. 2). Core-and-veneer architecture is such as mesas, buttes, and canyons. This typical of great house construction (Lek- area was home to the Chaco Anasazi, who son 1984:21; Vivian and Mathews 1965). A emerged in the tenth century as heirs to number of veneers, or facing styles, are cultural traditions extending back hun- recognized (Hawley 1934, 1938; Judd 1927, dreds of years. Many of the characteristics 1964; Lekson 1984:17–19). Great houses ex- of Chacoan culture, including grayware hibit larger rooms and higher roofs than pottery; a reliance on cultigens; and ag- small house sites. Most great houses con- gregated villages with large, communal tain enclosed kivas or circular rooms con- pit structures appear as early as the Bas- structed inside rectangular rooms. There ketmaker III period (500–700 A.D.) (Wills are at least four great-house construction and Windes 1989). The earliest construc- episodes within the canyon. During the tion at Chaco Canyon great houses dates Late Pueblo I period and the first part of to the late Pueblo I period (850–900 A.D.) the Early Bonito phase, construction was (Windes and Ford 1992:77). Chacoan flo- initiated at Una Vida, Pueblo Bonito, and rescence in Chaco Canyon and the San Pen˜ asco Blanco (Lekson 1984:64–66; Juan Basin spans the Pueblo II (900–1100 Windes and Ford 1992:77). During the lat- A.D.) and early Pueblo III (1100–1140 ter part of the Early Bonito phase, addi- A.D.) periods. Researchers with the Chaco tions were made to these three structures, Center, a long-term cooperative venture and construction was begun at Hungo between the University of New Mexico Pavi, Chetro Ketl, and Pueblo Alto (Lek- and the National Park Service, have sub- son 1984:66–70). Construction reached its divided the Chacoan heyday into the zenith during the Classic Bonito phase Early (900–1040 A.D.), Classic (1040–1100 with the building of Pueblo del Arroyo A.D.), and Late (1100–1140 A.D.) Bonito and additions to Pueblo Bonito, Pen˜ asco phases (Windes 1987:244). The Bonito Blanco, Hungo Pavi, Chetro Ketl, and phases are, in general, characterized by Pueblo Alto (Lekson 1984:70–72). During construction of masonry pueblos, agricul- the Late Bonito phase, Wijiji, New Alto, tural intensification, increasing site den- Casa Chiquita, Kin Kletso, and Tsin Klet- sity, and production of decorated white- zin were built (Lekson 1984:72). These ware and corrugated grayware ceramics. buildings differ in layout from earlier Most outlying Chacoan great houses were great houses; the latter four are discrete, founded in the Classic Bonito phase. Tree- new structures exhibiting blocky McElmo- ring dates are available for sites in Chaco style masonry. The estimated floor area of Canyon and for a few outlying great Classic Bonito phase canyon great houses houses, but the dating of most outliers ranges from 8,025 m2 (Hungo Pavi) to and communities is based on surface ce- 23,395 m2 (Chetro Ketl); McElmo-style ramics. Ceramics are used as temporal great houses are much smaller, with floor markers for local phase schemes. areas ranging from 1184 m2 (Tsin Kletsin) to 1,460 m2 (Casa Chiquita) (Powers et al. Chaco Canyon 1983:Table 41). All canyon great houses are constructed to face the south/south- The Bonito phases are represented in east, a pattern common to many Anasazi Chaco Canyon by great houses (Lekson sites. Because prevailing winds are from 1984) and small house sites (McKenna and the northwest in the San Juan Basin, a Truell 1986). Great houses are large, mas- southeast exposure would have ensured EVALUATING BONITO-STYLE ARCHITECTURE 475 that trash middens in front of the pueblo nito-style architecture, following the term were downwind. A southern exposure that originated with Gladwin (1945). would have taken advantage of the low, There are hundreds of small-house sites southern winter sun angle, helping keep in the canyon dating to the Bonito phases rooms warm and light during colder sea- (Hayes et al. 1981; McKenna and Truell sons. 1986). Small-house sites are single storied, Great houses often are associated with with relatively few (i.e., less than 25) one or more great kivas—large, semisub- rooms. They differ markedly in layout terranean circular structures usually con- from great houses—small-house sites do sidered to represent communal, public, not exhibit planned construction but ap- integrative architecture. Great kivas in pear to have grown by accretion. Single or Chaco Canyon range between 14 and 19 m compound masonry walls characterize in diameter, with an average diameter of wall construction, although core-and-ve- 16 m. Great kiva size does not appear to be neer construction is occasionally present. related to associated great house size Like great houses, small-house sites have (Pearson’s r ϭ 0.107 for n ϭ 5). All Classic formalized kivas, but they may also con- Bonito phase canyon great houses except tain pit rooms of variable design. Trash Pueblo Alto and Pueblo del Arroyo are middens, but not earthworks or roads, are associated with great kivas, but not all associated with small-house sites. Small- great kivas are associated with great house sites tend to be located on the south houses. At least three great kivas in Chaco side of Chaco Canyon. Canyon are associated with small house sites; these include: Casa Rinconada, on The Outliers the south side of the canyon across from Pueblo Bonito and Chetro Ketl (Vivian and Reiter 1960); an unnamed great kiva Thanks largely to broad regional survey opposite Wijiji (Vivian 1990:294); and an efforts during the 1970s and 1980s, close to unnamed great kiva at small site 29SJ1253 100 Chacoan outliers are currently known (McKenna and Truell 1986:238; Marshall from the San Juan Basin and surrounding et al. 1979:273). areas (Fig. 2) (Fowler et al. 1987; Marshall Many canyon great houses, including et al. 1979; Marshall and Sofaer 1988; Pow- Pen˜ asco Blanco, Pueblo Alto, Pueblo Bo- ers et al. 1983). At least 73 of these are nito, and Chetro Ketl, are associated with documented as associated with communi- earthworks, formally constructed plat- ties. Outlier great houses and communi- forms or berms made of trash, construc- ties exhibit substantial spatial variability tion debris, or sterile materials (Lekson (Fig. 3). Each outlier contains, at mini- 1984:74; Windes 1987:561–667). Road seg- mum, a massive, large-roomed great ments, or cleared, linear alignments are house that usually exhibits multiple sto- associated with all Classic Bonito-phase ries and core-and-veneer masonry. One canyon great houses except Una Vida or more great kivas, earthworks, and road (Kincaid 1983; Nials et al. 1987; Roney segments may also be present. As in 1992; Vivian 1997; Windes 1987:529–555). Chaco Canyon, the term “Bonito-style ar- Ranching, erosion, and other disturbances chitecture” refers to these features collec- in the canyon may well have obscured tively (Marshall et al. 1982:1227–1230). additional road traces in some areas (Viv- Spatial and temporal variability among ian 1983). Great houses, great kivas, earth- outliers comprise the basis of Marshall et works, and road segments form the group al.’s (1982:1231) distinction between “an- of features collectively referred to as Bo- cestral” outliers, in which Bonito-style 476 RUTH M. VAN DYKE

FIG. 2. Twelve Chaco Canyon great houses (after Lekson 1984). Although the plans represent the structures’ ultimate configurations, only the Classic Bonito phase portions of Una Vida, Pueblo Bonito, Pen˜ asco Blanco, Hungo Pavi, Chetro Ketl, Pueblo Alto, and Pueblo del Arroyo were used in the analysis. architectural elements were introduced (Powers et al. 1983:Table 41). Size does not into communities that had existed for sev- correlate inversely with distance from eral previous centuries, and “scion” outli- Chaco Canyon, however. Powers et al. ers, which were established as Chacoan (1983:313–315; 344–345) defined three colonies during the Classic Bonito phase. great house size classes and used that Existing communities tend to be large and framework to argue that great houses dispersed over a wide area, whereas scion could represent a three-tiered, integrated communities are small and tend to be settlement system. concentrated around the great house. Although outlier great kiva diameters Like canyon great houses, outlier great range between 12 and 22 m, most great houses tend to be oriented toward the kiva diameters fall within the confines of south/southeast (Van Dyke 1998:219). 15 to 17 m. Several communities contain Outlier great house floor area is extremely multiple great kivas. Great kivas are more variable, ranging between 145 m2 (Half- likely to be absent in outliers within 40 or way House) and 15,030 m2 (Aztec West) 50 km of Chaco Canyon (Van Dyke 1998: EVALUATING BONITO-STYLE ARCHITECTURE 477

FIG. 3. Examples of Classic Bonito phase outlier great houses included in the analysis, illustrat- ing the structures’ spatial diversity (after Carlson 1966; Marshall et al. 1979; Martin 1936; Morris 1939; Peckham 1969; Pippin 1987; Powers et al. 1983; Sigleo 1981; Van Dyke 1999).

215–217) A Mann–Whitney U test rejected close to Chaco Canyon, suggesting that the null hypothesis of no relationship be- relationships between Chaco Canyon and tween great kiva presence/absence and outlier communities changed as distance distance from Chaco Canyon with a prob- increased from Chaco Canyon. Possibly ability value of 0.038. Outliers lacking people living relatively near to Chaco great kivas tend to be located relatively Canyon traveled to the canyon for ritual 478 RUTH M. VAN DYKE activities and hence did not need great Gladwin 1945). Chaco Canyon looms large kivas in their own communities. in popular and archaeological conscious- The presence of roads and earthworks ness in part because of the role played by is more likely at outliers in the southern Chaco in the history of the development part of the San Juan Basin (Van Dyke 1998: of Americanist archaeology. Further, can- 218–219); however, outlier surveys of the yon great houses constitute the largest 1970s and early 1980s (Marshall et al. 1979; and most concentrated appearance of Bo- Powers et al. 1983) predate the systematic nito-style architecture. Given these fac- recognition and recording of outlier earth- tors, it is not surprising that Chaco Can- works, and much of the work that has yon is usually envisioned as the hub from recognized these features (Fowler et al. which Bonito-style architecture emanated. 1987; Marshall and Sofaer 1988) has been However, there are several possible expla- concentrated in the southern part of the nations for the spread of the architecture. basin. Stein and Lekson (1992) have sug- The appearance of Bonito-style architec- gested that modifications such as roads ture at outliers as the result of direct in- and earthworks are ubiquitous to outliers volvement on the part of a centralized, and represent a ritual Chacoan landscape Chacoan entity is only one of these. Alter- focused on great houses. Where they are natively, local people could have been present, earthworks define the space in emulating Bonito-style architecture they front of the great house. This barrier is saw either at Chaco or in neighboring symbolic rather than practical, however. communities, or Bonito-style architecture The recognition that roads link some of could have been imported into Chaco the great house communities with the Canyon from its origins in outlier commu- canyon was part of the impetus for the nities. The first two possibilities are con- recognition of a Chacoan system. With the sidered here in light of a range of compet- exception of the North and South Roads, ing models for the Chacoan system. many roads can be located only in short segments near outlier communities, call- Competing Explanations for the Chacoan ing into question the assumption that the System roads define and link a Chacoan system (Roney 1992). Contemporary explanations for the con- The pattern of large, Bonito-style struc- struction of Bonito-style architecture in tures with associated communities, great Chaco Canyon and in outlier communities kivas, and road segments does not stop at are numerous and contradictory (Sebas- the edge of the San Juan Basin. There are tian 1992:82–97; Vivian 1990:391–419). Al- large sites with Bonito-style structures in though most models were developed to northeast Arizona and southeast Utah; the explain canyon rather than outlier com- relationships between these sites to Chaco munity dynamics, the models do have ei- Canyon and to outliers within the basin ther inferred or directly stated implica- are subject to debate. tions for relationships between the outliers and Chaco Canyon. A Chacoan EXPLAINING THE DISTRIBUTION origin for outlier great houses is in accord OF BONITO-STYLE ARCHITECTURE with models proposed by Vivian (1989, 1990) and Wilcox (1993). In both of these Relationships between canyon and out- scenarios, people from Chaco are directly lier Bonito-style architecture have long responsible for the construction of Bonito- been a focus of archaeological inquiry style architecture outside Chaco Canyon. (Roberts 1932; Martin 1936; Morris 1939; Vivian (1989, 1990), building upon the EVALUATING BONITO-STYLE ARCHITECTURE 479 work of Kluckhohn (1939), contends that ered the canyon as the center of a redis- great houses and small-house sites in tribution network for subsistence goods; Chaco Canyon represent two separate but the system protected the San Juan Basin coexistent egalitarian cultural traditions. Anasazi against crop shortfalls. Outliers Grounded in principles of dualism, the were linked to the canyon in a relation- San Juan tradition gave rise to great ship that was primarily economic, al- houses organized in terms of a rotating though it may have been legitimated by a sequential hierarchy (after Johnson 1982). ritual veneer. Adaptation to the environ- By contrast, within the small-house site ment is seen as the motivating force be- Cibola tradition, social relationships were hind the evolution of a Chacoan regional organized in terms of lineages. For Vivian, system. A ranked social hierarchy involv- outliers were established when represen- ing managerial elites who control eco- tatives from canyon groups moved into nomic and ritual activities is envisioned. the basin in search of better farmland, In Judge’s later (1989) “pilgrimage fair” sometimes joining existing communities. model, redistribution is held to be respon- Wilcox (1993) proposes a military Cha- sible for the initial development of Chaco, coan state replete with social hierarchy but control of turquoise is the key to the and economic and religious centralization. canyon’s growth into a ritual center under Canyon great houses are barracks, and the purview of canyon-based elites. outliers are military installations estab- Toll (1984, 1985) sees Chaco as an egal- lished to obtain agricultural tribute from itarian system characterized by the move- and retain control over the countryside. ment of goods to counter environmental Great kivas are focal points for tribute col- shortfalls. Periodic gatherings took place lection, and roads are for the efficient at Chaco without an aegis of elites. movement of Chacoan armies as they go “Alarming” quantities of smashed utility about this business. However, there are vessels in the earthworks at Pueblo Alto serious material problems with envision- (Toll 1985:185) are the detritus of cyclical ing Chaco as a state. The highly visible feasting (Toll 1985:369-406). Large quanti- hallmarks of centralization present in ties of trachyte-tempered pottery traveled other ancient states, including a unified from the Chuskan slope to Chaco Canyon, currency, standardized weights and mea- but Chaco appears to have been primarily sures, and a system of record keeping a terminus rather than a redistribution (Runciman 1982:361), are entirely lacking center for the pottery (Toll 1984:130) as in the Chacoan system. well as Chuskan lithic materials (Cameron In contrast with the direct canyon ori- 1984; Jacobson 1984). gins for outliers suggested by the work of At the opposite end of the explanatory Vivian and Wilcox, a more ambiguous spectrum, the models of Mathien (1993), view is proffered in Chaco Center and Sebastian (1992), and Stein and Lekson related models (Judge 1979, 1989; Judge et (1992; Lekson 1991) allow for local devel- al. 1981; Schelberg 1984; Toll 1984, 1985). opment of Bonito-style architecture. The For these authors, canyon-directed outlier canyon is the primary focus of Mathien’s construction does not seem mandatory, and Sebastian’s work, but both authors yet direct interaction between the outliers suggest outlier Bonito-style architecture and Chaco Canyon remains an important might reflect aggrandizing strategies un- component of the outliers’ raison d’etre. dertaken by local leaders seeking to estab- Chaco Center archaeologists (Judge 1979; lish alliances. Following the work of C. Judge et al. 1981; Powers 1984; Powers et Smith (1976a, 1976b), Mathien (1993) fo- al. 1983; Schelberg 1984) originally consid- cuses on exchange as a means of under- 480 RUTH M. VAN DYKE standing Chacoan socioeconomic organi- sazi pattern. Bonito-style elements are zation. Prior to 1050 A.D., exchanged seen as symbolic links connecting outliers items were not rare but were sometimes in a religious koine that crosscuts ethnic or unpredictable in availability. The subse- linguistic boundaries. Stein and Lekson quent introduction of scarce prestige conceive of a Chacoan ritual landscape in items led to a bounded system controlled which great houses, great kivas, road seg- by a “big man.” In a Chacoan bounded ments, and encircling earthworks had system, scarce and critical resources could symbolic meaning that was replicated have been exchanged between parties in a over a wide area. The spatial extent and situation similar to a Melanesian “Kula meaning of Bonito-style architecture is ring” (C. Smith 1976b). Control of water not necessarily congruent with the extent and better farmland would have given and the nature of Chacoan regional inter- some basin Anasazi economic power dur- action (Lekson 1991:32). ing environmentally difficult times. Those in less fortunate circumstances could have Style and Social Identity contributed their labor to construct roads or great houses in exchange for food. By definition, Bonito-style architecture Nascent Anasazi social differentiation in Chaco Canyon and in outlier commu- gave rise to client–patron relationships in nities involves a number of shared, easily Chaco Canyon in a model developed by recognizable architectural characteristics. Sebastian (1992). Correlations between es- A diverse range of social relationships be- timated corn crop yields and great house tween outlier and canyon great houses construction dates form the basis of a sce- may be represented by the stylistic affin- nario in which the building of the canyon ities among the structures. The compli- great houses assisted emerging canyon cated connections between style and so- elites in solving problems of leadership. cial identity have been the focus of Architecture, roads, and earthworks uti- considerable recent investigation (Carr lized labor debts, constituted physical ev- and Neitzel 1995; Conkey and Hastorf idence of leaders’ power, and ultimately 1990; Hegmon 1992; Stark 1998). Stylistic became a medium of competition be- choices may be emblemic, intentional at- tween elites seeking to impress and attract tempts to communicate group member- followers. Outlier communities represent ship (Wiessner 1983, 1984, 1985, 1990) or indigenous developments of patron–cli- enculturative, passive by-products of ent relationships similar to those evolving shared learning frameworks (Sackett 1982, in Chaco Canyon, and canyon elites estab- 1985, 1990). Bonito-style architecture obvi- lished alliances with outlier leaders. ously has a great deal of communicative, Stein and Lekson (1992) directly address symbolic potential, but any investigation the appearance of Bonito-style architec- into the nature and directionality of pos- ture in outliers with a model that can in- sible symbolic messages also must ad- corporate a diverse spectrum of canyon– dress the critical issue of Chaco Canyon outlier relationships. Communities dating involvement in the construction of the ar- from the eleventh and twelfth centuries chitecture in outlier communities. First, with Bonito-style architecture are found the nature of social connections between outside the San Juan Basin over a large canyon and outlier Bonito-style builders area difficult to view as one coherent sys- must be assessed; only then can possible tem (Lekson 1991:46). Hence, characteris- communicative or symbolic purposes be tic outlier attributes may simply represent examined. In other words, the encultura- a late Pueblo II–early Pueblo III pan-Ana- tive information contained in the architec- EVALUATING BONITO-STYLE ARCHITECTURE 481 ture must be teased out before its sym- possible canyon–outlier builder relation- bolic content can be meaningfully ships are organized into two camps: infor- examined. mation about the construction of Bonito- For Carr (1995a, 1995b), visibility is the style architecture may have traveled key to the distinguishing among commu- directly from Chaco Canyon to outlier nicative and enculturative stylistic at- communities, or people in outlier commu- tributes. Artifacts and features with high nities may have attempted to emulate the visibility, such as polychrome bowls or architecture they saw in Chaco or in monumental architecture, have a high neighboring communities. The first possi- communicative potential. In contrast, pat- bility is considered under the rubric of terning among low-visibility artifacts and directed construction; the second is consid- features, such as plainware vessels and ered under local construction. A third pos- domestic architecture, is likely to reflect sibility—that Bonito-style architecture shared enculturative backgrounds. Fol- originated in outlier communities and lowing Carr’s (1995b) logic, because Bo- traveled to Chaco Canyon—is not ad- nito-style architectural forms are highly dressed in this study. The reasons behind visible, they are likely to have had com- the spread of Bonito-style architecture municative or emblemic purposes. How- and the mechanisms and directionality of ever, this does not preclude the possibility the flow of information are of consider- that enculturative information also is con- able importance for evaluating some of tained in the structures. Regardless of the the existing Chacoan explanatory models high-visibility communicative potential of reviewed above. Bonito-style architecture, it should be pos- Directed construction. If the information sible to look at the low-visibility aspects of for construction of Bonito-style architec- construction to see if the builders of can- ture at outliers emanates from a central, yon and outlier structures relied upon Chacoan source, several kinds of relation- shared learning frameworks. The French ships between the canyon and the outliers school of technological style, or technologie are possible. First, the architecture could (Lechtman 1977; Lemonnier 1986), avoids have been built by migrants from Chaco the problematic issue of the intentionality Canyon, a` la Vivian (1990). In this case, of stylistic communication by focusing not Bonito-style architecture represents group on the finished product but rather on the identification with Chaco Canyon that learning frameworks and sequence of may or may not be conscious. Colonists specific creative activities used in produc- probably would have maintained social tion. Learning-framework information ties with Chaco and may or may not have should be reflected in the internal con- been considered independent of Chacoan struction and organization of Bonito-style social authority. If Bonito-style architec- architectural features. ture at outliers represents the intrusive The many possible relationships be- presence of migrants from Chaco Canyon tween the builders of canyon and outlier into extant basin communities, the archi- Bonito-style architecture can be organized tecture reflects a specific and probably as an explanatory continuum. In the en- conscious Chacoan aesthetic juxtaposed suing analysis, I focus on the opposing with local Anasazi traditions. ends of this continuum. Although this is Second, Bonito-style architecture in admittedly a somewhat reductionist ap- outlier communities might represent in- proach, it does facilitate the empirical formation conveyed by specific individu- evaluation of possible relationships be- als from Chaco Canyon who did not make tween canyon and outlier builders. Thus, the community their permanent home. 482 RUTH M. VAN DYKE

Masons may have traveled from the can- rapidly throughout the San Juan Basin yon to the outliers for the express purpose and adjacent areas in a “down-the-line” of designing and constructing Bonito-style manner. Ethnographic sources suggest architecture in the communities. Masons that large-scale construction projects are might have been members of a Chacoan often communal endeavors for social, if elite or attached specialists serving Cha- not practical, reasons (Tuzin 1976). coan elites (Ames 1995). Ancient political In all these scenarios, the techniques of authorities often established military and Bonito-style architecture were not com- governmental installations in high places, mon knowledge, but were known only to constructing visually massive, vertically certain individuals or small groups who impressive buildings accompanied by art- presumably plied their trade primarily in work that dramatized the regime’s inevi- Chaco Canyon. If knowledge about Bo- tability and continuity with the natural nito-style construction was restricted, it is order. If Bonito-style architecture is a likely that individuals or factions either physical manifestation of Chacoan domi- from the canyon or from the local commu- nation, we might expect the kinds of spa- nity attempted to use exclusive access to tial practices and representations docu- that knowledge in the establishment or mented in other parts of the world as legitimation of power. concomitant with political expansion (Al- Local construction. At the opposite end cock 1993; A. Smith 1996). A nascent Cha- of the spectrum of explanations for spatial coan polity placing its stamp upon local similarities between the canyon and the communities coincides with Wilcox’s outliers is the idea that local factions (1993) concept of a Chacoan state. sought to emulate Bonito-style architec- Chacoan masons may have constituted ture visible either among their neighbors a relatively independent guild or group or at Chaco Canyon. A possible rationale that possessed exclusive knowledge for local emulation of Chacoan architec- sought by locals. These masons could tural forms is suggested by Renfrew and have been invited to ply their trade in Cherry’s (1986) peer–polity interaction outlier communities, or local would-be model. According to this model, in the builders could have traveled to Chaco to absence of a strong central political au- consult with them. Information about Bo- thority, neighboring communities observe nito-style construction might have been and compete with one another. A well- available to locals who traveled to Chaco developed set of prestige symbols seen in to participate in ritual gatherings, as envi- one community will appeal to neighbors, sioned by Toll (1985) and Judge (1989). If who will attempt to imitate it if their own Chacoan elites sought to expand an elite symbolic system is less developed and network to include the leaders of outlier does not conflict. Renfrew and Cherry communities, perhaps masons were sent (1986:8) call this process “symbolic en- to work for those leaders on a part-time or trainment.” Kintigh (1994:134-136) sug- a full-time basis as part of the consolida- gests that the appearance of quasi-Bonito- tion of network ties. style architectural elements such as large, Neighboring communities might have unroofed great kivas at post-Chacoan learned about Bonito-style architecture sites stems from peer–polity interaction in from one another through cooperative, the absence of a central, Chacoan influ- rather than competitive, activities. If labor ence. Prestige in the Cibola (Zuni) area for large-scale construction projects was during the thirteenth century A.D. was drawn from neighboring communities, ar- gained by emulating the Chacoan past. chitectural information could have moved Could prestige also have been gained by EVALUATING BONITO-STYLE ARCHITECTURE 483 emulating the Chacoan present? Applica- also be true—if outlier great houses were tion of the peer-polity interaction model constructed under entirely local auspices, to the Chacoan era is predicated on a they should exhibit great variability, dif- weak central role for the canyon itself. The fering both from canyon great houses and peer-polity interaction model is cited by from each other. As discussed above, be- Blanton et al. (1996:5) as an example of a cause external, high-visibility similarities network strategy to power. Network strat- are likely to contain communicative infor- egies, in which neighboring elites ex- mation, they are not useful for purposes of change but also compete with one an- analysis. However, low-visibility aspects other, are similar to what both Mathien of construction should contain encultura- (1993) and Sebastian (1992) have in mind tive information. Similarities among inter- when they discuss interaction between nal, low-visibility attributes of great house Chaco Canyon elites and outlier leaders. architecture are likely to indicate shared Construction of Bonito-style architecture learning frameworks. Local emulation in an attempt to foster prestige and legit- should be distinguishable from directed, imate inequalities within local outlier Chacoan construction in that although the communities could fit either of these mod- general form and appearance of Bonito- els. style architecture might be emulated, spe- In another twist on local emulation, Bo- cific, internal precepts of the style cannot. nito-style architecture could have been If Bonito-style architecture were con- built in outlier communities by Chacoan structed under local auspices, internal de- migrants who did not possess technical tails of outlier great houses should differ architectural knowledge but who sought from each other and from internal details to emulate the canyon style to emphasize of canyon great houses. Furthermore, if their origins or ties with the canyon. Peo- architectural information was transmitted ple who emigrate are usually not the peo- in a down-the-line fashion from Chaco ple who hold power in the society of their Canyon, internal variable diversity should origin. Perhaps some outlier communities increase with increased distance from were founded by disenfranchised Cha- Chaco Canyon. coan people who attempted to replicate Following these assumptions, seven the great houses of the canyon but who Classic Bonito phase Chaco Canyon great were not privy to specific tenets of canyon houses were used as the template against architecture. If Bonito-style architecture which contemporaneous outlier great represents a communal, symbolic state- houses were compared. Using the fine- ment of affinity with Chaco Canyon or grained temporal resolution afforded by with neighboring communities, this fits tree-ring dates, only those canyon great the perspective of Bonito-style architec- house configurations dating to the second ture as ideational bond as put forth by half of the eleventh century were included Stein and Lekson (1992). in the analysis. Five internal, or low-visi- bility, architectural variables characteris- ARCHITECTURAL ANALYSIS tic of canyon great houses were chosen for comparison. These include core-and-ve- How do directed and local construction neer walls, banded facing, symmetry, ele- differ archaeologically? If outlier great vated kivas, and kiva/room ratio (Table 1). houses were built according to directed Core-and-veneer walls and banding fac- Chacoan information, then they should ings are particularly likely to represent resemble the great houses of Chaco Can- common learning frameworks. Symmetry, yon and each other. The converse should elevated kivas, and kiva/room ratio com- 484 RUTH M. VAN DYKE

TABLE 1 Classic Bonito-Phase Canyon Great Houses and Variables Used in the Analysisa

Great house Core-and-veneer Banding Symmetry Elevated kiva K:R ratio

Chetro Ketl 1 1 1 1 0.028 Hungo Pavi 1 1 1 1 0.007 Pen˜ asco Blanco 1 1 1 1 0.033 Pueblo Alto 1 1 1 0 0.115 Pueblo Bonito 1 1 1 1 0.047 Pueblo del Arroyo 1 1 1 1 0.052 Una Vida 1 1 1 1 0.038

a Based on data in Lekson (1984); 0 ϭ absent, 1 ϭ present. municate aspects of social organization. the five internal variables should be sim- All five are aspects of construction that ilar in their distribution across space. could not be easily emulated by outsiders They should co-occur in a majority of lacking access to the specifics of great- cases. If the converse is true, and outlier house construction or layout. The choice Bonito-style architecture is a result of local of internal variables was mitigated by the emulation, then these variables should ex- fact that the study dealt mainly with sur- hibit great disparity. Little patterning or face data. Thus, potential internal vari- co-occurrence should be evident. If com- ables such as kiva features, or technical munal labor practices led to Bonito-style aspects of construction such as mortar architectural information being passed composition or stone shaping, could not from community to community, then di- be considered given the constraints of the versity among internal variables is ex- data. Following a review of the data set pected to decrease as distance from Chaco and methods used in the analysis, each Canyon increases. variable is explained and its results dis- In order to identify spatial patterning, cussed in sequence below. internal variables were compared with A total of 61 Classic Bonito phase great each other and with region, distance in houses from 55 outlier communities pro- kilometers from Chaco Canyon, and vided a data set for examining the distri- great-house area. Region is a categorical bution of internal variables (Table 2; Fig. variable assigned to each great house 3). Where excavation data allowed great- based on location. The greater San Juan house temporal components or construc- Basin may be divided physiographically tion episodes to be separated (e.g., into 20 or more subregions (e.g., Marshall Guadalupe, Lowry, Wallace), only compo- et al. 1979:20-22), but this level of compart- nents from the Classic Bonito phase were mentalization yields outlier counts for used in the analysis. Most unexcavated each area that are too small for meaning- great houses were dated using midden ful comparisons of the variables. Thus, surface ceramics, although tree-ring dates three general regions were defined by were available in a few cases (e.g., Kin means of two east–west boundaries drawn Bineola, Kin Klizhin). Great houses with approximately at the north edge of the ambiguous architectural definition or with Chaco Plateau and the south edge of the substantial later components were omit- South Chaco Slope (Fig. 4). The north (Re- ted in order to minimize sources of error. gion 1), central (Region 2), and south re- If outlier Bonito-style architecture is a re- gions (Region 3) contain 13, 30, and 18 sult of directed, canyon influences, then outlier great houses respectively. A sec- TABLE 2 Classic Bonito-Phase Outlier Great Houses and Variables Used in the Analysis

Core-and- Elevated Kiva:room Great housea Region Distanceb Areac veneerd Bandingd Symmetryd kivad ratio Major references

1. Allentown 3 163.8 1813 1 — — 0 0.053 Marshall et al. (1979); Powers et al. (1983); Roberts (1939, 1940) 2. Andrews 3 76.3 470 1 0 1 1 0.238 Marshall et al. (1970); Van Dyke (1999) 3. Bee Burrow 2 30.9 430 1 0 1 0 0.125 Marshall et al. (1979); Powers et al. ARCHITECTURE BONITO-STYLE EVALUATING (1983) 4. Bluff 1 190.1 — 1 0 0 0 — Cameron (1996); Cameron and Lekson (1996) 5. Casa Escondida 2 17.6 346 1 1 — 0 0.143 Marshall and Sofaer (1988) 6. Casamero 3 73.0 365 1 1 0 0 0.143 Harper et al. (1988); Marshall et al. (1979); Powers et al. (1983); Sigleo (1981) 7. Cerro Prieto 3 — 1758 1 — 0 0 0.025 Fowler et al. (1987) 8. Chimney Rock 1 147.8 1461 1 — 1 0 0.036 Eddy (1977); Jeancon (1922); Jeancon and Roberts (1923); Powers et al. (1983) 9. Coolidge 3 79.5 731 1 — 1 0 0.190 Marshall et al. (1979) 10. Cove 1 133.3 300 0 — — — — Reed and Hensler (1998) 11. Coyotes Sing Here 3 66.2 349 1 — 1 0 0.133 Marshall et al. (1979) 12. Cuatro Payasos 2 35.0 378 1 — 0 0 0.071 Marshall and Sofaer (1988) 13. Dalton Pass 2 53.7 753 1 — — 0 0.214 Powers et al. (1983) 14. El Rito 3 93.9 — 0 0 — 0 0.073 Allan and Gauthier (1976); Powers et al. (1983) 15. Escalante 1 179.5 424 1 0 0 0 0.040 Hallasi (1979); Powers et al. (1983) 16. Escalon 2 34.8 268 1 1 — 0 0.050 Marshall and Sofaer (1988); Marshall (1994a) 17. Fort Wingate 3 94.3 844 0 0 0 0 0.091 Peckham (1958); Marshall et al. (1979) 18. Gonzales Well 3 140.5 1129 1 — — 0 — Fowler et al. (1987) 19. Great Bend 2 49.4 409 1 — 0 0 0.375 Marshall et al. (1979) 20. Greenlee 2 17.1 250 1 0 0 0 0.067 Powers et al. (1983) 21. Grey Hill Spring 2 51.1 188 1 1 0 0 0.200 Marshall et al. (1979); Powers et al. 485 (1983) 486 TABLE 2—Continued

Core-and- Elevated Kiva:room Great housea Region Distanceb Areac veneerd Bandingd Symmetryd kivad ratio Major references

22. Guadalupe 3 95.3 335 1 1 1 0 0.091 Baker (1984); Irwin-Williams and Baker (1991); Marshall et al. (1979); Pippin (1987); Powers et al. (1983); Washburn (1974) 23. Haystack 3 84.2 700 1 — 1 1 0.200 Marshall et al. (1979); Powers et al. (1983) 24. Hinkson Ranch 3 — 265 1 — 1 0 — Fowler et al. (1987) 25. Ida Jean 1 148.0 542 1 1 1 0 0.074 Powers et al. (1983) 26. Kin Bineola 2 16.6 3360 1 0 1 1 0.050 Marshall et al. (1979); Powers et al.

(1983) DYKE VAN M. RUTH 27. Kin Hocho’i 3 129.6 697 1 — 0 0 0.138 Fowler et al. (1987) 28. Kin Klizhin 2 11.5 350 1 1 1 1 0.188 Marshall et al. (1979); Powers et al. (1983) 29. Kin Nizhoni 3 81.2 Marshall et al. (1979) Upper 330 1 1 0 0 0.200 Powers et al. (1983) Lower 305 1 0 1 0 0.200 30. Kin Ya’a 2 50.4 864 1 1 1 1 0.100 Marshall et al. (1979); Powers et al. (1983) 31. Lowry 1 189.5 562 1 1 1 0 0.125 Kendrick and Judge (1996); Martin (1936); Powers et al. (1983) 32. Morris 39 1 91.9 403 1 0 1 0 0.050 McKenna and Toll (1991); Morris (1939); Powers et al. (1983) 33. Morris 41 1 104.0 1069 1 1 0 0 0.040 McKenna and Toll (1991); Marshall et al. (1979); Morris (1939) 34. Muddy Water 2 46.8 Marshall et al. (1979); Powers et al. LA 10959 367 1 — — 0 0.111 (1983) LA 10716 387 1 — — 0 0.067 LA 17257 418 1 1 — 1 0.286 35. Navajo Springs 3 189.6 760 1 0 — 0 0.063 Graves (1990); Warburton and Graves (1992) 36. Newcomb 2 81.2 448 1 — 0 0 0.077 Marshall et al. (1979) 37. Padilla Well 2 8.6 200 1 — 0 0 0.000 Marshall and Sofaer (1988) 38. Peach Springs 2 55.8 752 1 1 1 1 0.033 Powers et al. (1983) 39. Pierre’s 2 12.1 Powers et al. (1983); Harper et al. House A 265 1 0 0 0 0.200 (1988) House B 336 1 1 0 0 0.077 El Faro 505 1 0 0 0 0.111 40. Pueblo Pintado 2 51.0 4415 1 1 0 0 0.168 Marshall et al. (1979); Powers et al. (1983) 41. Red Willow 2 68.6 1036 1 — 0 1 0.300 Marshall and Sofaer (1988) 42. Salmon 1 72.6 3452 1 1 1 1 0.010 Irwin-Williams and Shelly (1980); VLAIGBNT-TL ARCHITECTURE BONITO-STYLE EVALUATING Irwin-Williams et al. (1975); Powers et al. (1983) 43. San Mateo 3 83.5 1765 1 0 0 1 0.048 Marshall et al. (1979) 44. Sanostee 2 66.2 Marshall et al. (1979) House 1 346 1 — 1 0 0.222 House 2 72 1 — 1 0 0.000 45. Section 8 2 47.9 290 1 0 1 0 0.400 Marshall et al. (1979) 46. Skunk Springs 2 77.1 1530 1 — 1 1 0.095 Marshall et al. (1979); Peckham (1969); Powers et al. (1983) 47. Squaw Springs 1 109.8 440 1 0 0 0 0.176 Marshall et al. (1979); Powers et al. (1983) 48. Standing Rock 2 43.3 415 1 0 1 0 0.000 Marshall (1994b); Marshall et al. (1979); Powers et al. (1983) 49. Sterling 1 70.0 — 1 0 — 0 0.040 McKenna and Toll (1991); Powers et al. (1983) 50. Toh La Kai 2 99.8 818 1 1 0 1 0.065 Marshall et al. (1979) 51. Twin Angels 1 60.0 364 1 — 0 0 0.118 Carlson (1966); Powers et al. (1983) 52. Upper Kin Klizhin 2 15.0 351 1 — 0 1 0.040 Powers et al. (1983) 53. Village of the 3 120.2 349 1 1 1 1 0.111 Powers et al. (1983); Roberts (1932) Great Kivas 54. Wallace 1 163.5 522 0 0 1 0 0.111 Bradley (1974, 1993); Powers et al. (1983) 55. Whirlwind 2 48.5 378 1 0 0 1 0.048 Marshall et al. (1979)

a Numbers correspond to Fig. 4. b Linear kilometers to Pueblo Bonito. c Total roofed area in two dimensions (square meters). d ϭ ϭ

0 absent; 1 present. 487 488 RUTH M. VAN DYKE ond categorical variable, north/south, also Core-and-Veneer Walls was used to compare the outliers in the San Juan/Animas Valley and points fur- Core-and veneer walls are described by ther north (Region 1) to those in the cen- Lekson (1984:21) as “hallmark(s) of Cha- tral and south part of the San Juan Basin coan building” and are present at all (Regions 2 and 3). Distance is a metric vari- seven Classic Bonito phase canyon great able referring to kilometers between each houses. Core-and-veneer walls consist of outlier great house and Chaco Canyon, for two parallel simple walls separated by a which Pueblo Bonito is used as the arbi- space filled with various materials. For trary central point. The metric variable “fill core” core-and-veneer walls, this area was assigned to each great house on space is filled with mud, earth, trash, or the basis of two-dimensional roofed area rubble placed in the cavity after several in square meters. Canyon great houses courses of wall were completed. For “solid are at the large end of the great-house size core” core-and-veneer walls, rubble is laid continuum, so inclusion of this measure between the two facing walls as the fac- helps assess whether any internal vari- ings are built. Chacoan walls rely on ables are common only to very large great width, rather than other techniques, such houses. Measurements were made on as buttressing, for stability. The core-and- great-house drawings scanned or digi- veneer technique may have been devel- tized into AutoCAD. Plazas were excluded oped to facilitate the construction of very wide ground floor walls that could bear because not all great houses have walled the load of the multiple stories above; the or well-defined plazas and plazas can in- widths of upper-story walls were succes- flate the areal measurements for great sively reduced (Lekson 1984:15). Although houses without addressing the reasons core-and-veneer construction provides behind this difference. Multiple stories functional benefits enabling the creation were also excluded for the purposes of of load-bearing walls, the technique is not this analysis because outlier upper story a necessary requirement for construction estimates based on surface data are unre- of multiple-story structures. Early, multi- liable and quite variable. ple-story construction portions of Una Comparisons of dichotomous categori- Vida, Pen˜ asco Blanco, and Pueblo Bonito cal variables, including presence/absence were built of simple and compound ma- of core-and-veneer walls, banding, sym- sonry (Judd 1964:58; Lekson 1984:21–22, metry, and elevated kivas, were made us- 90–91, 104). Core-and-veneer construction ing Fisher’s exact test. Comparisons of the is considered an internal architectural polytomous categorical variable, region, characteristic because wall interiors were made using Pearson’s chi-square would not have been readily visible to test. Comparisons of metric variables, in- casual passers-by. It is unlikely that core- cluding distance from Chaco Canyon, and-veneer construction was spontane- great-house area, and kiva/room ratio, ously independently invented throughout were made using the Mann–Whitney U the San Juan Basin. If core-and-veneer test and Kruskal–Wallis one-way analysis construction is considered evidence of a of variance. Probabilities resulting from relationship with Chaco, then the pres- these comparisons are presented in Table ence of this technique at nearly all of the 3; p values of .05 or less are considered outlier great houses in the sample consti- statistically significant. Selection, analysis, tutes support for directed, Chacoan con- and results for each of the five internal struction. variables are discussed below. Despite the reduced condition of some EVALUATING BONITO-STYLE ARCHITECTURE 489

FIG. 4. Distribution of five internal variables among Classic Bonito phase outlier great houses included in the analysis. 490 RUTH M. VAN DYKE

TABLE 3 Probability Values for Cross-Tests of Internal Variables

Core-and- Elev. Region N/S Dist. Area veneer Band. Symm. kiva

Core-and-veneer 0.113b 0.196a 0.037e 0.930c ———— Banding 0.733b 1.000a 0.838c 1.000c 0.230a ——— Symmetry 0.600b 0.742a 0.419c 0.845c 1.000a 0.724a —— Elevated Kiva 0.475b 0.713a 0.292c 0.009e 1.000a 0.476a 0.754a — Kiva:room ratio 0.148d 0.050e 0.524d 0.378d 0.886c 0.831c 0.806c 0.528c

a Probability value for Fisher’s exact test. b Probability value for Pearson’s chi-square test. c Probability value for Mann–Whitney U test. d Probability value for Kruskal–Wallis one-way ANOVA. e Statistically significant value (all are Mann-Whitney U tests). of the great houses in the sample, core- Rito) and 163.6 km (Wallace) from Pueblo and-veneer construction is recorded as Bonito. These findings would seem to present at all except four—El Rito, Fort support the directed, Chacoan construc- Wingate, Cove, and Wallace (Fig. 4; Table tion of most Classic Bonito phase outlier 2). El Rito, Fort Wingate, and Wallace are great houses, with the exception of a all multiple-story structures. El Rito ex- handful in the distant hinterlands. How- hibits masonry similar to Hawley’s (1934, ever, no significant associations were 1938) Type 1 found at Una Vida (Powers et found between core-and-veneer masonry al. 1983: 216, 222–224), suggesting that the and any other variables. structure actually dates to the Early Bo- nito–phase. The only visible masonry at Banded Facing Style Fort Wingate is compound (Marshall et al. 1979:155), but the reduced state of the rub- Classic Bonito phase masonry has a ble mound makes it impossible to say with banded appearance created by the use of certainty that core-and-veneer masonry is alternating bands of thin, tabular sand- absent. Both Cove and Wallace have firm stone and thick, blocky sandstone. The occupations during the Classic Bonito– use of sandstone by canyon and outlier phase but lack core-and-veneer masonry. great house builders reflected the mate- At Cove, the rubble mound is reduced, rial’s availability and its functional and but Reed and Hensler (1998) nevertheless aesthetic properties. In Chaco and many were able to identify compound masonry. outlier locations, both types of sandstone Excavations at Wallace have determined are locally available in Tertiary or Creta- that Pueblo II rooms were built of simple ceous sedimentary exposures. The soft and compound masonry (Bradley 1974; sandstone is easily worked, and both soft Powers et al. 1983:163). When the pres- blocks and hard tabular pieces are easily ence/absence of core-and-veneer ma- stacked, facilitating construction and sta- sonry was compared with the other vari- bility of tall, multistoried walls. Sandstone ables, it was found to be significantly facings would have provided smooth sur- associated with the dependent variable faces and sharp angles for plastering, and distance (Table 3), reflecting the fact that tabular sandstone is especially conducive the four cases where core-and-veneer ma- to producing the banded appearance of sonry is absent fall between 93.9 km (El much Chacoan architecture. EVALUATING BONITO-STYLE ARCHITECTURE 491

Banding is exhibited by Types 2 and 3 in visibility attribute because most canyon the widely adopted Judd/Lekson facing- great-house walls were probably coated style typology (Lekson 1984:17–19). Band- with mud or plaster both inside and out ing is not a necessary outcome of building (Lekson 1984:29). Excavated outlying great with coursed sandstone, as it does not houses commonly exhibit interior wall characterize all canyon great-house ma- plaster [e.g., Aztec West (Morris 1928:120, sonry. However, banded veneers may 272–273, 289, 294], Bis sa’ani (Breternitz et have provided some functional benefit. al. 1982:187, Table 29], Casamero (Sigleo Core-and-veneer masonry with Type 2 1981:3), Guadalupe (Pippin 1987), Morris facing represented a technological im- 39 (Morris 1939:52), and Salmon Ruin (Ir- provement over the earlier, simple or win-Williams et al. 1975:52, 120)]. Exam- compound Type 1 masonry because the ples of exterior wall plaster are rare due to use of more spalls than mortar in the the vagaries of preservation, but it is joints strengthened walls by decreasing known from buried or remodeled great- mortar exposure and by transferring most house walls at Chetro Ketl (Lekson 1984: of the load of the wall from the core to the 173) and Bis sa’ani (Breternitz et al. 1982: facing. Increased stone-on-stone contact 185, 187). Ethnographic evidence also resulted in stronger walls (Lekson 1984:23; supports the practice of plastering exte- Reiter 1933:67). Stone-on-stone contact is rior pueblo walls (Mindeleff 1989: maximized in Type 4 masonry, however, 137–138). Mindeleff (1989:45) observes that which lacks the banded appearance of Zuni walls are plastered but not banded Types 2 and 3. and adds that “it is not to be expected that Mindeleff (1989:140) suggested banded walls would be carefully constructed of facings were inspired by the juxtaposed banded stonework when they were to be seams of tabular and blocky sandstone in subsequently covered with mud.” The fact the walls of Chaco Canyon or were an that canyon great houses do exhibit care- accidental result of using the material ex- ful and aesthetically pleasing banded ve- pediently at hand. However, Powers et al. neers suggests that banding had stylistic (1983:317) note that veneer variability ev- meaning for the builders. If exterior walls ident through time within one structure or were plastered, the presence of banding between adjacent structures indicates that might reasonably be assumed to indicate building material did not determine style. shared knowledge of a specific Chacoan A raw materials-based interpretation for construction style, whether or not it was banding does not explain why banding imbued with additional symbolic value. was intentionally created at some great The presence/absence of banded ve- houses where suitable tabular and blocky neers was tabulated for 38 outliers where sandstone were not readily available. At masonry is exposed. Type 2 or Type 3 Casamero (Sigleo 1981:3), limestone masonry was considered synonymous blocks and chinking were used to produce with banding. Data sources are sometimes banding, suggesting that the creation of ambiguous, describing facings as banded veneers must have been at least “coursed” or “Chaco-style.” Pictures were partially an aesthetic choice. Banding may consulted when available. In cases where have been intentionally created to in some ambiguity could not be resolved, the great way symbolize “Chaco-ness” or “great house was excluded from the analysis. house,” or the style may have been part of Where bands were intentionally created the received knowledge of Chacoan build- through the use of other raw materials ers. besides sandstone, banding was counted Banded veneers are an excellent low as present. Banding is present at 20 and 492 RUTH M. VAN DYKE absent at 18 outlier great houses in the and possibly the presence of a rotating data set (Fig. 4; Table 2). However, no sequential hierarchy as a mechanism for significant associations were found be- organizing tasks, decision making, ritual, tween banding and other internal vari- labor, or other aspects of society. Vivian ables (Table 3). Banding is not patterned (1990:299) further contends that symmetry in distribution, and the fact that banding and concomitant dual social divisions ex- was probably hidden behind plaster fur- hibit a patterned appearance expressed ther suggests that the builders of struc- most strongly in the southern and eastern tures exhibiting banding shared informa- parts of the San Juan Basin. tion and learning frameworks. A directed, If a great house is comprised of two Chacoan origin is suggested for the great equal, opposing parts, it is considered to houses where banding is present, but possess symmetry. Symmetry may also be what of those where it is absent? Tempo- present in the guise of two equal, oppos- ral distortion and inconsistencies in the ing roomblocks or discrete structures. The data might account for the absence of universal presence of symmetry at great banding in a few cases, but it seems likely houses would be considered support for a that some of these great houses might be directed, Chacoan origin for the struc- of local construction. tures. A patterned distribution of great houses with symmetry might support Symmetry Vivian’s (1990) interpretation or might suggest other, variable relationships be- All canyon great houses exhibit symme- tween outliers and the canyon. A com- try, defined as correspondence in size, plete lack of patterning in the distribution shape, and relative position of parts on of symmetry may be considered as sup- opposite sides of a dividing line, center, or port for local great-house construction, es- axis. In outlying great houses, symmetry pecially in those cases where symmetry is may be expressed by two facing kivas absent. within one rectangular roomblock, by two Architectural information permitted the joined roomblocks, or by two separate examination of seven Classic Bonito phase roomblocks (Vivian 1990:298). Vivian canyon great houses and 49 Classic Bonito (1970) contends that symmetry is an archi- phase outlier great houses for the presence tectural manifestation of dual social orga- of symmetry. Great houses were discarded nization that characterizes canyon great from the data base if their plans were too houses but not small-house sites. Dual so- ambiguous or too obscured by rubble to cial organization is expressed most make an assessment. There may be some strongly among the Eastern Pueblos such conflation of noncontemporaneous build- as the Tewa (Ortiz 1965, 1969). Ortiz (1965: ing episodes among the remaining sample, 389) describes Tewa dual organization as but this should not be detrimental to the “a system of antithetical institutions with analysis if symmetry during the Classic Bo- the associated symbols, ideas, and mean- nito phase was accomplished to further ings in terms of which social interaction symbolic ends. Symmetry is present at 25 takes place.” Apparent asymmetries are and absent at 24 outlier great houses (Fig. 4; balanced during an annual or biennial Table 2). No significant associations were course of rotation in what Vivian (1990: found between symmetry and any other 431–435), following Johnson (1982), refers variable (Table 3). Like banding, symmetry to as a rotating sequential hierarchy. Bi- is neither ubiquitous nor patterned in dis- lateral architectural symmetry is seen as a tribution. Diversity among these variables direct indicator of the existence of dualism could be argued to support a local origin for EVALUATING BONITO-STYLE ARCHITECTURE 493 some Bonito-style architecture, especially heights and numbers of stories, and it where symmetry is absent. provides a sufficient sample size for com- parison with other variables. A patterned Elevated Kivas distribution of elevated kivas might indi- cate information exchange among a sub- As discussed above, most Chacoan set of outlier great houses. great houses contain enclosed kivas. Some Tower kivas are documented at three great houses also contain elevated kivas, or outliers, including Kin Klizhin, Kin Ya’a, enclosed kivas are found at a second-story and Haystack (Marshall et al. 1979:15–16). level or higher. In rare cases, a series of up Second-story kivas are documented for 11 to four enclosed kivas are vertically outliers, bringing the total number of out- stacked on top of one another to create a liers with elevated kivas to 14. The pres- tower kiva (Marshall et al. 1979:18). All ence/absence of elevated kivas was found Classic Bonito phase canyon great houses, to be significantly associated with area with the exception of single-storied (Table 3). Elevated kivas are more com- Pueblo Alto, contain elevated kivas. There mon at larger sites. A total of 60 cases are a number of possible reasons for the where both elevated kiva presence/ab- construction of elevated kivas. Vivian sence and great house area are available (personal communication, October 1997) were ranked by great house area (see Ta- proposes a functional explanation, con- ble 2). Great house two-dimensional 2 tending that second-story kivas were built roofed area ranges from 72 to 4415 m with 2 to maximize the space available within the a median of 427 m . A total of 69% of the confines of the great house. Other re- elevated kivas occur in the upper half of 2 searchers have proffered the idea that the great-house sample (area Ͼ427 m ). A tower kivas functioned as part of a signal- total of 46% of the elevated kivas occur in 2 ing network (Hayes and Windes 1975). the upper quarter (area Ͼ818 m ). Signaling alone does not seem a sufficient Although no statistically significant as- explanation because, as Lekson (1984:52) sociation was found between elevated ki- states, “a similar height could have been vas and location, all except two elevated attained without the Tower Kiva.” Mar- kivas (Salmon and Morris 41) are found in shall et al. (1979:204) offer the following the south (Fig. 4). Elevated kivas also are intriguing speculation credited to Fewkes. significantly associated with road seg- ments. A Pearson’s chi-square test found It must be remembered that the ceremonial that elevated great kivas and road seg- room or kiva, in modern mythology, represents the underworld out of which...theearly races ments are not independent at a probabil- of men emerged. The tower kiva at Kin Ya’a may ity level of 0.046, although the association have been four kivas, one above another, to rep- between the two variables is weak (␾ ϭ resent the underworlds in which the ancestors 0.264). The implications of these results for of the human race live in succession before directed vs local construction at outliers emerging into that in which we now dwell. are ambiguous at best—although some Tower kivas and second-story kivas are patterning is present, there are no signif- considered here together as elevated kivas icant associations between elevated kivas to circumvent the confusion surrounding and other internal variables. the two terms that are sometimes used interchangeably (e.g., Lekson 1984:52). Kiva:Room Ratio This also avoids misinterpretations re- lated to the difficulties inherent at many Kiva:room ratio represents the number sites with reduced walls in deciding wall of kivas (excluding great kivas) against the 494 RUTH M. VAN DYKE total number of rooms within the struc- U test found the two sets of data to be ture. For example, a site with 5 kivas and significantly different (p ϭ 0.001), suggest- 25 rooms would have a kiva:room ratio of ing differences in social organization and 5:25, or 0.20. Although no direct assign- kiva function between the two types of ment of function can be made to rectan- structures (Van Dyke 1998:253–254). A gular or round rooms without excavation comparison of canyon and outlier great- data, kiva:room ratio uses some basic as- house kiva:room ratios should help deter- sumptions about the nature of these mine whether canyon and outlier great rooms to provide a way to get at possible houses are likely to have served similar differences in overall great-house func- functions without actually having to ad- tion. At Pueblo II Anasazi sites, rectangu- dress the issue of what those functions lar rooms are commonly found to have were. That is, if both canyon and outlier been used for habitation and/or storage. great houses were used for the same Round rooms, or kivas, are commonly thing, whether it be habitation, storage, or considered to have been used for ritual ritual activities; if kivas served similar purposes at least some of the time (Adler functions at canyon and outlier great 1989; Lekson 1988). Although there is a houses; and if canyon and outlier social wide range of variability, small-house or structure were similarly organized, then habitation sites frequently consist of three kiva:room ratios of both outlier and can- to six rooms associated with a single kiva. yon great houses should be similar, re- Because these small sites often are gardless of other differences in overall equated with single households, the logi- size. Variability among the set of outlier cal conclusions are (1) that each kiva was great house kiva:room ratios would sup- used by a small group of related people port local or diverse origins for outlier and (2) that any ritual activities conducted great houses. in the kiva were organized on the basis of Kiva:room ratios were calculated for 57 kinship. Steward (1937) noted a kiva:room Classic Bonito phase outlier great houses. ratio of 1:6 to 1:5 to be characteristic of This measure included second-story room Pueblo II Anasazi habitation sites and re- counts. The dependent metric variable flected that this indicated every lineage kiva:room ratio was compared against the constructed its own kiva and conducted its other variables in the analysis using non- own ceremonies. A later decrease in the parametric tests. Kiva:room ratios were number of kivas per room, with kiva:room found to be significantly different in the ratios of 1:15 to 1:25, was interpreted by north and south (Table 3). This result re- Steward as reflecting a change in the or- flects the fact that great houses located in ganization of ritual. At this point, kivas the south (n ϭ 46) have a median kiva: were considered to move beyond their room ratio of 1:9, whereas great houses role as structures for familial ritual and to located in the north (n ϭ 11) have a me- acquire a new function as settings for in- dian kiva:room ratio of 1:20. tegrative social activities (Lipe and Heg- The 1:20 kiva:room ratio figure for mon 1989). northern great houses is similar to that Powers et al. (1983:Table 41) lists kiva: obtained by Lipe (1989:56, Table 1), who room ratios for a number of great houses estimates a kiva:room ratio of 1:15.2 for and small sites. Classic Bonito-phase can- nine Chacoan great houses in the Mesa yon great houses (n ϭ 7) have a median Verde area, although contemporaneous, kiva:room ratio of 1:26.3. Canyon small local “Mesa Verde Anasazi” pueblos ex- sites (n ϭ 9), in contrast, have a median hibit kiva:room ratios of 1:6.5. This pattern kiva:room ratio of 1:6.1. A Mann–Whitney suggests Chacoan great house kivas func- EVALUATING BONITO-STYLE ARCHITECTURE 495 tioned at a large-scale, integrative level, regional patterning exists, it suggests that whereas kivas in local pueblos were more some outlier groups shared information likely used by households or extended with each other and with Chaco, but other families (Lipe 1989:59, 64). In the current groups did not. study, significant differences in kiva:room Kiva:room ratios in the north (Region 1) ratios between northern and southern are significantly lower (i.e., fewer kivas outliers may reflect differences in social per room or more rooms per kiva) than organization, kiva function, and/or rela- those in the south (Regions 2 and 3), sug- tionship with Chaco Canyon. Classic Bo- gesting differences in social organization, nito phase great houses may have had great-house function, and/or relation- more of an integrative community func- ships with Chaco Canyon between the tion in the north than in the south. As with two areas. Northern median kiva:room ra- elevated kivas, the results of the analysis tios are nearer the canyon median than with respect to the directed-vs-local ques- southern kiva:room ratios, which might tion are ambiguous. imply a closer relationship between the canyon and the northern great houses. Summary of Results However, most other evidence links the canyon with the south. Architectural at- Results of the comparative analysis in- tributes more likely in the south include dicate the Chacoan outliers found across roads and earthworks, larger great the San Juan Basin and in adjacent areas houses, elevated kivas, and higher kiva/ should not be viewed as one homogenous room ratios (i.e., more kivas per room or entity. Rather, a variety of relationships fewer rooms per kiva). This area includes appear to have existed between canyon the Chuskan slope, the central basin, the and outlier great houses. Some outlier south Chaco slope, the Red Mesa valley, great houses bear close resemblance to and points farther south and west. Such canyon great houses and probably were subregional divisions were not imple- built under canyon influence or direction, mented for the current analysis because but others may be local. Regionally pat- the subsequent sample sizes for each sub- terned distributions may indicate the ex- region were too small to elicit meaningful istence of diverse, regionally discrete re- results. However, these topographically lationships between outliers and the defined subregions bear further investiga- canyon. Findings for each of the five vari- tion as discrete or semidiscrete interactive ables are summarized in Table 4. Core- entities. and-veneer masonry is nearly ubiquitous The central basin (Region 2) does not in distribution, indicating that this con- exhibit discrete patterning with respect to struction technique was widely shared. the internal architectural variables. How- Banding and symmetry occurred in ap- ever, the central basin contains most of proximately half the cases, but they did the small outlier great houses, and ab- not co-occur in a patterned relationship, sence of great kivas is more likely. Be- and no discernible patterning was found cause the central basin is the area closest in the locational distribution of these vari- to Chaco, the people living there un- ables or in their association with other doubtedly would have had more intensive variables. A local origin could be argued interactions with Chaco than people living for outlier great houses where the vari- in more distant areas of the San Juan Ba- ables are absent. Kiva:room ratio and ele- sin and might have used canyon ritual vated kivas are not ubiquitous but do ex- facilities rather than constructed their hibit some regional patterning. Where own. This pattern resonates with Breter- 496 RUTH M. VAN DYKE

TABLE 4 Summary of Findings for Analysis of Internal Variables

Statistically significant Ubiquitous Patterned Directed or local Internal variable associations Meaningful findings distrib.? distrib.? construction?

Core-and-veneer Distance More likely to be absent Yes Yes Directed when distance from Pueblo Bonito is great Banding None Equally likely to be No No Local present or absent in any region, sometimes created by alternative means Symmetry None Equally likely to be No No Local present or absent in any region Elevated kiva Area More likely at large No Yes Ambiguous great houses Kiva:room ratio North/south Higher kiva:room ratio No Yes Ambiguous (fewer rooms per kiva) in the south nitz et al.’s (1982; Doyel et al. 1984) idea of To further explore the continuum of re- a Chaco Halo, where sites located in the lationships between outlier communities immediate environs of Chaco Canyon are and Chaco Canyon, an index of Chacoan considered economically and demograph- similarity was devised using distance ically synonymous with sites in the can- from Chaco Canyon and the five internal yon. variables. Table 5 represents an attempt to quantitatively arrange 36 outlier great THE COMPARATIVE STUDY houses along a continuum between local IN PERSPECTIVE and directed Chacoan extremes. Great houses from Table 2 were included only The results of the analysis suggest ar- when all relevant information was avail- chitectural information followed subre- able. Index scores can range between 0 gional networks of interaction across the and 5, with 0 representing an outlier great Chacoan world, but the specific nature of these networks is not defined by the cur- house differing in the extreme from can- rent study. Bonito-style architecture was yon great houses and 5 representing an not spread in a homogenous manner, and outlier great house completely at home in the presence of Bonito-style architecture Chaco Canyon. One point was assigned does not necessarily entail direct contact for presence of each of the 4 categorical with Chaco Canyon. Large-scale architec- variables. One point was assigned when tural similarities need not be interpreted kiva:room ratios fell within 1.24 standard to indicate participation in one coherent, deviations of the mean canyon great centralized Chacoan system. Some outlier house kiva:room ratio of 0.046. Finally, communities may be local, in situ devel- logged distances from Chaco (ranging be- opments. A similar interpretation was tween 1.06 and 2.28 based on logged val- reached by Marshall et al. (1979:337) and ues of distances listed in Table 2) were Powers et al. (1983:341). subtracted from the maximum subtotal of EVALUATING BONITO-STYLE ARCHITECTURE 497

5 points so that those outliers nearest the TABLE 5 ؍ canyon received the least “distance de- Outliers (n 30) Ranked from Most (High Scores) to Least (Low Scores) Interaction with Chaco Canyon merits” and vice versa. Outlier great Based on Internal Architectural Attributes and houses categorized in this manner earned Distance scores ranging from Ϫ1.28 (Bluff) to 3.30 (Kin Ya’a). Great houses with low scores Outlier Index score can be considered more local, and those Kin Ya’a 3.30 with high scores can be considered more Peach Springs 3.25 Chacoan. Of course, these rankings could Salmon 3.14 vary considerably if one or more variables Kin Klizhin 2.94 were given more weight or if other vari- Village of the Great Kivas 2.92 ables were incorporated into the index. Pierre’s 2.92 Ida Jean 2.83 Table 5 is not meant as the last word on Kin Bineola 2.78 the directed vs-local question, but it does Lowry 2.72 provide one way to illustrate the range of Pueblo Pintado 2.29 variability present among outliers. Many Kin Nizhoni 2.09 of the sites toward the Chacoan end of the Morris 39 2.04 Guadalupe 2.02 spectrum are qualitatively different from Toh La Kai 2.00 the rest of the outlier pack. Outlier socio- Morris 41 1.98 political dynamics in communities far re- Navajo Springs 1.72 moved from the canyon environs might Bee Burrow 1.51 have been quite different from sociopoliti- Escalon 1.46 Muddy Water 1.33 cal organization near to or inside the can- Whirlwind 1.31 yon. Further investigation into communi- Grey Hill Spring 1.29 ties or subregions near the local end of the Andrews 1.12 scale should incorporate other lines of ev- San Mateo 1.08 idence for or against local, in situ devel- Wallace 0.79 Greenlee 0.77 opment (Van Dyke in press). Casamero 0.14 Escalante Ϫ0.25 Implications for Chacoan Models Fort Wingate Ϫ0.98 Squaw Springs Ϫ1.04 Ϫ Results of the analysis indicate that sub- Bluff 1.28 stantial subregional diversity is contained under the rubric of what is commonly held to constitute the Chacoan “system.” one model is needed to explain the ap- The fact that Bonito-style architectural pearance of Bonito-style architecture in characteristics (great houses, great kivas, Chaco Canyon as well as throughout the and surrounding communities) are found San Juan Basin and adjacent regions. Dif- over a wide area should not be held to ferent explanatory theories may be appro- indicate that all those communities were priate for different subsets of outliers. The integrated into a holistic Chacoan system, models of Vivian (1989, 1990), Wilcox nor does it necessarily follow that they (1993), Toll (1984, 1985), and the Chaco were all in communication with one an- Center archaeologists (Judge 1979, 1989; other or with Chaco Canyon. Although Judge et al. 1981; Schelberg 1984) that some outliers may have interacted inten- posit direct canyon involvement with out- sively with Chaco Canyon, others may lier great-house construction may be ap- have interacted rarely with the canyon or propriate for outliers near the upper end not at all. Thus, it is likely that more than of the “Chacoan similarity” index. These 498 RUTH M. VAN DYKE outliers presumably derived Bonito-style loosely suggested for the outliers by the architecture by virtue of direct interaction work of Mathien (1993) and Sebastian with Chaco Canyon. But how was Bonito- (1992). It is also consonant with the ritual style architectural information transmit- landscape idea proposed by Stein and ted among outlier communities at the Lekson (1992). If, as Stein and Lekson sug- lower end of the index? No statistically gest, Bonito-style architecture was con- significant relationships were identified structed to function as a ritual setting, between internal variables and distance social inequalities could have been ex- from Chaco Canyon, so it does not appear pressed, legitimated, and created through that architectural information moved in a construction and through ceremonies or “down-the-line” manner from the canyon. activities enacted in the facilities. Control This does not support a scenario in which of ritual knowledge goes hand-in-hand neighboring communities cooperated to with the legitimation and construction of construct Bonito-style architecture, nor political, economic, and social forms of does it suggest that shared labor is re- power (Bradley 1984; Hayden 1995; White- sponsible for the spread of Bonito-style ley 1986). It is likely that the individuals or architectural forms. factions who organized construction of the Peer–polity interaction could account architecture played highly visible roles in for the appearance of Bonito-style archi- any ceremonial activities that took place tecture over a wide area without necessi- there. Because the production of both big tating shared labor or direct participation buildings and ritual is expensive, it is in events at Chaco Canyon. In a peer– likely that these leaders had an economic polity situation, Bonito-style architecture power base of sorts from which to work. In would have spread through observation a scenario similar to that proposed by Se- and competition between neighboring bastian (1992) for Chaco Canyon, leaders communities. Construction of Bonito-style might have been descendants of the first architecture would have benefited local settlers in the area who, by virtue of being leaders seeking to bolster personal pres- firstcomers, farmed the most productive tige through competition on a regional land and therefore produced and con- scale. Great houses are impressive fea- trolled access to most community surplus. tures; once they appeared in one commu- Construction of dramatic, public ritual nity, leaders of a neighboring community space could have validated these factions’ might have had little trouble convincing exclusive access to ritual knowledge and their populace that competitive emulation increased their prestige or social power. was necessary. Everyone in the commu- nity would have contributed their labor or CONCLUSIONS other resources, but the economically ad- vantaged members of the community The presence of common architectural would have been able to contribute more. elements across the San Juan Basin and This would have specifically enhanced the adjacent areas need not be interpreted to status of those with economic resources to indicate that the entire area participated spare, despite an overarching rubric of a in one coherent, centralized Chacoan sys- shared, egalitarian, communitywide en- tem. The analysis results suggest that a deavor. In this manner, individual or fac- perspective in which all Chacoan outliers tional power would be reflexively con- are integrated into one Chacoan system structed as well as expressed. may well obfuscate more than explain. Peer–polity interaction elaborates on The picture that emerges from the analy- the sorts of local aggrandizing strategies sis is of a set of outliers characterized by EVALUATING BONITO-STYLE ARCHITECTURE 499 diversity rather than homogeneity, re- terns often fail to appear as requested. quiring a range of interpretations. Like One drawback to this type of analysis is many outlier studies, this work has fo- that when patterning is ambiguous or ab- cused on Bonito-style architectural ele- sent, myriad explanations could account ments at the expense of the surrounding for the ambiguous results. In this study, I communities. Outlier great houses and as- have argued that ambiguity indicates out- sociated features are generally well de- lier relationships are more complex than scribed, but information for the surround- initially assumed and hence are only ru- ing communities often tends to be dimentarily identifiable here. Incomplete, sketchy. This reflects the daunting nature missing, or erroneous data also may have of the investigative task—several of the 58 muddied the waters, but those sorts of outliers used in the present study are problems can be ameliorated through known to be surrounded by upward of 50 continued outlier community investiga- small sites. Although the picture is slowly tions. Assumptions about the connections improving (Harper et al. 1988; Kantner between social relationships and architec- and Mahoney in press; Kendrick and ture are not altogether satisfactory but Judge 1996; Marshall 1994a, 1994b; Reed et could be refined through continued eth- al.; Van Dyke 1999), advances in our un- nographic research. derstanding of outlier dynamics are pred- If the construction of archaeological icated upon the kinds of information that knowledge merely entailed acquiring only community studies can provide. enough of the right data and identifying Massive, visually impressive architec- the proper linkages between material cul- tural features such as Chacoan great ture and human activity, our course would houses obviously contained symbolic be laborious but direct. However, despite meanings for prehistoric builders and the assumptions of this study, architec- viewers. In trying to explain why large- tural patterning does not passively reflect scale architecture appeared contempora- social organization. Relationships be- neously over a wide area, as in the Cha- tween society and the built environment coan case, one important part of the are interactive rather than reflective. So- puzzle involves determining the social cial forms are constituted, legitimated, mechanisms through which the architec- and changed by the ongoing, recursive tural information was transmitted. Carr interaction between spatial perception, (1995b) discusses the communicative po- representation, and experience (Lefebvre tential of highly visible architecture, but 1991). A categorical approach can but rep- this does not preclude the structures from resent a simplistic and incomplete at- also containing enculturative information tempt to construct information through in the form of low-visibility, internal at- pattern identification. To understand the tributes that could not have been easily uses and meanings of prehistoric build- emulated by outsiders and that reflect ings within their social contexts, creative how people shared information as they methods must be devised to examine the built the structures. Patterning among ways in which people interact with spatial low-visibility, internal attributes should forms. Despite the limitations of material help identify shared learning frameworks culture, investigations into such issues as and concomitant social affinities even labor organization, communication of within the context of high-visibility archi- building techniques, uses of spaces, intra- tecture. community social dynamics, or represen- Of course, given the tremendous com- tations of space in nonarchitectural media plexity of social and material culture, pat- are within our grasp and should assist in 500 RUTH M. VAN DYKE this regard. The ultimate goal should be to tract Archeology, University of New Mexico, interpret the meanings architectural Albuquerque. forms held for those who built, used, and Ames, K. M. perceived them in prehistory. 1995 Chiefly power and household production on the northwest coast. In Foundations of social inequality, edited by T. D. Price and G. M. ACKNOWLEDGMENTS Feinman, pp. 155–188. Plenum, New York. Baker, L. L. The research presented here was funded by a Dis- 1984 Stabilization of Guadalupe Ruin. Eastern New sertation Improvement Grant from the National Sci- Mexico University Contributions in Anthro- ence Foundation (SBR-9510037, R. Gwinn Vivian, pology 11, Portales. Principal Investigator). Outlier data was compiled Blanton, R. E., G. M. Feinman, S. A. Kowaleski, and through original fieldwork and archival research as P. N. Peregrine well as through the use of published sources. Many 1996 A dual-processual theory for the evolution individuals, including Cathy Cameron, Jim Ken- of Mesoamerican civilization. Current An- drick, Bob Powers, Paul Reed, and John Roney gra- thropology 37(1):1–14. ciously made unpublished data available to me for Bourdieu, P. this effort. Joyce Raab at the Chaco Archives and Tim 1971 The Berber house or the world reversed. In Seaman at the Laboratory of Anthropology facili- Echa´nges et Communications, edited by J. tated access to archived materials. Fieldwork in the Pouillon and P. Maranda. Mouton, The Andrews community was carried out with the per- Hague. mission of Jim Walker of the Archaeological Conser- vancy and with the assistance of Joshua Jones and Bradley, B. Sarah Herr. Helpful comments on earlier versions of 1974 Preliminary report of excavations at Wallace this manuscript were provided by Beth Grindell, Ruin, 1969–1974. Southwestern Lore 40(3–4): Sarah Herr, Louise Senior, Tom Windes, and Lisa 63–71. Young. I especially thank Chuck Adams, Barbara 1993 Wallace Ruin: Implications for outlier stud- Mills, and Gwinn Vivian for their continuing support ies. In The Chimney Rock Archaeological Sym- and intellectual contributions to the ideas contained posium, edited by J. McKim Malville and G. herein. Any errors or misinterpretations are of Matlock, pp. 72–75. U.S. Department of Ag- course the responsibility of the author. riculture Forest Service General Technical Report RM-227, Washington, DC. REFERENCES CITED Bradley, R. 1984 The social foundations of prehistoric Britain: Themes and variations in the archaeology of Adams, E. C. power. Longman, London/New York. 1983 The architectural analogue to Hopi social organization and room use, and implica- Breternitz, C. D., D. E. Doyel, and M. P. Marshall tions for prehistoric northern Southwestern (editors) culture. American Antiquity 48:44–61. 1982 Bis sa’ani: A late Bonito phase community on 1991 The origin and development of the pueblo Escavada Wash, northwest New Mexico. Na- katsina cult. University of Arizona Press, vajo Nation Papers in Anthropology No. 14, Tucson. NNCRMP, Window Rock. Adler, M. A. Cameron, C. M. 1989 Ritual facilities and social integration in 1984 A regional view of chipped stone raw mate- nonranked societies. In The architecture of rial use in Chaco Canyon. In Recent research social integration in prehistoric pueblos, edited on Chaco prehistory, edited by W. J. Judge by W. D. Lipe and M. Hegmon, pp. 35–52. and J. D. Schelberg, pp. 137–152. Reports of Occasional Papers of the Crow Canyon Ar- the Chaco Center No. 8, Division of Cultural chaeological Center 1, Crow Canyon. Research, National Park Service, Albuquer- que. Alcock, S. E. 1996 Archaeological testing at the Bluff Great House, 1993 Graecia capta: The landscapes of Roman Greece. southeastern Utah, October 1995. Manuscript Cambridge University Press. on file, Department of Anthropology, Uni- Allan, W. C., and R. Gauthier versity of Colorado, Boulder. 1976 Preservation and conservation of archaeological 1998 Coursed adobe architecture, style, and so- resources. Manuscript on file, Office of Con- cial boundaries in the American Southwest. EVALUATING BONITO-STYLE ARCHITECTURE 501

In The archaeology of social boundaries, edited State Historic Preservation Division, Santa by M. T. Stark, pp. 183–207. Smithsonian Fe. Institution Press, Washington, DC/London. Giddens, A. 1999 Hopi dwellings: Architecture at Orayvi. Univer- 1984 The constitution of society: Outline of the theory sity of Arizona Press, Tucson. of structuration. University of California, Cameron, C. M., and S. H. Lekson Berkeley. 1996 Preliminary report on the 1996 excavations on Gladwin, H. S. the Bluff Great House site. Manuscript on file, 1945 The Chaco branch: Excavations at White Mound Department of Anthropology, University of and in the Red Mesa Valley. Medallion Papers Colorado, Boulder. No. 33, Gila Pueblo, Globe. Carlson, R. L. Graves, D. K. 1966 Twin Angels pueblo. American Antiquity 31: 1990 Navajo Springs: An examination of the great 676–682. house and surrounding community. M.A. the- Carr, C. sis, Northern Arizona University, Flagstaff, 1995a Building a unified middle-range theory of Arizona. artifact design. In Style, society, and person: Hallasi, J. A. Archaeological and ethnological perspectives, 1979 Archaeological excavation at the Escalante edited by C. Carr and J. Neitzel, pp. 151–170. Site, Dolores, Colorado, 1975 and 1976. In Plenum, New York. The Archaeology and stabilization of the 1995b A unified middle-range theory of artifact Dominguez and Escalante Ruins, edited by design. In Style, society, and person: Archaeo- A. D. Reed, J. A. Hallasi, A. S. White, and logical and ethnological perspectives, edited by D. A. Breternitz. Colorado State Office, Bu- C. Carr and J. Neitzel, pp. 171–258. Plenum, reau of Land Management, Cultural Re- New York. source Series 7. Carr, C., and J. Neitzel (editors) Harper, R., M. K. Swift, B. J. Mills, J. Brandi, and J. C. 1995 Style, society, and person: Archaeological and Winter ethnological perspectives. Plenum, New York. 1988 The Casamero and Pierre’s outliers survey: An archaeological Class III inventory of the BLM Conkey, M. W., and C. A. Hastorf lands surrounding the outliers. Office of Con- 1990 Introduction. In The uses of style in archaeol- tract Archeology, University of New Mexico, ogy, edited by M. W. Conkey and C. A. Has- Albuquerque. torf, pp. 1–4. Cambridge University Press, Cambridge. Hawley, F. M. 1934 The significance of the dated prehistory of Doyel, D. E., C. D. Breternitz, and M. P. Marshall Chetro Ketl, Chaco Canyon, New Mexico. 1984 Chacoan community structures: Bis sa’ani Monographs of the School of American Re- pueblo and the Chaco halo. In Recent re- search No. 2. search on Chaco prehistory, edited by W. J. 1938 The family tree of Chaco Canyon masonry. Judge and J. D. Schelberg, pp. 37–54. Re- American Antiquity 3(3):247–255. ports of the Chaco Center No. 8, Division of Cultural Research, National Park Service, Hayden, B. Albuquerque. 1995 Pathways to power: Principles for creating socioeconomic inequalities. In Foundations of Eddy, F. W. social inequality, edited by T. D. Price and 1977 Archaeological investigations at Chimney Rock G. M. Feinman, pp. 15–86. Plenum, New Mesa, 1970–72. Memoirs of the Colorado Ar- York. chaeological Society No. 1. Hayes, A. C., and T. C. Windes Ferguson, T. J. 1975 An Anasazi shrine in Chaco Canyon. In Pa- 1996 Historic Zuni architecture and society: An ar- pers in honor of Florence Hawley Ellis, edited chaeological application of space syntax. An- by Theodore R. Frisbie, pp. 143–156. Papers thropological Papers of the University of Ar- of the Archaeological Society of New Mexico izona No. 60, University of Arizona Press, No. 2, Santa Fe. Tucson. Hayes, A. C., D. M. Brugge, and W. J. Judge Fowler, A. P., J. R. Stein, and R. Anyon 1981 Archaeological surveys of Chaco Canyon, New 1987 An archaeological reconnaissance of west-cen- Mexico. Publications in Archaeology No. 18a, tral New Mexico: The Anasazi monuments Chaco Canyon Studies, The Chaco Center, project. Manuscript on file, New Mexico National Park Service, Albuquerque. 502 RUTH M. VAN DYKE

Hegmon, M. Proceedings of the First Conference on Scientific 1992 Archaeological research on style. Annual Re- Research in the National Parks 2, edited by view of Anthropology 21:517–536. Robert M. Linn. U.S. Government Printing Hunter-Anderson, R. Office, Washington, DC. 1977 A theoretical approach to the study of house 1989 Chaco Canyon—San Juan Basin. In Dynam- form. In For theory building in archaeology, ics of Southwest prehistory, edited by L. S. edited by L. R. Binford, pp. 287–315. Aca- Cordell and G. J. Gumerman, pp. 209–261. demic Press, New York. Smithsonian Institution, Washington, DC. Irwin-Williams, C., and L. Baker (editors) Judge, W. J., W. B. Gillespie, S. H. Lekson, and H. W. 1991 Anasazi puebloan adaptation in response to cli- Toll matic stress: Prehistory of the Middle Rio Puerco 1981 Tenth century developments in Chaco Can- valley. Manuscript on file, Bureau of Land yon. In Collected papers in honor of Erik Keller- Management, Albuquerque. man Reed. Papers of the Archaeological So- Irwin-Williams, C., and P. H. Shelley (editors) ciety of New Mexico No. 6, Archaeological 1980 Investigations at the Salmon site: The structure Society Press, Albuquerque. of Chacoan society in the northern Southwest. Kantner, J., and N. Mahoney (editors) Eastern New Mexico University Printing in Great house communities across the San Juan Services, Portales. press Basin. Anthropological Papers of the Univer- Irwin-Williams, C., R. Adams, P. C. Johnson, M. A. sity of Arizona, University of Arizona Press, Bennett, E. P. Morenon, and G. A. Davis Tucson. 1975 Salmon Ruin archaeological investigations. Ar- Kendrick, J. W., and W. J. Judge chaeological Completion Report Series No. 1996 The Lowry Community: Testing great house 8, Division of Grants, Office of Archaeolog- models on the Chaco frontier. Paper presented ical and Historic Preservation, National Park at the 61st Annual Meeting of the Society for Service, Department of the Interior, Wash- American Archaeology, April 13, 1996, New ington, DC. Orleans. Jacobson, L. A. Kent, S. 1984 Chipped stone in the San Juan Basin: A distri- 1990 A cross-cultural study of segmentation, ar- butional analysis. Unpublished M.A. thesis, chitecture, and the use of space. In Domestic Department of Anthropology, University of architecture and the use of space, edited by S. New Mexico, Albuquerque. Kent, pp. 127–152. Cambridge University Jeancon, J. A. Press, Cambridge. 1922 Archaeological research in the northeastern San Juan Basin during the summer of 1921. The Kincaid, C. (editor) State Historical and Natural History Society 1983 Chaco roads project, Phase I: A reappraisal of of Colorado and the University of Denver. prehistoric roads in the San Juan Basin. Bureau of Land Management, Albuquerque. Jeancon, J. A., and F. H. H. Roberts, Jr. 1923 Further archaeological research in the Kintigh, K. W. northeastern San Juan Basin of Colorado. 1994 Chaco, communal architecture, and Cibolan The Colorado Magazine 1(1):3–10. aggregation. In The ancient Southwestern com- Johnson, G. A. munity: Models and methods for the study of 1982 Organizational structure and scalar stress. prehistoric social organization, edited by W. H. In Theory and explanation in archaeology: The Wills and R. D. Leonard, pp. 131–140. Uni- Southhampton Conference, edited by C. Ren- versity of New Mexico Press, Albuquerque. frew, M. Rowlands, and B. Segraves, pp. Kluckhohn, C. 389–421. Academic Press, New York. 1939 Discussion. In Preliminary report on the 1937 Judd, N. M. Excavations: Bc50-51, Chaco Canyon, New 1927 The architectural evolution of Pueblo Bo- Mexico, edited by C. Kluckhohn and P. nito. Proceedings of the National Academy of Reiter, pp. 151–162. University of New Mex- Sciences 13(7):561–563. ico Bulletin 345, Anthropological Series 3(2), 1964 The architecture of Pueblo Bonito. Smithso- Albuquerque. nian Miscellaneous Collections 147(1). Lechtman, H. Judge, W. J. 1977 Style in technology—some early thoughts. 1979 The development of a complex cultural eco- In Material culture, styles, organization, and system in the Chaco Basin, New Mexico. In dynamics of technology, edited by H. Lecht- EVALUATING BONITO-STYLE ARCHITECTURE 503

man and R. Merrill, pp. 3–20. West, New 1994b Standing Rock community study. In A study York. of two Anasazi communities in the San Juan Lefebvre, H. Basin, by R. J. Bradley and R. Sullivan. Uni- 1991 The production of space, translated by D. Ni- versity of New Mexico Office of Contract cholson-Smith. Basil Blackwell, Cambridge. Archeology and Maxwell Museum of An- thropology, Albuquerque. Lekson, S. H. Marshall, M. P., J. R. Stein, R. W. Loose, and J. E. 1984 Great pueblo architecture of Chaco Canyon, Novotny New Mexico. University of New Mexico 1979 Anasazi communities of the San Juan Basin. Press, Albuquerque. Public Service Company of New Mexico, Al- 1988 The idea of the kiva in Anasazi archaeology. buquerque. The Kiva 53(3):213–234. Marshall, M. P., D. E. Doyel, and C. D. Breternitz 1991 Settlement patterns and the Chaco region. 1982 A regional perspective on the late Bonito In Chaco and Hohokam: Prehistoric regional phase. In Bis sa’ani: A late Bonito phase com- systems in the American Southwest, edited by munity on Escavada Wash, northwest New P. L. Crown and W. J. Judge, pp. 31–55. Mexico, edited by C. D. Breternitz, D. E. School of American Research, Santa Fe. Doyel, and M. P. Marshall, Volume 3, pp. Lemonnier, P. 1227–1240. Navajo Nation Papers in Anthro- 1986 The study of material culture today: Toward pology No. 14, NNCRMP, Window Rock. an anthropology of technical systems. Jour- Marshall, M. P., and A. Sofaer nal of Anthropological Archaeology 5:147–186. 1988 Solstice project: Archaeological investigations in Lipe, W. D. the Chacoan province, New Mexico. Manu- 1989 Social scale of Mesa Verde Anasazi kivas. In script on file, Laboratory of Anthropology, The architecture of social integration in prehis- Museum of New Mexico, Santa Fe. toric pueblos, edited by W. D. Lipe and M. Martin, P. S. Hegmon, pp. 53–71. Occasional Papers of 1936 Lowry Ruin in southwestern Colorado. In the Crow Canyon Archaeological Center Field Museum of Natural History, anthropolog- No. 1, Cortez, Colorado. ical series, Volume 23(1). Chicago. Lipe, W. D., and M. Hegmon (editors) Mathien, F. Joan 1989 The architecture of social integration in prehis- 1993 Exchange systems and social stratification toric pueblos. Occasional Papers of the Crow among the Chaco Anasazi. In The American Canyon Archaeological Center No. 1, Southwest and Mesoamerica: Systems of prehis- Cortez, Colorado. toric exchange, edited by J. E. Ericson and T. G. Baugh, pp. 27–63. Plenum, New York. McGuire, R. H., and M. B. Schiffer 1983 A theory of architectural design. Journal of Mindeleff, V. Anthropological Archaeology 2:277–303. 1989 A study of pueblo architecture in Tusayan and [1891] Cibola. Smithsonian Institution Press, Wash- McKenna, P. J., and H. W. Toll ington, DC. 1991 Regional patterns of great house develop- Morris, E. H. ment among the Totah Anasazi, New Mex- 1928 Notes on excavations in the Aztec ruin. In ico. In Anasazi regional organization and the Anthropological papers of the American Mu- Chaco system, edited by D. E. Doyel. Maxwell seum of Natural History, Volume 26(5). Amer- Museum of Anthropology Anthropological ican Museum of Natural History, New York. Papers No. 5. 1939 Archaeological studies in the La Plata district: McKenna, P. J., and M. Truell Southwestern Colorado and northwestern New 1986 Small site architecture of Chaco Canyon, Mexico. Carnegie Institution of Washington New Mexico. In Publications in archaeology, Publication No. 519, Washington, DC. Volume 18D. Division of Cultural Resources, Nials, F., J. Stein, and J. Roney National Park Service, Santa Fe. 1987 Chacoan roads in the southern periphery: Re- Marshall, M. P. sults of phase II of the BLM Chaco roads project. 1994a El Llano-Escalon community study. In A Bureau of Land Management, Albuquer- study of two Anasazi communities in the San que. Juan Basin, edited by R. J. Bradley and R. Ortiz, A. Sullivan. University of New Mexico Office of 1965 Dual organization as an operational concept Contract Archeology and Maxwell Museum in the pueblo Southwest. Ethnology 4(4):389– of Anthropology, Albuquerque. 396. 504 RUTH M. VAN DYKE

1969 The Tewa world: Space, time, being, and becom- als. Bureau of American Ethnology Bulletin ing in a pueblo society. University of Chicago No. 126. Press, Chicago. Roney, J. R. Peckham, S. L. 1992 Prehistoric roads and regional integration in 1958 Salvage archaeology in New Mexico 1957– the Chacoan system. In Anasazi regional or- 1958: A partial report. El Palacio 65(5):161– ganization and the Chaco system, edited by 163. D. E. Doyel, pp. 123–131. Maxwell Museum 1969 An archaeological site inventory of New Mexico, of Anthropology Anthropological Papers 5, Part I. Manuscript on file, Laboratory of An- University of New Mexico, Albuquerque. thropology, Museum of New Mexico, Santa Runciman, W. G. Fe. 1982 Origins of states: The case of archaic Greece. Comparative studies in society and history 24: Pippin, L. C. 351–377. 1987 Prehistory and paleoecology of Guadalupe Ruin, New Mexico. University of Utah Anthropo- Sackett, J. D. logical Papers No. 107, Salt Lake City. 1982 Approaches to style in lithic archaeology: Journal of Anthropological Archaeology 1:59– Powers, R. P. 112. 1984 Regional interaction in the San Juan Basin: 1985 Style and ethnicity in the Kalahari: A reply The Chacoan outlier system. In Recent re- to Wiessner. American Antiquity 50:154–159. search on Chaco prehistory, edited by W. J. 1990 Style and ethnicity in archaeology: The case Judge and J. D. Schelberg, pp. 23–36. Re- for isochrestism. In The uses of style in archae- ports of the Chaco Center No. 8, Division of ology, edited by M. W. Conkey and C. A. Cultural Research, National Park Service, Hastorf, pp. 32–43. Cambridge University Albuquerque. Press, Cambridge, England. Powers, R. P., W. B. Gillespie, and S. H. Lekson Schelberg, J. D. 1983 The outlier survey: A regional view of set- 1984 Analogy, complexity, and regionally-based tlement in the San Juan Basin. In Reports of perspectives. In Recent research on Chaco pre- the Chaco Center 3. U.S. Department of the history, edited by W. J. Judge and J. D. Schel- Interior, National Park Service, Albuquer- berg, pp. 5–21. Reports of the Chaco Center que. No. 8, Division of Cultural Research, Na- tional Park Service, Albuquerque. Reed, P. F., and K. N. Hensler (editors) 1998 Anasazi community development in Redrock Schlanger, S. H. Valley: Final report on the Cove Archaeological 1986 Population studies. In Dolores archaeological Project along the N33 road in Apache County, program: Final synthetic report, compiled by Arizona. Navajo Nation Papers in Archaeol- D. A. Breternitz, C. K. Robinson, and G. T. ogy No. 33, Navajo Nation Archaeology De- Gross, pp. 493–501. United States Depart- partment, Window Rock. ment of the Interior, Bureau of Reclamation, Denver. Reiter, P. Sebastian, L. 1933 The ancient pueblo of Chetro Ketl. Unpub- 1992 The Chaco Anasazi: Sociopolitical evolution in lished M.A. thesis, University of New Mex- the prehistoric Southwest. Cambridge Univer- ico, Albuquerque. sity Press, Cambridge. Renfrew, C., and J. F. Cherry Sigleo, A. C. 1986 Peer polity interaction and socio-political 1981 Casamero: A Chacoan site in the Red Mesa Val- change. Cambridge University Press, Cam- ley, New Mexico. Unpublished manuscript bridge. on file, Chaco Center, National Park Service, Roberts, F. H. H., Jr. Albuquerque. 1932 Village of the Great Kivas on the Zuni Reserva- Smith, A. T. tion, New Mexico. Bureau of American Eth- 1996 Imperial archipelago: The making of the Urar- nology Bulletin 111. tian landscape in southern Transcaucasia. Ph.D. 1939 Archaeological remains in the Whitewater Dis- dissertation, Department of Anthropology, trict, eastern Arizona. Part 1: House types. Bu- University of Arizona, Tucson. UMI Publi- reau of American Ethnology Bulletin No. cations, Ann Arbor. 121. Smith, C. A. 1940 Archaeological remains in the Whitewater Dis- 1976a Regional economic systems: Linking geo- trict, eastern Arizona. Part 2: Artifacts and buri- graphical models and socioeconomic prob- EVALUATING BONITO-STYLE ARCHITECTURE 505

lems. In Regional analysis I: Economic systems, Vivian, R. G. edited by C. A. Smith, pp. 3–63. Academic 1970 An inquiry into prehistoric social organiza- Press, New York. tion in Chaco Canyon, New Mexico. In Re- 1976b Exchange systems and the spatial distribu- constructing prehistoric pueblo societies, edited tion of elites: The organization of stratifica- by W. A. Longacre, pp. 59–83. School of tion in agrarian societies. In Regional analysis American Research, Advanced Seminar Se- II: Social systems, edited by C. A. Smith, pp. ries, University of New Mexico Press, Albu- 309–374. Academic Press, New York. querque. Stein, J. R., and S. H. Lekson 1983 Appendix A, discovery and description: 1992 Anasazi ritual landscapes. In Anasazi re- Chacoan road field studies, 1963–1980. In gional organization and the Chaco system, ed- Chaco Roads Project, Phase I: A reappraisal of ited by D. E. Doyel, pp. 87–100. Maxwell prehistoric roads in the San Juan Basin, 1983, Museum of Anthropology Anthropological edited by C. Kincaid, pp. A1–A34. Bureau of Papers 5, University of New Mexico, Albu- Land Management, Albuquerque. querque. 1989 Kluckhohn reappraised: The Chacoan sys- tem as egalitarian enterprise. Journal of An- Steward, J. 45: 1937 Ecological aspects of Southwestern society. thropological Research 101–113. Anthropos 32:87–104. 1990 The Chacoan prehistory of the San Juan Basin. Academic Press, New York. Toll, H. W. 1997 Chacoan roads: Morphology. Kiva 63(1):7– 1984 Trends in ceramic import and distribution 34. in Chaco Canyon. In Recent research on Chaco prehistory, edited by W. J. Judge and J. D. Warburton, M., and D. K. Graves Schelberg, pp. 115–135. Reports of the 1992 Navajo Springs, Arizona: Frontier outlier or Chaco Center No. 8, National Park Service, autonomous great house? Journal of Field Ar- Albuquerque. chaeology 19(1):51–69. 1985 Pottery, production, and the Chacoan Anasazi Washburn, D. system. Ph.D. dissertation, University of Col- 1974 Nearest neighbor analysis of Pueblo I-II set- orado. University Microfilms, Ann Arbor. tlement patterns along the Rio Puerco of the Tuzin, D. F. East, New Mexico. American Antiquity 39(2): 1976 The Ilahita Arapesh: Dimensions of unity. Uni- 315–335. versity of California Press, Berkeley. Whiteley, P. M. Van Dyke, R. M. 1986 Unpacking Hopi “clans,” II: Further ques- 1998 The Chaco connection: Bonito-style architecture tions about Hopi descent groups. Journal of in outlier communities. Ph.D. dissertation, De- Anthropological Research 42(1):69–79. partment of Anthropology, University of Ar- izona. UMI Publications, Ann Arbor. Wiessner, P. 1999 The Andrews community: A Chacoan out- 1983 Style and social information in Kalahari San lier in the Red Mesa Valley, New Mexico. projectile points. American Antiquity 48:253– Journal of Field Archaeology 26. 276. in A Chacoan ritual landscape: The view from 1984 Reconsidering the behavioral basis for style: press the Red Mesa Valley. In Great house commu- A case study among the Kalahari San. Jour- nities across the San Juan Basin, edited by J. nal of Anthropological Archaeology 3:190–234. Kantner and N. Mahoney. Anthropological 1985 Style or isochrestic variation? A reply to Papers of the University of Arizona, Univer- Sackett. American Antiquity 50:160–166. sity of Arizona Press, Tucson. 1990 Is there a unity to style? In The uses of style in Vivian, G., and T. W. Mathews archaeology, edited by M. W. Conkey and 1965 Kin Kletso: A Pueblo III community in Chaco C. A. Hastorf, pp. 105–112. Cambridge Uni- Canyon, New Mexico. Southwest Monuments versity Press, Cambridge. Association, Technical Series No. 6, Part 1, Wilcox, D. Globe, Arizona. 1993 The evolution of the Chacoan polity. In The Vivian, G., and P. Reiter Chimney Rock Archaeological Symposium, ed- 1960 The great kivas of Chaco Canyon and their re- ited by J. McKim Malville and G. Matlock, lationships. Monographs of the School of pp. 76–90. United States Department of Ag- American Research and the Museum of riculture Forest Service General Technical New Mexico No. 22, Santa Fe. Report RM-227. 506 RUTH M. VAN DYKE

Wills, W. H., and T. C. Windes ies, Publications in Archeology No. 18F, Na- 1989 Evidence for population aggregation and tional Park Service, Santa Fe. dispersal during the Basketmaker III period Windes, T. C., and D. Ford in Chaco Canyon, New Mexico. American 1992 The nature of the early Bonito phase. In Antiquity 54(2):347–369. Anasazi regional organization and the Chaco Windes, T. C. system, pp. 75–86, edited by D. E. Doyel. 1987 Investigations at the Pueblo Alto complex, Chaco Maxwell Museum of Anthropology Anthro- Canyon, New Mexico, 1975–1979: Volume I, pological Papers 5, University of New Mex- volume II, parts 1 and 2. Chaco Canyon Stud- ico, Albuquerque.