Reevaluating the Organization of Lapidary Production at Chaco Canyon

Abstract

Several decades ago, the National Park Service's Chaco Project revealed evidence for widespread ornament manufacture at small sites (small houses) in Chaco Canyon, New Mexico, as well as possible workshop-scale production at two of these locations. Given that consumption of finished jewelry items is clearly concentrated at large sites (great houses), it was suggested that lapidary production was part of a larger corporate political strategy, in which goods produced in surrounding small houses were used to sustain communal events related to construction activities and ritual performances at great houses. This article addresses a critical gap in this narrative—ornament production at great houses. Using Pueblo Bonito as a case study, I present the results of a systematic analysis of lapidary tools from the site and characterize the nature of on-site ornament manufacture. I find evidence that significant jewelry-making was occurring at Pueblo Bonito, at least on par with previously documented small-house jewelry workshops, and that a portion of this was embedded within elite households. These results require us to reconsider the role of ornament production in Chacoan political economy.

Resumen

Hace varias décadas, el Proyecto de Chaco del Servicio Nacional de Parques halló evidencia de la fabricación de adornos personales en “casas pequeñas” en el Cañon de Chaco (Nuevo Méjico), incluyendo la producción a escala de taller especializado en dos de estos yacimientos. Puesto que el consumo de objetos de adorno está concentado en “las casas grandes”, se ha sugerido que la producción lapidaria haya sido parte de una estrategía política corporativa más amplia en la cual los bienes producidos en las casas pequeñas cercanas fueron empleados para sostener eventos comunales relacionados con las actividades de construcción y actuaciones rituales en las casas grandes. Este artículo pretende rellenar un hueco en esta narrativa: la producción de adornos en casas grandes. Con el Pueblo Bonito como caso de estudio, presento los resultados de un análisis sistemática de los utensilios lapidarios del sitio y caracterizo la naturaleza de la fabricación de adornos en este yacimiento. Encuentro evidencia significativa de fabricación en el Pueblo Bonito, al menos tanto como se había documentado anteriormente en talleres de las casas pequeñas, y que una porción de esta producción tomó lugar dentro de residencias de familias de élite. Estos resultados nos obligan a reconsiderar el papel de la producción de adornos en la economía política de la sociedad de Chaco.

Our understanding of craft production in Chaco Canyon, New Mexico—and its role in the ritual and political economy of ninth- through twelfth-century Chacoan society—is largely based on work conducted on small sites during the National Park Service's (NPS) Chaco Project. Imbalances in the production and consumption of turquoise ornaments at small and large sites (“small houses” and “great houses”) identified during that effort directly inform key arguments proposed for Chacoan political strategy, system financing, and the organization of production (e.g., Cameron and Toll 2001; Toll 2006). Based in part on these patterns, researchers propose that Chaco was a corporate chiefdom in which production was diffuse and principally organized at the small-site household level, whereas consumption was highly centralized, oriented toward communal ritual activities at great houses, and coordinated by elites or ritual leaders (Earle 2001; Hagstrum 2001; Mathien 2001; Peregrine 2001; Renfrew 2001; for more recent discussions of ritual production, see Gruner 2019; Hanson 2023; Mills 2023; Plog et al. 2017).

There is a critical gap in these accounts, however: an evaluation of ornament production activities at great houses. Although much has been written about the quantities of turquoise and shell jewelry in elite burials and ritual deposits at Pueblo Bonito (e.g., Ditto 2017; Giomi and Peeples 2019; Gruner 2019; Heitman 2015; Mattson 2016; Mills 2008; Plog and Heitman 2010), lapidary tools have not been systematically identified in the site's artifact assemblages. How can we fully understand the configuration of jewelry manufacture and use at Chaco without fully considering the evidence for lapidary production at great houses in a manner similar to that at small houses? In an effort to begin to address this problem, I present the results of an analysis of lapidary tools from Pueblo Bonito—the largest Chacoan great house—and compare them to data collected for two probable small-house jewelry workshop sites (29SJ629 and 29SJ1360) by the Chaco Project. I then interpret these results in the context of the larger organization of ornament manufacture in the canyon using broad parameters of production (Costin 1991) and discuss the implications for previously proposed models of Chacoan corporate political strategy and the role of ornaments in those models.

The Organization of Production in Chaco Canyon

Chaco Canyon, located in the San Juan Basin of northwestern New Mexico, is widely understood as the center of a distinctive Ancestral Pueblo cultural system associated with monumental architecture, elaborate and highly structured ritual practice, a unique suite of material culture, the exchange of imported goods, and the most hierarchical social organization in the region before or since. First coalescing by the early AD 800s, Chacoan developments profoundly influenced the social trajectory of the northern Southwest, even well after the decline of the canyon as a center place in the region in the late twelfth century (Plog and Heitman 2010; Plog et al. 2022; Watson et al. 2015). The economic and sociopolitical structure represented at Chaco Canyon remains a topic of debate, though there seems to be a growing scholarly consensus in favor of greater social complexity, a more defined social hierarchy, and a higher local residential population (Lekson 2018; Mills 2023; Plog 2018).

The classic archaeological signature of Chacoan communities is a two-tiered settlement hierarchy consisting of great houses—large, multistoried, and architecturally distinctive pueblos—and surrounding smaller and usually single-storied small houses (Kantner and Mahoney 2000; McKenna and Truell 1986; Powers et al. 1983). Great houses are variously interpreted as large residential pueblos, elite complexes, or public ritual centers (Crown 2016; Ditto 2017; Fowler and Stein 1992; Heitman 2007, 2011, 2015; Lekson 1999, 2006; Mills 2015; Plog 2018; Renfrew 2001; Wills 2001; Wills et al. 2021). These prominent structures are located across the Colorado Plateau but are clustered within Chaco Canyon itself (Figure 1). Of the 12 great houses in the canyon, Pueblo Bonito is the largest, with approximately 650 rooms and 35 kivas (circular, semisubterranean ritual structures). Although its function is debated, its size, architectural complexity, large number of dedicated ceremonial spaces, high-status interments, and concentration of imported and socially valuable material goods make it clear that Pueblo Bonito was an important focus of Chaco identity and practice (Crown 2016; Lekson 1999, 2006, 2018; Neitzel 2003). Researchers generally agree that small houses served as residences for nonelite, extended families engaged in farming and craft production.

Figure 1. Satellite image of Chaco Canyon with sites mentioned in the text. Drafted by Hannah V. Mattson.

The relationship between great houses and small houses is at the heart of models proposed for the economic maintenance of Chaco society in the core area. Several researchers view Chaco as a chiefdom with a corporate political strategy (Blanton et al. 1996; Feinman 2000) used to mobilize resources from the population. Most scholars agree that the primary purpose of this corporate strategy, and one that drove the organization of production in the canyon, was the management and funding of communal ritual events. Although the production of goods of all kinds is argued to be diffuse across canyon communities, particularly at small house sites (Hagstrum 2001), their consumption appears to be consolidated at great houses (Ditto 2017; Peregrine 2001:37; Plog et al. 2017:293–294; Watson et al. 2015). Earle (2001:26–27) argues for a system of staple financing that relied on contributions of surplus agricultural produce, utilitarian goods, and labor for large-scale constructions. Peregrine (2001:41) and Toll (2006:147) propose that leaders had an active role in promoting the production of votive goods, as well as other materials and labor. This view is also supported by the recent work on the distribution of imported materials and ritual paraphernalia at Pueblo Bonito (Bishop 2019; Bishop and Fladd 2018; Bishop et al. 2023; Gruner 2019; Hanson 2023; Mills 2023; Watson et al. 2015; Weiner 2018). Taking a somewhat different view, Renfrew (2001) views high-value goods consumed at great houses—referred to as places of “high devotional expression”—as offerings supplied by (and therefore presumably produced by) religious visitors or pilgrims within a sacred economy.

Ornament Production and Consumption

Chaco's great houses are remarkable in the quantity of jewelry and imported raw materials they contain compared to both contemporary small houses and other archaeological sites in the Ancestral Pueblo region. Early excavations at Pueblo Bonito recovered over 100,000 finished ornaments of turquoise, marine shell, jet, and shale; although these were found in virtually all contexts at the site, they were particularly concentrated within sumptuous burials and dedicatory offerings in ritual spaces (Judd 1954; Pepper 1920).

Whereas previous research on the meaning and value of turquoise and ornaments at Chaco is chiefly based on great-house assemblages, our knowledge of lapidary production relies on work conducted during the NPS Chaco Project. This archaeological research program, with fieldwork carried out in the 1970s, included excavation of 12 small houses with Bonito-phase (AD 850–1150) components, in addition to portions of the Pueblo Alto great house (e.g., Lekson 2006; Mathien, ed. 1997; Mathien and Windes 1987; McKenna 1984; Windes 1993). Although ornaments and lapidary debris were recovered from nearly all the sites investigated, over 80% of the more than 5,000 ornaments collected are from just two small houses—29SJ629 and 29SJ1360. These two sites also yielded greater quantities of lapidary tools, production debris, and unfinished ornaments (Mathien 2001; Windes 1992).

29SJ629, located on the south side of the canyon in an area known as Marcia's Rincon, consists of nine masonry rooms and three pit structures (Windes 1993). Two floors of one of the pit houses and features associated with a plaza were found to contain thousands of pieces of turquoise debris, along with lithic microdrills, ground stone lapidary abraders and files, and ornaments in various stages of manufacture (Mathien 2001; Windes 1993). One floor alone included more than 3,000 pieces of turquoise, and one of the pits in the plaza enclosed more than 30 lapidary abrading tools and microdrills (Windes 1993:307–308). The contexts associated with these artifacts date to AD 925–1050, the early Pueblo II period / Early Bonito phase.

The site of 29SJ1360, located just north of Fajada Butte, dates to the same period (AD 920–1020) and consists of two room blocks (including approximately 16 rooms total) and three pit structures. Lapidary evidence was recovered from one pit structure, which may be a possible kiva, and the floor of one of the plaza areas (Mathien 1984, 2001; McKenna 1984; Windes 1993:312). Though not as abundant as at 29SJ629, this material includes unfinished ornaments, turquoise debris, and a probable jeweler's tool kit (Akins 1997; McKenna 1984:275).

Although only these two sites were identified as intensive jewelry-manufacturing areas, indications of low-level turquoise ornament production were also identified at other small houses, including the Early Bonito sites of 29SJ626 and 29SJ627, the Classic Bonito–phase (AD 1020–1120) component of 29SJ423, and the Late Bonito (AD 1120–1220) small houses of Bc 51 and Bc 59. Based on his examination of turquoise on the surfaces of unexcavated sites, particularly on anthills, Windes (Windes 1992:161) notes that 94% of the Early Bonito–phase sites in the canyon contain turquoise debris, and over 40% have beads that were broken during the drilling process.

The NPS findings figure prominently in the major models proposed for the organization of production in the canyon. Turquoise ornament manufacture is generally interpreted as the product of dispersed, part-time, and independent household specialization across the small–house population, carried out at the encouragement—but not under the direct control—of elites (Earle 2001; Hagstrum 2001; Peregrine 2001; Plog et al. 2017:293–294; Toll 2006). In her study of craft production at small houses, Hagstrum (2001:52) sees such contributions as an obligatory “underwriting [of] the chiefdom” around which households had to schedule their other activities.

However, there is also evidence of jewelry production at canyon great houses throughout the entire temporal sequence that these studies do not consider. Using published excavation reports and museum catalogs at the American Museum of Natural History (AMNH), Mathien (1984:177–186, Table 3) notes the presence of possible lapidary workshops at Pueblo Bonito, Pueblo del Arroyo, Una Vida, Kin Kletso, and Kin Nahasbas, and she concludes that jewelry production took place on a part-time basis at both small and great houses (see also Gruner 2019:279; Mills 2023:234; Plog et al. 2017:293; Windes 1992).

Methodology

Sample

Between 2019 and 2022, I identified and analyzed potential lapidary tools from excavated contexts within Pueblo Bonito housed at the AMNH and the National Museum of Natural History (NMNH). These objects were collected by the Hyde Exploring Expedition (1896–1899) and the National Geographic Society Expedition (1920–1924), projects that completed excavation of almost the entirety of the structure (Judd 1954; Pepper 1920). I also included data from, but did not reanalyze, additional tools recovered from the two middens in front of Pueblo Bonito by University of New Mexico's (UNM) Chaco Stratigraphy Project (Hegberg and Crown 2016; Wills and Okun 2016). These were collected from three trenches placed through the mounds by Neil Judd in the 1920s. Material from the trenches was analyzed during the Pueblo Bonito Mounds project (Crown 2016).

Probable lapidary tools were initially identified based on similarity to published descriptions and images of ground stone and flaked stone tools found in direct association with production debris in workshop contexts documented at the sites of 29SJ629 and 29SJ1360 by the Chaco Project (Akins 1997; Lekson 1993, 1997; Mathien 2001; McKenna 1984; Windes 1993).

Lapidary Tool Classification

Ancestral Pueblo lapidary tools largely consist of four main classes of artifacts: passive abraders, active abraders, drills, and saws. Passive abraders include ground stone lapidary abraders, lapstones, and grooved abraders/bead-shapers; active abraders include ground stone files and reamers; and lapidary drills include both formal, bifacial tools and expedient flake tools. Saws are either ground stone and tabular or flaked stone and serrated.

Lapidary abraders and lapstones are tabular netherstones against which stone or shell ornaments are ground to shape. They have flat grinding surfaces and are both smaller and thinner than metates or grinding slabs. Lapidary abraders identified at the two small-house jewelry workshop sites are produced from unshaped local sandstone and are generally of sizes conducive to holding in the hand (McKenna 1984:Table 4.21; Windes 1993:Table 4.4; see also Hedquist 2017). Lapstones are larger, thinner, and typically more formal than lapidary abraders, often displaying edge modification in the form of grinding and/or flaking. Based on archaeological and experimental evidence, use wear on lapidary abraders and lapstones consists of clusters of deep scratches and striations that do not have a consistent orientation and do not extend across the entire work surface. It should be noted that lapstones, given their larger sizes, may have served as more generalized work surfaces compared to lapidary abraders, and they were used to passively grind a variety of materials to shape, including wood, bone, and horn/antler.

Grooved passive abraders are often assumed to represent shaft straighteners or awl shaping/sharpening stones, but some of these tools may have been used for shaping strings of beads (Jernigan 1978:204, Figure 95). I included grooved abraders in the Pueblo Bonito lapidary tool sample when they exhibited U-shaped grooves of widths consistent with disc bead diameters from the site.

Files are small, handheld abrading tools that are typically rectangular or cylindrical in shape (Windes 1993:227–228). They display grinding on both surfaces and edges, and beveling or faceting is common; striations on these surfaces are typically diagonal (versus parallel) to the short or long axes. Reamers are a special kind of file used for widening and shaping large holes, such as for rings and bracelets (Jernigan 1978:201). Given that reamers were not separated from files in the Chaco Project analysis, I also combined these categories for comparison.

Lapidary drills are those with blade dimensions consistent with the sizes of ornament perforations. Most of the drills recovered from lapidary contexts at 29SJ629 are expedient microdrills, produced by retouching one or two margins of a flake to exaggerate a projection. The majority are fashioned from silicified wood, a material sufficiently hard to drill turquoise. It is unknown if these drills were hafted, although one intact example from Pueblo Bonito suggests that they were. Although a few bow-drill components dating to the eleventh to thirteenth centuries have been identified in southeast Utah and northeast Arizona (Curcija 2018), it does not appear that bow drills were used widely across the Southwest before the introduction of metal. Given that there is currently no evidence of bow drills in the Chaco area, it is assumed that microdrills were hafted and used as palm drills—whose shaft was spun between the hands—similar to those used by the Chumash (e.g., Blake 2010). This is also supported by ethnographic research at Cochiti Pueblo (Lange 1959:168–169, quoting Frederick Starr [Notebook, New Mexico, 1897, Manuscript diary, Box 30, Folder 1, University of Chicago Library, Chicago]), where the drilling of beads with a “simple drill shaft” is noted as preceding the introduction of pump drills around AD 1880.

Although neither tabular nor serrated saws were identified as such in the 29SJ629 or 29SJ1360 assemblages, both were likely components of ancient Southwestern lapidary tool kits (Jernigan 1978:201, Figure 95). These tools would have been used for cutting or scoring raw material from matrix or into smaller pieces. Tabular saws are palm-sized pieces of very thin sandstone with at least one tapered edge; longitudinal striations are usually visible along this cutting edge. Serrated saws are generally smaller than tabular saws, produced from various cryptocrystalline materials, and they have flaked teeth along one margin.

Data Collection

After preliminary identification of potential lapidary tools in the AMNH and NMNH collections, additional attributes related to material type, design, manufacture, size, work surface topography, and manner and intensity of use and reuse were collected. Particularly relevant to this study is use wear, which was analyzed both under low and high magnification using a digital microscope. The presence, distribution, and orientation of scratches and striations were especially important in verifying use consistent with lapidary activities for abraders, whereas the location of microflakes and step fractures, rounding, and polish were used to confirm drilling use on potential expedient microdrills.

Results

The lapidary tool assemblage from Pueblo Bonito totals 758 objects, including 335 from the structure and 423 from the mounds (Table 1). Lapidary abraders and lapstones are the most common tool types, followed by drills and files (Figure 2). The assemblage contains almost eight times the number of lapidary tools as 29SJ629 and more than three times the combined assemblages from all the Chaco Project sites. This total probably still significantly underrepresents the full lapidary tool assemblage from Pueblo Bonito, given that the material from the two mounds is derived only from the three trench samples, which represent a small percentage of the total mound volume. It is also likely that a portion of the tools from the structure were not recognized as significant and were therefore discarded by early excavators.

Table 1. Summary of Pueblo Bonito Assemblage by Lapidary Tool Type.

^a This includes objects from the mounds curated at AMNH and NMNH and those excavated from Judd's trenches by UNM.

^b Active abraders with complete dimensions consistent with files recorded from structure.

^c This does not include tools with grooves that have alternate primary uses.

^d Three formal bead shapers and five grooved abraders with groove widths consistent with bead diameters.

^e Classified as “lapstone/netherstone” in analysis; those with complete dimensions are consistent with lapidary lapstones.

Figure 2. Frequencies of major lapidary tool types at Pueblo Bonito, 29SJ629, and all National Park Service (NPS) Chaco Project sites combined. Note that microdrills were not defined as a distinct category in the Chaco Project synthesis (Lekson 1997).

Lapidary Abraders and Lapstones

A total of 179 lapidary abraders and 325 lapstones were recovered from Pueblo Bonito (Figures 3 and 4). Immediately apparent is the large number of lapstone fragments from the mounds (n = 242), classified by Hegberg and Crown (2016). Due to possible differences in the way lapstones were defined between different projects, only those with dimensions consistent with lapstones from the structure, 29SJ629, and 29SJ1360 were included. Material types are dominated by sandstone for both lapidary abraders (95%) and lapstones (99%), with fine-grained textures most common for the former (68%), and medium-grained textures most common for latter (81%). The majority of these tools from the structure are formal (86%), whereas most of those from the mounds are expedient (75%). It is possible that the large number of expedient lapstones from the middens reflects the multipurpose nature of these work surfaces, as described earlier. However, because it is likely that many were also used for lapidary purposes, as indicated by the association of ground stone with similar dimensions with lapidary debris and other kinds of lapidary tools at 29SJ629 and 29SJ1360, they were not removed from the following analyses.

Figure 3. Lapidary abraders from (left) Pueblo Bonito and (right) 29SJ629 (Windes 1993:Plates 4.6 and 4.9). Catalog numbers for items from Pueblo Bonito: (clockwise from upper left) H/5899, H/5722, H/5521, H/5759, and A335877. Courtesy of the Division of Anthropology, American Museum of Natural History; and the Department of Anthropology, Smithsonian Institution. Photographs by Hannah V. Mattson.

Figure 4. Examples of lapstones from Pueblo Bonito: (top to bottom) H/4762, H/5249, and H/6023). Images courtesy of the Division of Anthropology, American Museum of Natural History. Photographs by Hannah V. Mattson.

Of the passive abraders from the structure that were deliberately modified, over half are trapezoidal or sandal shaped, consistent with the three main styles of sandals worn at the site: square-toed, rounded/tapered-toed, and jog-toed/-shaped (Bellorado 2020; Hays-Gilpin 1998:35; Jolie 2018; Teague and Washburn 2013). Although lasts—solid forms used as patterns for making footwear—were not necessary for sandal construction, it is possible that some of these abraders served this purpose concomitantly. However, there is little evidence that lasts were commonly used in the Ancestral Pueblo region or anywhere in the Americas for open-plan footwear, especially for those made from fiber (versus leather) and outside of specialized or workshop-scale modes of production (Edward Jolie, personal communication 2024). Sandal-shaped stones at Pueblo Bonito, and elsewhere in the northern Southwest, also occur in sizes inconsistent with footwear, indicating that they served other purposes, such as altar pieces (Benjamin Bellorado, personal communication 2024; see also Bellorado 2020; Snyder 1899). Consequently, this shape appears to be an iconographic reference, perhaps related to traditional migration histories or founding ancestors (Bellorado 2020; Crown et al. 2016).

Of the passive lapidary abraders and lapstones from 29SJ629 and 29SJ1360, 63% have no modification, and another 23% have only slight modification (Akins 1997). Only eight items from the two small house sites exhibit moderate-to-extensive work; of these, three lapstones from 29SJ1360 are sandal shaped like those from Pueblo Bonito (McKenna 1984:Figures 4.30–4.32)

Of the formally shaped passive abraders excavated from intramural contexts at Pueblo Bonito, 41% have residues of colorful minerals—red, yellow, and white. Although this might be interpreted as evidence that these tools also served as palettes, or were used to both grind pigments and work ornaments, traces of pigments were also found on other kinds of lapidary tools, such as files and bead shapers. I propose elsewhere (Mattson and Jones 2020) that pigments, especially red hued, may have been used as polishing compounds or grit in the ornament manufacturing process, in addition to having intrinsic animating properties in Ancestral Pueblo ontology (Mattson 2021). I have since also identified a prevalence of white mineral residue—likely selenite or gypsum—on lapidary tools from the site, suggesting that this material was also used in the same manner. Windes (1993:227) makes a similar suggestion for 29SJ629, noting that ground selenite was found in a ceramic bowl in one of the pithouses, and larger quantities were concentrated near mealing equipment, where the material was presumably ground for use, perhaps for ornament production. Pigment residue is not noted specifically for the small-house lapidary abraders.

The lapidary abraders from the structure are significantly longer and thicker than those from the mounds.¹ Combined (structure and mounds), they are significantly larger than the lapidary abraders from 29SJ629.² These differences are even more significant when comparing those from the Bonito structure to the small-house assemblages (Table 2). Lapstones from Pueblo Bonito display the same intrasite pattern, with those from the structure significantly outsizing those from the mounds, on average.³ However, comparing the Pueblo Bonito lapstones to those from 29SJ629 and 29SJ1360, the only significant difference in size is that those from the two small houses are wider.⁴ This may also reflect differences in shape, particularly because so many of the passive abraders from room contexts at Pueblo Bonito are sandal shaped and therefore longer than they are wide.

Table 2. Metric Data for Complete Dimensions of Lapidary Tools from Pueblo Bonito (Structure and Mounds).

Note: Format adapted from Windes [1993:Table 4.4] for data comparability.

Both lapidary abraders and lapstones from Pueblo Bonito also exhibit more wear than those from the two small-house jewelry workshop sites. The majority of the lapidary abraders and lapstones from the structure have heavy use wear (81% and 86%, respectively), with moderately used tools comprising most of the remainder. Those from the mounds primarily display moderate use wear (62% and 84%, respectively), with about one-third of the lapidary abraders recorded as having heavy use wear. By comparison, only four lapidary abraders or lapstones from 29SJ629 and 29SJ1360 (less than 5% of the total) have heavy wear, and 67% have moderate wear.

The tools from the Bonito mounds are also quite fragmentary, with more than half of the lapstones and 81% of the lapidary abraders representing less than half of the original tool. The original analysts argue that this “reflects patterns of discard typical of exhausted artifacts” (Hegberg and Crown 2016:151). By contrast, most of the passive lapidary tools from the structure (89%) are complete.⁵ Typically, only general provenience information is provided for materials excavated from the structure. Only one-third of the lapidary tools have provenience information beyond room number, with “debris” the most common, followed by “floor.” The debris category was used liberally during the Hyde Exploring Expedition, and it appears to encompass a range of stratigraphic levels, including near-floor (i.e., floor fill) contexts, as well as material from collapsed superjacent rooms. The limited evidence for the ritual deposition of lapidary tools in the structure combined with the abundance of complete, formal tools in intramural contexts suggests that most of these items were not intentionally discarded—as they were in the mounds—but instead represent use contexts (production or storage areas).

Lapidary Grooved Abraders / Bead Shapers

A total of 27 grooved passive abraders were recorded as part of the Pueblo Bonito lapidary tool assemblage. This includes 10 formal and 17 expedient grooved abraders with U-shaped grooves of sizes consistent with bead manufacture. Formal grooved lapidary abraders are heretofore referred to as bead shapers. All but one of the grooved abraders—an expedient example—are from excavated contexts within the structure. Many of the expedient abraders have multiple grooves, whereas formal examples only have a single groove each (Figure 5). All but one of the expedient abraders are made from coarse or very coarse sandstone, whereas the bead shapers include all textures.

Figure 5. Small-grooved abraders from Pueblo Bonito, both informal (left) and formal/bead shapers (right). Catalog numbers are as follows: (clockwise from upper left) H/2733, A335658, H/6541, and H/6413. Note red and white mineral residue on formal examples. Courtesy of the Division of Anthropology, American Museum of Natural History. Photographs by Hannah V. Mattson. (Color online)

Grooved abraders with U-shaped grooves are usually assumed to be arrow-shaft straighteners or smoothers. To ensure that tools used for this purpose were not incorrectly included in the lapidary tool assemblage, I compared the diameters of intact arrow shafts excavated from Pueblo Bonito with the widths of both formal and expedient “small-grooved” abraders, as well as with the diameters of disc beads from the site. The arrow-shaft sample includes 36 individual arrows, all wooden foreshaft elements.⁶ The disc-bead sample consists of 2,800 complete disc beads of turquoise, shell, and shale associated with a variety of contexts at the site (e.g., midden, room fill, room floors, kivas, burials; Mattson 2016). As displayed in Figure 6, the groove widths of the small-grooved abraders overlap with disc bead diameters⁷ but are significantly smaller than arrow shaft diameters.⁸

Figure 6. Box plot comparing the diameters (mm) of arrow shafts and disc beads with groove diameters for small-grooved abraders and formal bead shapers from Pueblo Bonito.

The majority (67%) of the lapidary grooved abraders from Pueblo Bonito exhibit heavy use wear, including 40% of the formal and 82% of the expedient tools. Mineral pigment residue was identified along the grooves of three of the most formal bead shapers, one with red ochre and the other two with selenite or gypsum. All three of these tools are made from fine-grained material. By contrast, none of the coarser-grained formal or expedient grooved abraders exhibit mineral residues. This supports the idea that certain minerals were used in the later stages of ornament production as polishing compounds.

In the analysis of the Chaco Project ground stone, Akins (1997) identified four items from 29SJ629 and 29SJ1360 as undifferentiated grooved abraders. Because the dimensions and contours of the grooves were not recorded, it is unknown if these are related to arrow, awl, or bead manufacture. Windes (1993:225) notes that 16 of the lapidary abraders from 29SJ629 have shallow grooves within their surfaces, most with widths measuring 1 mm to 4 mm, followed by 6 mm to 9 mm, both of which are consistent with bead diameters from the site.

Active Abraders/Files

The Pueblo Bonito assemblage includes 73 active lapidary abraders with morphology and use wear consistent with use as files (including reamers and rasps); 57 of these are from the structure, and 16 are from the mounds (Figure 7). As would be expected for files, a variety of materials and textures are represented, but more than half are sandstone and fine-to-medium grained. Almost 95% of the files from the structure are formal, and those from the mounds also appear to be intentionally shaped based on the information available. Most (76%) show signs of heavy use, including striations, edge rounding, and polish. Eight of the files, all heavily worn and all but one formal in design, have red or yellow ochre residue.

Figure 7. Examples of files from Pueblo Bonito: (left to right) A335617-1, A335617-2, A335616-1, and A335616-2. Courtesy of the Department of Anthropology, Smithsonian Institution. Photographs by Hannah V. Mattson.

Only six active lapidary abraders were excavated from 29SJ629 and 29SJ1360.⁹ Five of these are formal, and five exhibit moderate wear. A number of similar active lapidary abraders were also excavated from the Pueblo Alto (n = 12) and Pueblo del Arroyo (n = 25) great houses (Akins 1987; Judd 1959; Windes 1993:227). Of the files from these sites for which data are available (n = 22), most are formal (82%), and all display heavy or moderate wear. None of the active abraders from Pueblo del Arroyo or those recorded by NPS analysts exhibit residue. Overall, active lapidary abraders appear to be intentionally designed and intensively used tools in the canyon, regardless of whether they were used at great houses or small houses.

Drills/Microdrills

The Pueblo Bonito assemblage includes 121 drills, 71 from the structure and 50 from the middens (Figure 8). Approximately half of these (n = 62) are informal, and at least 40% (n = 47) could be classified as microdrills based on their sizes. One of the microdrills from the structure has intact hafting, and three of the drills are made from modified projectile points. Chert is the most common material represented (52%), particularly Chuska or Narbona Pass chert (21%), followed by silicified wood (38%). Interestingly, one drill made from turquoise was recovered from the mounds. Examination of the microdrills from the structure under high magnification revealed that over half exhibit pronounced use wear, including edge rounding, polish, and scarring along blade margins. Five of the tools have ground material embedded within flake scars—three with shell or a white mineral, and one each with what resembles jet and red shale. The proximal ends of drill shafts (i.e., drill bits), where present, have a mean maximum dimension of 1.5 mm (Table 2). Disc beads from Pueblo Bonito, from all contexts and of all materials, have exterior hole diameters ranging from 0.5 mm to 5.5 mm, with an average 1.6 mm.

Figure 8. Expedient drills from Pueblo Bonito: (top row) H/9714, H/9873, H/10187, and H/10300; (bottom row) H/7662). Courtesy of the Division of Anthropology, American Museum of Natural History. Photographs by Hannah V. Mattson.

A total of 26 drills, “fortuitous perforators,” and flakes with drill-like projections with blade dimensions consistent with possible ornament production were excavated from 29SJ629 and 29SJ1360, most from contexts associated with evidence of lapidary activity (Lekson 1993, 1997; McKenna 1984). In form and manufacture, these are almost identical to those from Pueblo Bonito,¹⁰ but all but two are produced from silicified wood. Compared to the sample from Pueblo Bonito, they are similar in length but significantly larger in blade width.¹¹ Lekson (1997:687) notes that drill-blade length likely correlates with the type of material worked, given that longer blades would be used not only on thicker materials but also on softer ones. This suggests that the drills from Pueblo Bonito and those from 29SJ629 and 29SJ1360 were used on similar materials, although those from Pueblo Bonito may have been used to create smaller perforations.

Saws and Tabular Tools

Thirty-three saws—13 flaked stone serrated saws and 20 ground stone tabular saws—were identified in the Pueblo Bonito assemblage (Table 1). The serrated saws are of sizes conducive to cutting relatively small/thin materials, and there is no evidence of hafting. More than half are unifacially retouched along a single margin. Two retain black residue, possibly jet, in the microflakes along their working edges. No serrated saws or “denticulate” tools were recovered from either 29SJ629 or 29SJ1360.

The tabular saws are primarily made from fine-grained sandstone or slate and have evidence of heavy wear, including parallel striations along their cutting margins. They are generally palm sized and very thin. Interestingly, four have colorful residue (red, yellow, and white) along their used edges, suggesting that either some of these tools were used to cut soft minerals for use in paint or ornament production, or that ground minerals were also used as grit during the sawing process. Six active abraders from 29SJ629 have dimensions consistent with use as tabular saws.

Spatial Distribution of Lapidary Evidence

Approximately half of the lapidary tools in the combined Pueblo Bonito assemblage are from the mounds, and half are from the structure. Eighty-two rooms and eight kivas contained at least one lapidary tool (Figure 9). These represent all portions of the pueblo, as well as those constructed at different times (Stein et al. 2003; Windes 2003). Although the distribution is widespread, the tools are particularly concentrated in the oldest part of the structure, or “Old Bonito,” built in the early ninth to early tenth centuries but used into the twelfth century. All of the rooms with eight or more lapidary tools (Rooms 2, 37, 38, 54, 78, 109, and 318)¹² and 60% of those with five or more tools are located in this part of the building. It has been suggested that two “houses” are represented by the architecture and distribution of material remains at the site, both in Old Bonito (Heitman 2007, 2011, 2015; Mills 2015). Within a house society model, houses are flexible, transgenerational corporate formations based on shared ancestors (real or fictive) and the perpetuation of both tangible and intangible resources (Joyce and Gillespie 2000; Lévi-Strauss 1982, 1987). They are anchored within specific, persistent architectural spaces, such as the northern and western burial crypts at Pueblo Bonito and their surrounding ritual storage rooms (Beck 2007; Heitman 2015; Mills 2015). Houses encompass spheres of both domestic and ritual activities, but they also embody shared cosmologies. Here, we see that rooms near the northern burial cluster, interpreted as the core of the primary house, contain the greatest concentration of lapidary tools.

Figure 9. Map of Pueblo Bonito showing density of lapidary tools by room. Colors denote densities as follows: dark blue = ≥10 items, light blue = 5–9 items, and pink = 3–4 items. Base image is from the Chaco Research Archives.

Another concentration of lapidary activity—albeit less pronounced—is located in the southeastern portion of the structure, particularly in Rooms 168, 170, and 173 (Figure 9). This part of the building was constructed between AD 1050 and 1070 but was used until AD 1150 (Stein et al. 2003; Windes 2003). All three of these rooms comprise the southernmost wall of the pueblo.

The rooms with the most lapidary tools each contain a diversity of objects, but common to all of them are abundant ground stone, flaked stone, and bone tools, as well as variable quantities of turquoise fragments and finished ornaments. Rooms 37, 78, and 318 included additional lapidary raw materials; Rooms 2, 54, 38, 109, and 173 contained ornaments in the process of manufacture; and ochre or gypsum/selenite were recovered from Rooms 2, 38, 78, 168, 170, and 173 (see also Mathien 1984:180; Pepper 1920:39).

At 29SJ629 and 29SJ1360, lapidary evidence is also found in spaces alongside materials associated with lithic and bone tool manufacturing and food processing. In contrast to the two small-house sites, evidence of jewelry making at Pueblo Bonito is not strongly associated with either plaza or kiva contexts (an interesting contrast to blue-green paint production [Hanson 2023]). Although seven of the rooms in the structure containing lapidary tools flank plazas (Rooms 40, 144, 165, 210, 288, 332, and 348), these yielded only nine items total. Turquoise debris and ornaments, both whole and broken, are common in Bonito's kiva pilaster deposits, yet only 4% of the lapidary tools from the site were recovered from kivas.

Discussion

Researchers have used a range of typologies to characterize past craft production systems, in which distinct modes of manufacture are distinguished based on location and distribution, intensity and level of investment, magnitude, and amount of elite sponsorship (e.g., Brumfiel and Earle 1987; Rice 1981; Sinopoli 1988). Below, I discuss my interpretations of the ornament production evidence from Chaco Canyon, now more complete with the addition of the Pueblo Bonito lapidary data, using Costin's (1991, 2001, 2005) production parameters: concentration, intensity, scale, and context.

The spatial organization of production comprises the dimension of concentration, with dispersed and nucleated at ends of the spectrum (Costin 1991:13–15, 2001:295). In dispersed production, manufacture occurs at each community of a certain size, depending on the nature of demand for the product. A widely accepted interpretation is that Chacoan ornament production occurred primarily at small houses, with a subset of these communities specializing in the craft (Hagstrum 2001; Peregrine 2001). Survey data supports the presence of some degree of lapidary activity at virtually all residential sites in Chaco (Windes 1992). That only a handful of small houses have been excavated—the two lapidary workshop sites among them—suggests the possibility that evidence for ornament manufacture on par with 29SJ629 and 29SJ1360 may exist at other, and perhaps even many, small-house sites. The evidence from Pueblo Bonito clearly indicates that significant lapidary production was occurring there, and both descriptions in excavation monographs and museum catalogs suggest that it was at other great houses as well. The combined available evidence from Chacoan sites points to a fairly dispersed mode of ornament production when viewed from the perspective of the entire canyon. From an intrasite perspective, production is also dispersed across Pueblo Bonito, with 90 discrete intramural contexts containing lapidary tools (Figure 9).

The relative time committed to production activities defines the dimension of intensity, with part time and full time at either end of the continuum (Costin 1991:16–18, 2001:289). Where production is conducted on a part-time basis, it is scheduled alongside other domestic activities. Full-time production, on the other hand, entails intensive time investment; at its extreme, this occurs to the near exclusion of other kinds of domiciliary and subsistence activities. Compared to full-time producers, it is expected that part-time producer households will include archaeological evidence of an assortment of subsistence and other economic activities (Costin 1991:32–35, 40, 2001:280, 289, 294, 2005:1064, 1068).

Both small-house and great-house lapidary production evidence in Chaco is found in contexts associated with a range of other household activities, suggesting that ornament manufacture was conducted on a relatively part-time basis throughout the canyon. At Pueblo Bonito, lapidary tools are most often found alongside those used for lithic reduction, weaving, paint production (particularly of colors other than blue-green; Hanson 2023), and food processing; there are no single-purpose or dedicated jewelry workshop areas. However, lapidary tools are also particularly concentrated within certain rooms, suggesting that production may have occurred more intensively in some parts of the site than in others. A comparatively greater intensity of ornament production at Pueblo Bonito compared to small houses is also supported by the presence of more formal and larger lapidary abraders; additional types of technology such as specialized bead shapers and use of mineral pigments as polishing agents; drills made of more expensive materials and that include smaller bit sizes; and tools that exhibit more evidence of heavy use.

The scale of production refers to the size of the production group and the degree to which kinship directs labor enrollment (Costin 2001:15–16). Smaller-scale production units typically consist of artisans within the same family, whereas larger-scale units are more likely to include unrelated individuals. Archaeologically, scale is tied to both the location and size of production loci. Small-scale, kin-based work areas tend to be relatively small in areal extent and are located within or near a family's residential space; large-scale workshops are, by definition, greater in size and will be situated outside of these domestic areas (Costin 1991:29). In all identified cases at Chaco, ornament production appears to be fairly small in scale, given that evidence for lapidary activities is found within or adjacent to domestic locales, whether in small houses or great houses. At 29SJ629 and 29SJ1360, jewelry-making occurred both within residential pithouses and in nearby plaza areas. As described above, at Pueblo Bonito, rooms with lapidary tools also contain evidence of other domestic activities.

Pertinent to this discussion is the size of the lapidary tool assemblage from Pueblo Bonito, which is much larger than that of either of the two small-house jewelry workshops. However, given that Pueblo Bonito is also a much larger site, production and consumption would naturally occur at a higher rate. Interestingly, statistical analysis of the relative frequency of passive and active lapidary abraders at Pueblo Bonito and 29SJ629 indicates no significant difference between the two assemblages (χ ² = 0.687; p = 0.406). Another way to compensate for higher consumption rates at larger sites is to use an index to some other kind of everyday material. Comparing the total number of lapidary tools to combined numbers of manos and metates, as a relative proxy for residential population size, yields indexes of 0.18 for Pueblo Bonito and 0.21 for 29SJ629 (Heitman 2016:480).¹³ It appears, then, that ornament production occurred at Pueblo Bonito at a scale and concentration comparable to that of 29SJ629, even after accounting for differences in population size.

Although ethnohistoric references to the organization of ornament production are few, where mentioned, it is described as occurring in household contexts and typically on a part-time basis (e.g., Lange 1959:171–172; Parsons 1932:12; White 1935:27). As it was in the past, the art of making jewelry is transmitted generationally within families, with some families now specializing in the craft but also conducting other activities—such as farming, raising livestock, producing other crafts—or having other employment. Even though a portion of artists now work as full-time lapidarists and produce a large output for the tourist market, they still typically work from their homes. Often, another family member, perhaps one who is actively learning, works alongside the artist, and they produce pieces as a team (see also Hedquist 2017:245, 247, 270).

There is archaeological evidence of a similar kind of intensive production within households in the region. For example, Mills (2007) calls attention to the importance of the household as a context of intensifying multicraft production, including both ceramics and ornaments, in east-central Arizona in the late thirteenth and fourteenth centuries, a time of intense in-migration and aggregation. In a similar vein, Feinman and Nicholas (2007) identify the presence of multicraft production in households in the Valley of Oaxaca during the Classic period, providing another case study of part-time, specialized production in domestic contexts (see also Hirth 2006, 2009).

The relationship between artisans and consumers conditions the dimension of context. This dynamic is often framed in terms of the degree to which production is sponsored or managed by elites or institutions, with independent and attached production at either end of the scale (Arnold and Munns 1994; Brumfiel and Earle 1987; Clark 1995; Hirth 2020). In independent production contexts, the manufacture and consumption of goods is unrestricted. Producers are not funded or supervised by those with greater power, and the use of their goods is not constrained by social status (Costin 2005:1070). These goods tend to be utilitarian in nature and are produced in nonelite contexts. Attached production entails manufacture in direct association with elite or otherwise institutionally controlled contexts. Goods tend to be those with social values relating to prestige and authority; the production of these items may be tightly controlled and involve more intricate or secret knowledge or technology (Costin 2001:298–299, 2005:1070–1072). Ornament production in Chaco Canyon seems to have largely occurred in relatively independent contexts, supported by evidence of lapidary activity in both small-house and great-house domestic contexts, in addition to that of widespread use of jewelry throughout the Chacoan population.

However, production also appears to be more concentrated and intensive in certain portions of Pueblo Bonito associated with the northern and western houses. The lapidary evidence in these rooms includes abraders of socially charged shapes (sandals); greater quantities of valuable and imported raw materials that may indicate differential access to exchange networks; and probable evidence of a more ritualized ornament production process. Along with domestic items, these rooms also contain other distinctive items associated with elevated social standing (e.g., house heirlooms) and restricted ritual knowledge or performance (Gruner 2019; Neitzel 2003). This suggests that ornament production was conducted within elite households by elite members—specifically, high-ranking house members (see also Plog et al. 2017:293–294).

Ames (1995:174) defines this as embedded production: artisans born into noble houses “produced elite goods, not as dependents of an elite class, but as part of their roles as elite individuals.” In this mode of production, high-status producers make items for their own consumption or that of other elite members. The main archaeological indicator of embedded production is evidence of manufacture within high-status domestic contexts (Ames 1995; Costin 2005; Inomata 2001; Janusek 1999; Menzies and Haller 2012; Ritchie et al. 2022). Similar to attached production, the items produced by embedded producers typically confer special rank or privilege, such as goods and equipment used in ritual displays associated with household/lineage roles and rights. In another example, Inomata (2007:125) finds evidence of the multicraft manufacture of “precious” objects, requiring specialized and esoteric knowledge, within elite households in Aguateca, a Classic Maya site in Guatemala. This includes headdresses and pyrite mirrors, painted codices and ceramics, carved stellae, and textiles, all of which were produced for fellow elites.

In a study of the distribution of certain imported materials and ceremonial objects in Pueblo Bonito, Gruner (2019) suggests that embedded production centered on the northern part of the structure (Bishop 2019; Bishop and Fladd 2018; Bishop et al. 2023; Crown 2020; Heitman 2007, 2011, 2015; Watson et al. 2015; Weiner 2018). This production is argued to be ritual in nature, given that it includes inalienable ritual paraphernalia and is spatially associated with burial, storage, or ceremonial rooms. The likelihood of specialized, embedded ornament production in the canyon is also noted by Plog and colleagues (2017:293–294) based on workshop evidence and the large quantity of shell, turquoise, and jet recovered (see also Mills 2023:234).

Some aspects of previously proposed models of Chacoan ornament production are supported—in particular, that jewelry manufacture was diffuse, household based, and part time (Earle 2001; Hagstrum 2001; Peregrine 2001; Plog et al. 2017). I depart from most of these models, however, in arguing that this includes great houses (see also Mathien 1984, 2001:106) and that the production of ornaments consumed at great houses was not necessarily the responsibility of small-house artisans or of those visiting great houses from outside the canyon. I propose that the turquoise and other ornament items used by great-house residents could have been produced by on-site artisans, both elite and otherwise.

Apart from production evidence, these previous models also rest on assumptions of differential consumption based on the relative scarceness of finished ornaments at small houses compared to great houses. These do not take into consideration issues of scale, site function, and object life history. Finished ornaments are consistently, if more sparsely, present at small houses, with a mean of 131 per site¹⁴ for those excavated by the Chaco Project and UNM Field Schools in the 1930s–1940s (Brand et al. 1937; Kluckhohn and Reiter 1939; Mathien 2001; Windes 1992). Though there are ritual caches at small houses, the density is much lower than it is for great houses (Ditto 2017:167). Small-house sites also have fewer kivas and no clearly high-status mortuary spaces—both contexts in which ornaments are amassed as offerings within Chacoan ritual practice. Despite these differences, however, jewelry items have been found in some small-house interments, including an individual buried at 29SJ1360 with a necklace of almost 4,000 shale beads (Mathien 1997; McKenna 1984) and other small house burials with ornaments, reported by Akins (1986). This suggests that additional interments with finished ornaments are likely present in unexcavated small sites.

Another way to examine local demand relative to production is to compare the number of finished ornaments to lapidary tools at excavated sites. The resulting indexes range widely—365.72 for 29SJ1360, 131.93 for Pueblo Bonito, and 1.34 for 29SJ629. The high index from 29SJ1360 is largely the result of the necklace from the mortuary context described above, whereas 29SJ629 lacked such a deposit and contained an unusually large number of lapidary tools. Combining all of the Chaco Project sites together (but excluding the Pueblo Alto great house) yields an index of 21.67, an average of “normal” consumption relative to production at small-house sites. The index for Pueblo Bonito is clearly very high compared to this. However, this must be treated with a degree of caution, given that only a small percentage of the mounds has been excavated. Based on the sample from the Judd trenches, these unexcavated deposits likely contain an abundance of lapidary tools but few finished ornaments (which are concentrated in intramural structured contexts), so the true index is likely significantly lower than what is calculated here. Regardless, it is clear that ornament consumption was higher at Pueblo Bonito than at other sites. At least some of this demand may have been met through more intensive production in certain contexts, such as through embedded manufacture in the northern portion of the structure.

Conclusion

In this article, I presented the results of an analysis of lapidary tools from the preeminent great house in Chaco Canyon—Pueblo Bonito. These new data shine a spotlight on great-house ornament manufacture, which has been overlooked in discussions of the organization of production in the canyon and Chacoan political economy. Previous models assign ornaments—particularly those made from turquoise—somewhat different roles in systems of institutional support. Many view the production and circulation of these items as too difficult to regulate to have been controlled directly by elites, as would be the case for prestige goods in a system of wealth finance. At the same time, the production of ornaments is often explained as being carried out at the behest of elites or ritual leaders, similar to obligatory contributions of labor for large constructions, utilitarian goods, and surplus produce.

Here, I argue that it was not necessarily incumbent upon small-house residents to supply the ornaments consumed by elites either personally or in ritual performances carried out at great houses such as Pueblo Bonito. Rather, ornament production in each community may have been generally consistent with its own demand and that of nearby related households. Because this demand was larger at great houses, due to the presence of high-status houses and greater numbers of ceremonial facilities, production was accordingly more intensive—though still part time—in these locations. It is also clear that ornaments were not all valued similarly. Certain forms, materials, and quantities were required by those with higher status for their own adornment—either in life or in death—or for their ritual responsibilities. I suggest that these items may have been made by members of high-ranking households through a system of embedded production. Greater access to long-distance exchange networks, and the imported and valuable raw materials they provided for ornament manufacture, likely reinforced these social differences (e.g., Gruner 2019; Plog 2018; Watson et al. 2015).

There are, of course, limitations to this study that must be acknowledged. This research focuses on Pueblo Bonito, the ornament production at which may not be representative of other Chaco great houses (for recent discussion of the “Pueblo Bonito Bias,” see Bernardini and Schachner 2022; Plog 2018). Likewise, most small houses remain unexcavated, and those that have been selected for investigation over the last 60 years may or may not be characteristic of the larger small-house population in the canyon. Furthermore, the strict dichotomy between small and great houses has been questioned by some (Heitman 2007, 2015). There are also issues working with legacy collections; the curated assemblages from Pueblo Bonito are almost certainly incomplete and biased against informal tools and manufacturing debris (e.g., Heitman 2016). Despite these caveats, however, extant museum collections remain a valuable source of information to address new anthropological questions. Maximizing the data potential of previously excavated material is particularly important as the discipline begins to emphasize less invasive investigations.

Outside of the American Southwest, this study contributes to at least two areas of broader significance. Archaeological research on ornaments frequently focuses on either consumption—especially the use of finished items in relation to social identity and power relations—or interregional trade. Fewer studies address the organization of ornament production. This study adds to our knowledge of variation in ornament manufacture in the past, including associated material residues and lapidary technologies excluding metal and mechanical devices (e.g., bow drills). It also contributes to our understanding of variation in craft production systems, adding to a growing body of archaeological examples of embedded production in nonstate societies.