The Northwest Coast of North America (NWC) has recaptured the popular (Graeber and Wengrow Reference Graeber and Wengrow2022) and academic (Holland-Lulewicz et al. Reference Holland-Lulewicz, Thompson, Birch and Grier2022) imagination regarding the structure of past social and political systems. How did the unique horizontal social relationships in the Salish Sea region of the NWC emerge?
Our previous work (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017) examined Salish Sea labret lip ornaments using social network analysis (SNA) to assess the chronology and distribution of the practice. Archaeologists in the American Southwest have used SNA to study change in active social symbols, examining social transformations with ceramic styles and sourcing (Birch and Hart Reference Birch and Hart2018; Borck et al. Reference Borck, Mills, Peeples and Clark2015; Brughmans Reference Brughmans2010; Gjesfjeld Reference Gjesfjeld2014, Reference Gjesfjeld2015; Knappett Reference Knappett2011, Reference Knappett2013; Mills Reference Mills2017; Mills et al. Reference Mills, Peeples, Randall Haas, Borck, Clark and Roberts2015). SNA has also been widely applied to lithic material sources (Buchanan et al. Reference Buchanan, Hamilton, David Kilby and Gingerich2016; Golitko and Feinman Reference Golitko and Feinman2015; Golitko et al. Reference Golitko, Meierhoff, Feinman, Patrick and Williams2012; Mills et al. Reference Mills, Clark, Peeples, Randall Haas, Roberts, Brett Hill and Huntley2013; Phillips Reference Phillips2011). SNA was also used to examine lithic tool styles (Rorabaugh Reference Rorabaugh2019) and spatial and temporal obsidian tool-stone patterns (Springer et al. Reference Springer, Lepofsky and Blake2018) for the Salish Sea.
SNA integrated with a geographic information system (GIS) enables archaeologists to examine social connections at multiple scales. Elsewhere we argued (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017) that labrets represent a credibility-enhancing display (CRED; e.g., Henrich Reference Henrich2009)—a physical manifestation of a “good name” critical to reciprocity among ethnographic-period Coast Salish political economies. That initial study explored diachronic change in labrets, using the artifacts as a proxy for Coast Salish social networks.
We argue that the emergence of these relationships, or social foraging (e.g., Smith and Grier Reference Smith and Grier2022), is signified at keystone cultural places (Lepofsky et al. Reference Lepofsky, Armstrong, Greening, Jackley, Carpenter, Guernsey, Matthews and Turner2017) by labrets. Past researchers (Ames Reference Ames, Price and Feinman1995; Keddie Reference Keddie1981, Reference Keddie1989) argued that labrets may have represented “proto elite” status or achieved status. Instead, we suggest that labrets are a physical manifestation of the emergence of horizontal social relationships in the Salish Sea. We contend that labrets are better characterized as reflections of House emergence in the Lévi-Straussian sense given their associated material, immaterial, and relational wealth in the Salish Sea. Here, we refer to Houses as a social institution not to be conflated with house structures, which are but one material reflection of the House as a social institution.
Background
In the 1980s, archaeologists studying house deposits and burial contexts in British Columbia saw labrets in the record by 3200 cal BP, much later than when social inequality and culturally complex systems emerged (Carlson Reference Carlson1987; Coupland et al. Reference Coupland, Bilton, Clark, Jerome, Cylbulski, Frederick, Holland, Letham and Williams2016; Keddie Reference Keddie1981, Reference Keddie1989). Elsewhere, we argued that both labrets and cranial modification served as unambiguous markers that people used to identify trusted individuals for reciprocation in emerging systems of resource management (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Here, we argue that labrets reflect the emergence and elaboration of Houses in the Salish Sea. The following section discusses the chronology of Houses in the Salish Sea (Figure 1).
Figure 1. Salish Sea regional chronology periods in study shaded.
The chronology contextualizes evolving understandings of Salish Sea lifeways using house structure and burial data to elucidate our discussion of social relationships. We use the term period as opposed to phase or culture type, acknowledging the numerous critiques of culture-historic sequence construction (Henry et al. Reference Henry, Angelbeck and Rizvi2017; Lyman et al. Reference Lyman, O'Brien and Dunnell1997; Oland et al. Reference Oland, Hart and Frink2012; Pestle et al. Reference Pestle, Curet, Ramo and Lopez2013). Ritchie and colleagues (Reference Ritchie, Lepofsky, Formosa, Porcic and Edinborough2016) demonstrated the utility of avoiding culture-historical categories when examining demographic variation through time in the Fraser Valley. Following critiques of sequence building and periodization in the Salish Sea (Grier Reference Grier2022; Morin Reference Morin2014; Rorabaugh et al. Reference Rorabaugh, Shantry and Brown2020), we refer to periods only as a frame of reference to examine diachronic change.
The role that large-scale procurement and storage of seasonal resources played in building Salish Sea traditions is well documented (Ames Reference Ames1994; Matson Reference Matson, Croes, Hawkins and Isaac1992). However, debates still exist regarding the timing of horizontal social organization and the degree of historic continuity in practices (Carlson Reference Carlson1991; Matson Reference Matson and Matson2010; Matson and Coupland Reference Matson and Coupland1994; Moss Reference Moss2011). By 2000 BP, there is clear archaeological evidence for hereditary forms of social inequality (e.g., Beattie Reference Beattie1981; Burley and Knusel Reference Burley and Knusel1989; Carlson and Hobler Reference Carlson and Hobler1993; Curtin Reference Curtin1991; Grier Reference Grier, Matson, Coupland and Mackie2003).
Researchers continue to challenge past narratives that argue cultural complexity increased with growth, elaboration, and specialization of subsistence economies through time (e.g., Bilton Reference Bilton2014; Bovy Reference Bovy2005; Butler and Campbell Reference Butler and Campbell2004; Campbell and Butler Reference Campbell, Butler and Dean2010; Daniels Reference Daniels2009; Deur and Turner Reference Deur and Turner2005; Orchard Reference Orchard2007; Orchard and Clark Reference Orchard and Clark2006; McKechnie Reference McKechnie2014; Monks and Orchard Reference Monks and Orchard2011; Moss Reference Moss2011; Moss and Cannon Reference Moss and Cannon2011). Archaeological evidence suggests that the Salish Sea is a mosaic of complex relationships where the increased utilization and diversification of resources profoundly patterned the historical trajectories of Coast Salish groups (e.g., Angelbeck Reference Angelbeck2009; Angelbeck and Grier Reference Angelbeck and Grier2012; Cannon Reference Cannon, Fitzhugh and Habu2002; Grier Reference Grier, Matson, Coupland and Mackie2003; Martindale and Letham Reference Martindale, Letham, Johansen and Bauer2011).
Despite continuity in many Holocene practices, marked shifts exist in the seasonality of early houses (5000–4000 BP) compared to later houses (2000 BP to contact; Grier and Kim Reference Grier and Kim2012; Lepofsky et al. Reference Lepofsky, Schaepe, Graesch, Lenert, Ormerod, Carlson, Arnold, Blake, Moore and Clague2009) and the material expressions of achieved and ascribed social status (Angelbeck and Grier Reference Angelbeck and Grier2012; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Here, we suggest that labrets and cranial modification are diachronic material correlates of status that reflect the emergence of Houses as a social institution in the Salish Sea.
Houses and House Societies in the Salish Sea
We follow Lévi-Strauss's concept of the House society, recognizing renewed theoretical interest in critically examining the role of household and kin organization (e.g., Andrews and MacDonald Reference Andrews and MacDonald2022; Beck Reference Beck2007; Carpenter and Prentiss Reference Carpenter and Prentiss2022; Ellis Reference Ellis2023; Gillespie Reference Gillespie2001; Joyce and Gillespie Reference Joyce and Gillespie2000). Lévi-Strauss based social and political dynamics on House operation and interaction (Gillespie Reference Gillespie2001:12–13), intending to differentiate Houses from other social structures often discussed by anthropologists, such as lineages and clans. English translations and Lévi-Strauss's (Reference Lévi-Strauss and Modelski1982, Reference Lévi-Strauss1987) own adherence to the importance of kinship (Reference Lévi-Strauss1987:153–154) have obscured many of the nuances of this approach. Lévi-Strauss noted the limitations of the Anglo-American “corporate group,” where individuals are bounded by descent and/or residence. Instead, Houses are groups that maintain estates perpetuated by member recruitment and relationships expressed in the language of kinship and affinity reinforced by purposeful actions and symbols. Individuals may not be necessarily consanguineal or affinal kin but are described using such terms. Ames (Reference Ames, Sobel, Trieu-Gahr and Ames2006) viewed Houses as corporate groups holding estates of material and immaterial wealth that perpetuate through hereditary inheritance of names, goods, and titles. Advancing this (sensu Gillespie Reference Gillespie2001), we suggest that displays such as labrets are more consistent with the original Lévi-Straussian articulation of Houses and House Societies without using the descent or residence restrictions of corporate groups. In the ethnohistoric period, Houses served as the primary socioeconomic unit of Coast Salish societies and a keystone social institution, as argued by Holland-Lulewicz and colleagues (Reference Holland-Lulewicz, Thompson, Birch and Grier2022).
Houses functioned as collective systems that integrated autonomous nuclear families connected to other Houses through marriage, seasonal rounds, and institutional events such as potlatches. The materiality of House relationships is expressed in Coast Salish plank houses, where each family within the structure had their own hearths and food stores (Coupland et al. Reference Coupland, Clark and Palmer2009; Grier Reference Grier, Sobel, Trieu-Gahr and Ames2006; Suttles Reference Suttles1951, Reference Suttles1958). People constructed plank houses in a modular manner so that sections could be added or removed to accommodate fluctuating populations. Families had economic autonomy within each plank house and were not politically constrained by the authority of house leaders (Coupland Reference Coupland, Sobel, Trieu-Gahr and Ames2006; Elmendorf Reference Elmendorf1960, Reference Elmendorf1971; Miller and Boxberger Reference Miller and Boxberger1994; Richardson Reference Richardson, Williams and Hunn1982; Suttles Reference Suttles1951, Reference Suttles1958; Suttles and Lane Reference Suttles, Lane, Suttles and Sturtevant1990). Individuals often identified primarily with their plank house residence, and they were defined by extensive kin relationships in multiple villages (Thom Reference Thom1995). House leaders or “House Chiefs” did not have a formal title, and status was assumed informally by a house member with the most prestige (Ames Reference Ames, Price and Feinman1995, Reference Ames, Sobel, Trieu-Gahr and Ames2006; Coupland Reference Coupland, Sobel, Trieu-Gahr and Ames2006). Chiefs earned leadership, prefaced by existing membership in a high-status family with nobility and inherited privileges. Political participation gave people access to corporeal and noncorporeal resources. Relationships between House Chiefs were fundamentally horizontal, exemplified by the phrase “They Recognize No Superior Chief,” as remarked by Manuel Quimper Benítez del Pino of the 1790 Quimper expedition (Wagner Reference Wagner1933:131).
Charles Period (5500–3500 BP)
The focus of our study is whether a House's social relationships are materially reflected in labrets and if those relationships are horizontal in nature. Beginning with the Charles period, settlement (Figure 2) consisted of smaller domiciles likely used for multiple seasons. There is considerable variation in house structure form during this period. Early structures on the Upper Fraser show considerable variation in size. These include DhRp-52 (Hoffman Reference Hoffman2017; Lyons et al. Reference Lyons, Tanja, Miller, Huddlestan, Leon and Squires2018); DgRn-23 Xá: ytem (Mason Reference Mason1994), and DhRk-8. The structures at Maurer appear to be rectangular plank houses with living floor size consistent with the smaller size range of multifamily winter structures of the later Marpole period (LeClair Reference LeClair and Carlson1976; Schaepe Reference Schaepe and Carlson2003).
Figure 2. Referenced Salish Sea residential sites.
Although there are similarities with later plankhouse structures in the Salish Sea, variation observed in Charles period structures likely reflect seasonality, group size, and architectural style (Ormerod Reference Ormerod2002; Ormerod and Matson Reference Ormerod and Matson2000). The expanding size of Charles period houses suggests single or extended families (Grier Reference Grier, Matson, Coupland and Mackie2003:9). Coupland and colleagues (Reference Coupland, Clark and Palmer2009) argued that the Houses had no clear status differentials and that family autonomy was materially reflected in these structures. In addition to large rectangular plank house structures, people also built pithouses, an example of which is at 45WH-34 Ferndale (Gillis Reference Gillis2007; Hutchings Reference Hutchings2004). Researchers have also suggested the presence of house structures at DgRr-2, the St. Mungo site (Ham et al. Reference Ham, Yip and Kuller1983).
Evidence for hereditary forms of social inequality are based on grave goods observed with subadults (Ames and Maschner Reference Ames and Maschner1999; Matson and Coupland Reference Matson and Coupland1994). However, there is clear time depth for complex social systems and forms of inequality (Clark et al. Reference Clark, Coupland and Cybulski2012; Coupland et al. Reference Coupland, Bilton, Clark, Jerome, Cylbulski, Frederick, Holland, Letham and Williams2016). One adult male interment was adorned with a mass of ground stone beads, an action requiring a substantial degree of nonkin labor. Similarly dated burials with copious beads exist at the Tsawwassen site (Curtin Reference Curtin1999) and Cowichan Bay (Hanson Reference Hanson1990; Yip Reference Yip1982). Some form of incipient elite was present, although families were economically autonomous. Social restrictions, in the vein of protocols and practices to prevent intergenerational hoarding and surplus control (Ames Reference Ames, Price and Feinman2010; Arnold Reference Arnold1993; Blake and Clark Reference Blake, Clark and Blake1999; Cashdan Reference Cashdan1980; Trigger Reference Trigger2003; Woodburn Reference Woodburn1982) may have hindered the intergenerational transfer of wealth. Status may have been achieved rather than ascribed. Without an entrenched system for intergenerational transfer of wealth and its associated material correlates, we suggest that based on current archaeological lines of evidence, Houses would not necessarily be present in this period and the structure of hereditary social inequality may not have been as systematized.
Locarno Beach Period (3500–2400 BP)
Current literature suggests continuity and change from the Charles to Locarno Beach periods (e.g., Bilton Reference Bilton2014; McKechnie Reference McKechnie2014; Orchard and Clark Reference Orchard and Clark2006). Well-documented residential structures include DgRr-1 Crescent Beach (Matson Reference Matson and Matson2010), DeRt-1 Pender Canal (Carlson Reference Carlson1987; Hanson Reference Hanson1990), 45WH-48 Simonarson (Gaston Reference Gaston1975), 45WH-55 Woodstock Farm (Campbell et al. Reference Campbell, Barg, Meidinger and Koetje2010; Lewis Reference Lewis2013), 45SJ-165 Decatur Island (Walker Reference Walker2003), DcRu-1151 Esquimalt Lagoon (Bowie and Kristensen Reference Bowie and Kristensen2011), 45CA-523 Čḯxwicən (Butler et al. Reference Butler, Bovy, Campbell, Etnier and Sterling2019), and 45CA-426 Sequim (Morgan Reference Morgan1996, Reference Morgan1998). Matson (Reference Matson and Matson2010) argued that in contrast to other periods with multiple autonomous families in one structure, Locarno Beach residences were primarily winter structures occupied by single nuclear families, in part due to the majority of these archaeologically reported structures being substantially smaller than structures from the preceding Charles period and subsequent Marpole period. Residences dating to the Locarno Beach period also have considerable variation in form, from rectilinear structures to ovoid ones.
Archaeologists assert that grave goods associated with subadults indicate ascribed status (Carlson Reference Carlson1987, Reference Carlson1991). Burley and Knusel (Reference Burley and Knusel1989) have argued that while evidence of ascribed status throughout the Salish Sea during this period is weak, there is widespread variability in burial practices.
People created art industries such as incised siltstone concretions (Mitchell Reference Mitchell1971; Morgan Reference Morgan1998; Walker Reference Walker2003), small ground stone objects termed “Gulf Islands complex objects” (Duff Reference Duff1956), and labrets. These may represent embedded craft specialization, with labrets commissioned by elites (Ames Reference Ames, Price and Feinman1995). Alternatively, Coupland (Reference Coupland, Thompson, Maria T. and Kense1985) argued that the art may have represented lineage markers.
Labrets predate hereditary forms of social inequality by several hundred years (Ames Reference Ames2003:31; Shantry Reference Shantry2014). We argue that labrets may serve as a material correlate for the emergence of Houses, yet we caution that earlier structures may not have reflected the material reality of Houses. Labrets as a credibility-enhancing display, however, may have helped facilitate the House connections and relationships.
Labrets and dental abrasion from wear have been suggested as a high-status marker people achieved (e.g., Ames Reference Ames, Price and Feinman1995, Reference Ames2001; Angelbeck and Grier Reference Angelbeck and Grier2012; Coupland Reference Coupland, Thompson, Maria T. and Kense1985; Cybulski Reference Cybulski, Matson and Pratt1991; LaSalle Reference LaSalle2008, Reference LaSalle2014; Matson and Coupland Reference Matson and Coupland1994; Moss Reference Moss2011; Shantry Reference Shantry2014). Other potential status markers include knob and top woven cedar hats—perishable items preserved in wet site contexts (Croes Reference Croes2009).
In the Locarno Beach period, people demonstrated status differentiation when negotiating emergent forms of hereditary social inequality. Social networks were not yet as widespread, rigid, or entrenched as during the Marpole period (Angelbeck and Grier Reference Angelbeck and Grier2012; Clark Reference Clark2010).
Marpole Period (2400–1000 BP)
Do Houses emerge during the Marpole period? Recent syntheses suggest that marine and terrestrial resource use did not become more specialized or intensified (Bilton Reference Bilton2014; Butler and Campbell Reference Butler and Campbell2004; Campbell and Butler Reference Campbell, Butler and Dean2010; Orchard and Clark Reference Orchard and Clark2006). Extended kin likely owned Marpole resource locations (Coupland Reference Coupland, Thompson, Maria T. and Kense1985; Grier and Kim Reference Grier and Kim2012). Lenert (Reference Lenert2007) argued for ownership of the salmon fishery at the Katz site in the Upper Fraser Valley. Lepofsky and colleagues (Reference Lepofsky, Lertzman, Hallet and Mathewes2005) forwarded a hypothesis that a drier climate allowed for more resource heterogeneity and intercommunity institutions, such as Houses, developed for the redistribution of resources.
By 2000 BP, large plank houses, ranging from single large houses to multihouse villages, are reported in villages throughout the Salish Sea. Substantial plank house structures appear by 2600 cal BP at the Scowlitz site on the Upper Fraser (Lepofsky et al. Reference Lepofsky, Blake, Brown, Morrison, Oakes and Lyons2000), and the variation in house structure styles noted in the Locarno Beach period also continued until 2300 BP (Johnstone Reference Johnstone1991, Reference Johnstone and Carlson2003), Estimated household size varied greatly; some had less than 20 individuals, whereas others had more than one hundred (Ames and Maschner Reference Ames and Maschner1999; Coupland et al. Reference Coupland, Clark and Palmer2009; Lepofsky et al. Reference Lepofsky, Schaepe, Graesch, Lenert, Ormerod, Carlson, Arnold, Blake, Moore and Clague2009; Sobel Reference Sobel, Sobel, Trieu-Gahr and Ames2006). House structures have been documented at DgRs-1 Beach Grove (Matson Reference Matson1980), 45SJ-25 Garrison Bay (Kenady et al. Reference Kenady, Saastamo and Sprague1973; Stein Reference Stein2000), 45SJ-24 English Camp (Kornbacher Reference Kornbacher1989; Stein Reference Stein1992; Taylor and Stein Reference Taylor and Stein2011), DgRw-4 False Narrows (Burley Reference Burley1988), DfRs-1 Whalen Farm (Smith Reference Smith1921), DgRv-3 Dionisio Point (Grier Reference Grier2001; Mitchell Reference Mitchell1971), 45KI-59 Tualdad Altu (Chatters Reference Chatters, MacEachern, Archer and Garvin1989), and 45WH-1 Xwe'chi'eXen (Blodgett Reference Blodgett1976; Grabert Reference Grabert1988). Archaeologists have excavated features likely consistent with house structures at DhRs-1 Marpole (Borden Reference Borden and Carlson1970) and 45SN-100 Biederbost (Nelson Reference Nelson1962; Nordquist Reference Nordquist and Croes1976). Both taphonomic processes and sampling strategies make it difficult to assess site structure (Taylor and Stein Reference Taylor and Stein2011:3–7, 170). Not all reported structures are winter residences; some, such as DhRt-5 Point Grey, are argued to represent spring occupations (Borden Reference Borden1947; Coupland Reference Coupland1991). Clear evidence exists for task and status differentiation within Marpole-dated structures (Chatters Reference Chatters, MacEachern, Archer and Garvin1989; Grier Reference Grier2001, Reference Grier, Matson, Coupland and Mackie2003, Reference Grier, Sobel, Trieu-Gahr and Ames2006) where hierarchical and horizontal social relations are materially reflected in the spatial order (Coupland et al. Reference Coupland, Clark and Palmer2009).
Grave goods reflecting status are more common and diverse, including the interment of subadults with elaborate grave goods (Ames Reference Ames, Price and Feinman1995; Burley and Knusel Reference Burley and Knusel1989). The accumulation of grave goods does not appear gendered (Ames Reference Ames, Price and Feinman1995; Beattie Reference Beattie1981; Burchell Reference Burchell2006), although gendered differences among higher-status adults may exist (McIlraith Reference McIlraith2012). Child burials are mostly devoid of grave goods, suggesting that achieved status prevailed during the Marpole period.
The practice of cranial modification reflects ascribed status and is theorized to evolve from an achieved status of labret adornment (Ames and Maschner Reference Ames and Maschner1999; Angelbeck and Grier Reference Angelbeck and Grier2012; Matson and Coupland Reference Matson and Coupland1994:214–215; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Elsewhere, we have more cautiously argued that both practices served as credibility-enhancing displays. Cranial modification is more spatially widespread and present among Wakashan-speaking peoples on the outer coast (Dingwall Reference Dingwall1931:164). We suggest that this reflects the establishment, negotiation, and entrenchment of hereditary elites (e.g., Angelbeck and Grier Reference Angelbeck and Grier2012; Coupland et al. Reference Coupland, Clark and Palmer2009; Schaepe Reference Schaepe2012). Houses provide the most complete context for understanding how adornment is engaged in the trade of prestige items, including nephrite, dentalia, copper, obsidian, and ground stone objects such as bowls (Angelbeck Reference Angelbeck2009; Burley Reference Burley1980; Clark Reference Clark2010; Grier Reference Grier, Matson, Coupland and Mackie2003, Reference Grier, Sobel, Trieu-Gahr and Ames2006; Matson Reference Matson and Matson2010; Matson and Coupland Reference Matson and Coupland1994; Morin Reference Morin2012; Springer et al. Reference Springer, Lepofsky and Blake2018). The prerogatives for production and consumption of prestige items may have also been restricted by hereditary lines.
Marked shifts in burial practices occurred toward the end of the Marpole period (1600–1000 BP; Thom Reference Thom1995) from subsurface interment to aboveground burial (Burley and Knusel Reference Burley and Knusel1989; Cybulski Reference Cybulski1993). The intentional manipulation of mortuary symbols, including the increased prevalence of cranial modification (Angelbeck and Grier Reference Angelbeck and Grier2012; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017), has been argued to reflect adoption by lower-status persons (Thom Reference Thom1995).
We argue that viewing labrets as a credibility-enhancing display (CRED) leaves the door open for broader interpretation in the context of House representation without predefining status as achieved or ascribed (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Our social network analysis intentionally eschews the periodization prefaced here to examine diachronic change. We hypothesize that the structure of labret social networks from 3500 to 2000 BP (Ames Reference Ames, Price and Feinman1995) reflects the development of House relationships through time in the region.
Methods
Social Network Analysis
Combining data visualization and mathematical methods, social network analysis (SNA) explores the structure and effects of social networks on its participants (Freeman Reference Freeman2004; Freeman et al. Reference Freeman, Borgatti and White1991). SNA examines and compares networks generated from archaeological data and can represent different spatial or temporal scales. Network characteristics include degree centrality, graph centralization, graph density and spatial length.
Degree centrality is a calculation of the total number of connections from a single node (Freeman Reference Freeman1977). Graph centralization is a measure of the tendency that any single node is more central in a given network. Graph density is an index of the number of connections in a network compared to the total number of possible connections. This study uses the spatial length of ties as a straight geodesic (shortest route) distance between archaeological sites.
Hypotheses
We examine diachronic change using arbitrary 500-year time periods, which not only avoid culture history frameworks but are sized appropriately for comparison given the available data. Our three general hypotheses address labrets as a practice reflecting Coast Salish House relationships.
(1) The Expansive Network is based on a high degree of regional interaction, expansion over time, and increasing network ties and expansive populations, reflecting a low variation in degree centrality. A high degree of communication between Houses would result in an intentionally or unintentionally shared symbolic language in the region (e.g., Rorabaugh and Shantry Reference Rorabaugh and Shantry2017).
(2) The Spatially Circumscribed Network assumes a lower degree of regional interaction and highly variable networks in degree centrality through time, with consistent centralized nodes reflecting shared local styles. This model assumes less communication and more territoriality, with labret styles used intentionally as local markers until replacement by cranial modification.
(3) Autonomized Networks are anarchic forms of social organization observed among Marpole hunters (Rorabaugh Reference Rorabaugh2019). The concept of anarchism is embedded in Western capitalist history. Therefore, we use autonomization to reflect the economic structures of Houses in traditional Coast Salish society. This model predicts a low variation in degree centrality, with a transition to weak network ties in later time periods. Here, labrets are used as a symbol of House autonomy, whereas cranial modification is introduced as a more inclusive identity in the region.
Sample
We examined dated labrets from 31 site components (Table 1; N = 96; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Compilation of this dataset is not straightforward due to factors including concerns related to Tribal sovereignty, Truth & Reconciliation, the Native American Graves Protection and Repatriation Act, and various curation study requirements. Prior to the LaSalle (Reference LaSalle2008) and Shantry (Reference Shantry2014) studies, no accounting of labrets in the Salish Sea had ever been compiled. Labret distribution is almost continuous on the east coast of Vancouver Island up to Comox, on the Gulf Islands, on the Fraser Delta, and in the San Juan Islands, and they are exceedingly rare in Puget Sound—likely a reflection of urbanism, survey intensity, and frequency of house excavations (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017:Appendix A). Salish Sea labret styles vary and have similarities to styles found farther north on the NWC and in the Arctic (Figure 3; LaSalle Reference LaSalle2008; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017; Shantry Reference Shantry2014).
Table 1. 500-Year Periods, Component Sample Size, Labret Sample Size, and Labret Stylistic Richness.
Figure 3. Labret styles (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017).
Organizing dated contexts by 500-year periods mitigates plateaus in regional radiocarbon curves from 2200 to 2460 BP (e.g., Hutchingson Reference Hutchingson2014; Wessen Reference Wessen2017) as well as uncertainties surrounding local marine reservoir corrections for marine shell dates. The study tested variable sizes and temporal breaks to confirm robustness. In addition, site components mitigate time-averaging effects given that many sites examined in this study have occupations spanning thousands of years.
We constructed three contemporaneous social networks (1500–2500 cal BP, 2500–3000 cal BP, 3000–2500 cal BP). The 1500 and 2000 bins were combined to increase sample sizes. We refrained from constructing a social network for the earliest component of the Pender Canal site as it had no paired site component.
Social Network Statistics
We used Brainerd-Robinson coefficients of stylistic classes to construct networks and a Monte Carlo simulation to assess variation resulting from sampling error. The Brainerd-Robinson similarity coefficient is independent of sample size and is a pairwise measure of the proportional similarity of nominal categories. A measure of 0 is no similarity, and 200 has identical proportions. Using Brainerd-Robinson coefficients to assess the similarity of archaeological assemblages has been the basis of many SNA studies (e.g., Mills et al. Reference Mills, Peeples, Randall Haas, Borck, Clark and Roberts2015). Although this study has a sample size substantially smaller than the 4.3 million examined by Mills and colleagues (Reference Mills, Peeples, Randall Haas, Borck, Clark and Roberts2015), the methods employed are equally suitable for small sample sizes.
We converted the results to a cost matrix in ArcGIS10.3.1, and the outdegree centrality statistic was used to assess the centrality of each network node based on the summed weights of all outbound edges of a node (Freeman Reference Freeman1977, Reference Freeman1979). To compare networks, we used the standardized form of the outdegree centrality statistic, which divided the summed outbound edge weights of a node by the sum total of all edge weights in a given network. For cross-comparison, we used Kruskal-Wallis tests and Spearman's R. Statistical analyses were performed in R 3.4.4.
Results
SNA resulted in three social networks (Figures 4–6). The oldest (3500–3000 cal BP) has strong assemblage similarity, and high Brainerd-Robinson coefficients (Supplemental Table 1) indicate similar proportions of labret styles (Figure 7). Later (1500–2500 BP) assemblages exhibit less stylistic similarity (Table 2).
Figure 4. 1500–2500 cal BP labret social network.
Figure 5. 2500–3000 cal BP labret social network.
Figure 6. 3000–3500 cal BP labret social network.
Figure 7. Pairwise component Brainerd-Robinson coefficients by 500 year cal BP period whiskerplot.
Table 2. Social Network Statistics.
Brainerd-Robinson coefficients also showed statistically significant bin differences at a 0.05 level (Kruskal-Wallis test N = 220, Χ2 = 7.983, df = 3, p = 0.046). Smaller sample sizes in later periods contributed to an increase in labret stylistic richness over time (e.g., Grayson Reference Grayson1981; Jones et al. Reference Jones, Grayson and Beck1983). However, the transition to lower pairwise component Brainerd-Robinson coefficients also appears to be a robust pattern among bins.
Brainerd-Robinson coefficients and geodesic distance using Spearman's R did not correlate at a statistically significant level for any constructed network (1500 BP N = 43, Spearman's R = 0.24, p = 0.121; 2500 BP N = 44, Spearman's R = 0.23, p = 0.134; 3000 BP N = 46, Spearman's R = 0.24, p = 0.290). This result is consistent with the low transportation cost of watercraft (Ames Reference Ames, Fitzhugh and Habu2002).
Graph density and graph centralization reflect the density, or connectivity, of social networks (Freeman Reference Freeman1979; Table 2). Network density values suggest that connectivity decreases through time. Degree centrality variance for the latest time period examined also increases, but this does not appear to be patterned by increased stylistic richness or sample size (Supplemental Tables 2 and 3).
Diachronically, labret social networks overlap multiple ethnohistoric linguistic areas, specifically Hul'qumi'num, Straits, and Lushootseed Coast Salish speakers. The increase of degree centrality variance is coupled with a decrease in labrets and a rise in cranial modification. Less connected sites with lower-degree centrality tend to be seasonal components such as a temporary fish-drying rack associated with labrets at Xwe'chi'eXen (45WH1).
The autonomization model—where labrets are a symbol of House autonomy and cranial modification follows as a more inclusive symbol of regional identity—fits the data best in terms of networks exhibiting low variation in degree centrality and a transition to weak network ties in later time periods. The Spatially Circumscribed model, suggesting that labrets were used as regional territorial markers with consistent centralized network nodes through time, does not accommodate the temporal trends. Similarly, the Expansive model—marked by an elaboration of increasing network ties with low variation in degree centrality reflecting a strong, shared, regional symbolic grammar—does not fit these data. The broader context, including intentional labret destruction and a transition to cranial modification, suggests a more nuanced process reflecting an interplay between the expansive and autonomization models through time.
Discussion
We argue that the materiality of Coast Salish Houses and their underlying relationships of status and access to resources are reflected in not only house structures but also labrets. The results of our SNA suggest network patterns consistent with labrets serving as a symbol of House autonomy. Labrets are understood to represent achieved status (Carlson Reference Carlson1987; Keddie Reference Keddie1981, Reference Keddie1989), and we have argued (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017) that labrets served as a symbol of emergent elite status reflecting the affinal ties of individuals. Labret social networks show strong ties throughout the Salish Sea by 3000 BP, suggesting rapid expansion over 2,000 years following the initial occurrence of labrets at the Pender Canal site around 5000 BP. Access to resources and cultural keystone places are crucial benefits of status.
Following arguments by Ames (Reference Ames, Price and Feinman1995:159), House networks can include access to noncorporeal resources such as songs, rituals, and dances. Our analysis affirms prior arguments that labrets reflected a symbolic language on the Salish Sea (LaSalle Reference LaSalle2008; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017; Shantry Reference Shantry2014). Diachronic network expansion suggests that labret wearing served as a material manifestation of existing social relationships. Although gender is outside the scope of this study, adult males, females, and subadults wore labrets, which may relate to stylistic variability (Cybulski Reference Cybulski and Matson2010:10, 17; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). Although the materiality of House relationships may be reflected in labrets, community and individuality are also defined and reflected at a more local scale.
It is also possible that as opposed to being a form of controlling “labret inflation,” given that cranial modification is a more expansive symbol, the intentional destruction of labrets was an act of disavowing the symbolic meanings of these heirloom items as the meanings of labrets shifted through time. Labrets and cranial modification are CREDs that may have signified access to both noncorporeal and corporeal resources as household estates expanded through time (Rorabaugh and Shantry Reference Rorabaugh and Shantry2017). The social construction and reproduction of access to noncorporeal resources includes moral teachings essential to Salish Sea ontologies.
We argue that labrets are not the manifestation of a pre–Coast Salish culture (e.g., Donald Reference Donald, Matson, Coupland and Mackie2003:307) but instead are historically situated within the lived experience of Coast Salish identity and cosmology. In essence, labret social networks may be the material manifestation of peer elite horizontal social networks. Dynamic horizontal relationships are not static over time: for example, Wakashan-speaking groups practiced cranial modification, but labrets are absent in the archaeological record.
One issue we could not address in this study is the potential use of other forms of personal adornment. One interpretation of what have been called Gulf Islands complex objects (Ames Reference Ames, Price and Feinman1995; Burley Reference Burley1980; Carlson Reference Carlson1960; Duff Reference Duff1956; LaSalle Reference LaSalle2008; Palmer Reference Palmer2012) is that they may have been part of compound labrets or pendants. However, other researchers have argued for deep time in wealth-based material inequalities based on evidence from beads in shíshálh territory (Coupland et al. Reference Coupland, Bilton, Clark, Jerome, Cylbulski, Frederick, Holland, Letham and Williams2016), and our study is consistent with their findings. Quartz crystal microblades may also be associated with labret piercings (Kannegard Reference Kannegaard2015). The only hafted quartz crystal microblade observed is from the Locarno Beach period Hoko River wet site, end hafted on a cedar stick handle (Croes Reference Croes2009). Makah elders suggested that end-hafted microblades were used in minor surgeries.
Conclusion
Through the use of SNA, we argued that labrets served as a marker of achieved House membership and mark the emergence of Houses and their horizontal relationships in the Salish Sea. A rare type of symbol, labrets demonstrated engagement in peer elite networks initially based on achieved status. This diachronic development incorporated a wider range of access to corporeal and noncorporeal resources. Instead of viewing labrets as a marker of the achieved status of proto-elites, we suggest that they represent a material correlate for House emergence. The transition toward cranial modification and the appearance of elite goods (Angelbeck Reference Angelbeck2009; Angelbeck and Grier Reference Angelbeck and Grier2012:559–563; Grier Reference Grier, Matson, Coupland and Mackie2003; Rorabaugh and Shantry Reference Rorabaugh and Shantry2017) is likely an elaboration of earlier peer elite social networks. To conclude, we note that viewing labrets as an archaeological manifestation of Houses rises to recent calls against periodization and to consider continuity through deep time (e.g., Grier Reference Grier2022).
Acknowledgments
We acknowledge the Coast Salish ancestors and their descendants, and we are grateful for the opportunity to work with their ancestral belongings. This study was inspired by the late Dr. Sarah K. Campbell, whose work at the Xwe'chi'eXen site motivated us to go down this path. We also thank Mark Collard and Matt Peeples for ideas and feedback on the social network analyses employed, and Albert Garcia-Piquer for translation assistance.
Funding Statement
This research received no specific grant funding from any funding agency or commercial or not-for-profit sectors.
Data Availability Statement
Data are available through tDAR id: 404209 (https://core.tdar.org/document/404209/credibility-enhancing-displays-and-the-changing-expression-of-coast-salish-social-commitments).
Competing Interests
The authors declare none.
Supplemental Material
For supplemental material accompanying this article, visit https://doi.org/10.1017/aaq.2023.98.
Supplemental Table 1. Brainerd-Robinson coefficient matrices by period.
Supplemental Table 2. Brainerd-Robinson Monte Carlo coefficient matrices by period.
Supplemental Table 3. Degree centrality measures by period.
