Archived version from NCDOCKS Institutional Repository http://libres.uncg.edu/ir/asu/

History, Monumentality, And Interaction In The Appalachian Summit Middle Woodland

By: Alice P. Wright

Abstract The Middle Woodland period in eastern North America witnessed a florescence of monumental architecture and material exchange linked to widespread networks of ritual interaction. Although these networks encompassed large geographic areas and persisted for several centuries, extant archaeological models have tended to characterize Middle Woodland interaction as an historically unitary process. Using new data from the Garden Creek site in North Carolina, I argue that these frameworks obscure important historical shifts in Middle Woodland interaction. Recent collections-based research, geophysical survey, targeted excavation, and 14C dating (including Bayesian modeling) of this site reveal two coeval diachronic changes: a shift from geometric earthwork construction to platform mound construction; and a shift from the production of special artifacts (mica, crystal quartz) to the consumption of exotic artifacts in association with platform mound ceremonialism. These data hint at important changes in interregional relationships between the Appalachian Summit, the Hopewellian Midwest, and the greater Southeast during the Middle Woodland period, and provide a springboard for considering how processes of culture contact contributed to precolumbian cultural change.

Wright, A. (2014). History, Monumentality, and Interaction in the Appalachian Summit Middle Woodland. American Antiquity, 79(2), 277-294. doi:10.7183/0002-7316.79.2.277. Publisher version of record available at: https://www.cambridge.org/core/journals/american-antiquity/article/history-monumentality-and- interaction-in-the-appalachian-summit-middle-woodland/8A6ABDE3030E29F367568A0F24692AD9 HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND

Alice P. Wright

The Middle Woodland period in eastern North America witnessed a florescence of monumental architecture and material exchange linked to widespread networks of ritual interaction. Although these networks encompassed large geographic areas and persisted for several centuries, extant archaeological models have tended to characterize Middle Woodland interaction as an historically unitary process. Using new data from the Garden Creek site in North Carolina, I argue that these frameworks obscure important historical shifts in Middle Woodland interaction. Recent collections-based research, geo­ physical survey, targeted excavation, and l4C dating (including Bayesian modeling) of this site reveal two coeval diachronic changes: a shift from geometric earthwork construction to platform mound construction; and a shift from the production of special artifacts (mica, crystal quartz) to the consumption of exotic artifacts in association with platform mound cere­ monialism. These data hint at important changes in interregional relationships between the Appalachian Summit, the Hopewellian Midwest, and the greater Southeast during the Middle Woodland period, and provide a springboard for con­ sidering how processes of culture contact contributed to precolumbian cultural change.

Durante el periodo Middle Woodland, la region este de America del norte experimento unaflorescencia de arquitectura monu­ mental e intercambio que fueron conectados a redes extensivos de interaction ritual. Aunque estas redes cruzaron regiones geogrdficas grandes y persistieron por varios siglos, modelos arqueologicos existentes han caracterizado interacciones Middle Woodland como un proceso historico y unitario. Usando information nueva coleccionada del sitio de Garden Creek en North Carolina, argumento que estos modelos ocultan cambios importantes e historicos en las interacciones de la epoca Middle Woodland. Investigaciones recientes que utilizan colecciones, tecnicas [prospeccion] geofisicas, excavation, y datos radio- carbonicos del sitio de Garden Creek revelan dos cambios contempordneos y diacronicos: el cambio de la construccidn de monticulos [de tierra] geometricos a la construccidn de monticulos plataformas; y un cambio de la production de artefactos preciosos (mica, cuarzo) al consumo de artefactos exoticos relacionados con ceremonialismo de monticulos plataformas. Esta information indica que existieron cambios importantes en las relaciones interregionales entre el Sumo Appalachian, el medio- oeste Hopewellian, y la mayoria del sureste durante la epoca Middle Woodland, y ofrezcan una punto de partida para la con­ sideration de procesos de contacto entre culturas y su contribution a cambios en tiempos precolombinos.

uring the Middle Woodland period, in- 1979), or, more recently, as a variety of regionally digenous communities across eastern specific archaeological signatures, such as DNorth America contributed to a flores- Kolomoki pattern platform mounds, Copenabur- cence of monumental architecture, interregional ial practices, and Swift Creek ceramic exchange exchange, and cosmologically meaningful mate- networks. This taxonomic shift is the result of rial culture that has long captured the attention several decades of Middle Woodland research in and imagination of archaeologists. In the Mid- the Southeast that have demonstrated consider- west, Middle Woodland mounds, earthworks, and able geographic variability among ceremonial exotic and iconographically distinctive ritual as- practices that were once attributed to Hopewellian semblages are called Hopewell. In the Southeast, influences. While the broad synchronicity of Mid- contemporaneous traditions have been described die Woodland developments in the Midwest and as local versions of the wider Hopewellian phe- Southeast led to its early depiction "as a period nomenon (Brose and Greber 1979; Seeman of panregional communality, we have come to

Alice P. Wright • Department of Anthropology, Appalachian State University, ASU Box 32016, Boone, NC 28608-2016 (wrightap2 @ appstate .edu)

American Antiquity 79(2), 2014, pp. 277-294 Copyright © 2014 by the Society for American Archaeology

277 278 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014 realize that many diverse cultures were actually ment with Hopewell may have varied from hesi­ present, onto which a thin veneer of Hopewellian tant acceptance or enthusiastic adoption, from exchange, iconography, and ritual was over­ the mutual creation of a hybrid set of ritual or lain" (Anderson and Mainfort 2002:9). exchange practices, to polite rejection and even That said, such formulations rarely address pronounced resistance. what sorts of on-the-ground mechanisms con­ An historical processual examination of the tributed to intercultural dynamics during the Mid­ Hopewellian Southeast first requires the empirical dle Woodland period; put another way, they do identification of intercultural interaction and, by not critically examine the ways by which certain extension, Middle Woodland culture-making. To dimensions of Hopewell came to "overlay" di­ that end, this article presents new information verse southeastern traditions. To tackle these is­ from the Garden Creek site in western North Car­ sues, archaeologists must critically evaluate the olina's Appalachian Summit, and offers a di- "Hopewell Interaction Sphere" concept (Caldwell achronic narrative of the emergence of different 1964), which not only pre-supposes a unitary vi­ forms of monumental architecture and the pro­ sion of interregional Hopewell, but also tends to duction and use of artifacts involved in Middle characterize Hopewell interaction as something Woodland interaction networks. Although evi­ that happens to people, rather than a set of dence from Garden Creek has previously been processes that people actively created and ma­ cited to support various models for Middle Wood­ nipulated. An historical processual approach that land interaction, new data produced through geo­ emphasizes how convergences among diverse physical survey, targeted excavation, and radio­ peoples relate to cultural change (sensu Alt 2006) metric dating indicate that the interpersonal nature is especially well suited to unpacking these issues. and geographic orientation of interregional con­ From this perspective, nections underwent substantial transformations in the first several centuries A.D. In short, a history is defined as the ongoing process of chronologically dynamic record of monumental- making culture through social interactions. In ity, craft production, and exotic material culture this sense, history is never merely a chronicle at Garden Creek allows for the provisional infer­ or narrative of human experience, but rather a ence of several processes of intercultural contact series of processes (e.g.,diaspora,coalescence, that shaped and were shaped by Middle Woodland and ethnogenesis) which, like all processes, people in the Appalachian Summit. Rather than are subject to comparison and generalization characterizing these diverse material remains as [Sassaman 2010:5; emphasis in original; see an "overlay" of interregional influences, this his­ also Cobb 2005; Pauketat 2001]. torically minded account underscores the potential In undertaking such a project, it is critical to multiplicity of interaction experiences in the Mid­ keep in mind that, in addition to varying across dle Woodland Southeast, and calls attention to space, Hopewellian interactions in the Southeast the agency of local constituencies in the negotia­ likely shifted through time. If, as suggested by tion of intercultural connections. Seeman (2004:58-59), midwestern Hopewell per­ To frame the historical processual approach sisted across 13 generations, the chances that the adopted in this article, I begin with a discussion many-times-great grandchildren of the earliest of extant perspectives on Middle Woodland in­ Hopewellian groups were interacting with exactly teraction in the Southeast. Although several dif­ the same people, involved with exactly the same ferent models have emerged over the years, this interregional networks, or executing ceremonies brief history shows that they have tended to be according to the exact same script as their prede­ applied to the Middle Woodland archaeological cessors are rather slim. Similarly, the ways that record with a fairly broad brush, with little to no Southeastern peoples engaged or did not engage consideration of temporal variability. I propose with Hopewellian and other nonlocal objects and that instead of deploying these static frameworks ideas could have changed dramatically through for interregional interaction during the Middle time. Depending on particular geographic and Woodland period, we generate detailed diachronic historical contingencies, Southeastern involve­ histories of particular places to see if, when, and Wright] HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 279 how particular scenarios played out. Focusing on traits was linked to widely shared religious beliefs architectural, artifactual, and temporal data from and ceremonial practices (Caldwell 1964) that in the site's Middle Woodland monuments, I attempt turn were viewed as a realm of activities entirely to carry out precisely this sort of project at Garden separate from everyday "domestic" life. Creek. The resulting narrative, though specific By the 1970s, archaeologists were striving to to this site and its historical contingencies, sug­ delineate a mechanism for widespread Hopewell gests that diverse forms of interregional culture material culture that was grounded in economic contact are important historical processes to crit­ rationale. Stuart Struever and Gail Houart (1972) ically and creatively investigate across the Middle argued that the Hopewell Interaction Sphere was Woodland Southeast and beyond. an exchange system, comprised of specialized production and ceremonial facilities in Ohio and Archaeologies of Southeastern Hopewell subsidiary transaction centers associated with far- flung raw material sources. Slightly later, David According to James Griffin, one of the forefathers Braun (1986) proposed a peer polity framework, of Eastern Woodlands archaeology, the seminal which postulated that Middle Woodland commu­ 1978 conference on Hopewell archaeology in nity members would seek out exotic objects in Chillicothe, Ohio amounted to a "revolt of the the context of competitive display. Mark Seeman South" (Pacheco 1996:vi). Contrasting with most (1995) revised and elaborated upon these ideas, of the previous century of Hopewell scholarship, arguing that Hopewellian material culture and the Chillicothe Conference expanded the discus­ iconography was an interregionally intelligible sion of exotic artifacts, ceremonial practices, and form of communication that could be used by in­ monuments dating to the Middle Woodland pe­ cipient leaders to acquire authority and prestige. riod beyond the Ohio Valley and into the greater Today, a combination of Seeman's ideas and Midwest and Southeast. However, this interre­ Mary Helms's (1988) perspectives on exotic ma­ gional free-for-all was not unproblematic. Bennie terial culture defines the predominant view of Keel, who chaired the presentation of papers on southeastern Hopewell, which emphasizes the southeastern states, opened the follow-up discus­ potential utility of nonlocal artifacts in contexts sion by noting "that in a strict sense, there was of aggrandizement. As David Anderson no Hopewell in the Southeast" (Brose and Greber (1998:287) explained, 1979:209). Yet if that was the case, why was that by being major players in the Hopewellian region addressed at the conference at all? world, the principals at these [southeastern] The short answer is that for many decades, centers could have been perceived as having archaeological research at Middle Woodland sites esoteric knowledge...and this, plus their con­ in the Southeast had yielded artifacts and exotic trol over desirable wealth items, may have raw materials that were (and, in most cases, still inspired people to their service over wide areas are) commonly associated with Hopewellian cer­ and at the same time led to their sanctification. emonialism. Though exact assemblages varied from site to site, they included Hopewellian ce­ One correlate of this framework has been the ramics and figurines; copper panpipes, earspools, identification of gateway centers for interregional breastplates, and pins; cut and uncut sheet mica; exchange along likely transportation corridors and various other diagnostic stone, shell, and (Keith 2013; Ruby and Shriner 2005). bone objects (see chapters in Brose and Greber While different in many ways, these diffusion­ 1979). The earliest interpretations of these as­ ist, economic, and sociopolitical scenarios tend semblages were decidedly diffusionist. The most to presuppose an historical unity to interregional profound interregional influences were thought Hopewell. By this, I mean that each of these to have been directed toward the Southeast from frameworks is at least implicitly presented as the the Hopewell core, though it was also suggested way that Hopewellian interaction transpired at a that a small number of traits may have diffused particular locality throughout the Middle Wood­ from south to north (e.g., Greenman 1938; Sears land period. For example, in the Appalachian 1962; Shetrone 2004 [1930]). This diffusion of Summit, evidence for Hopewell interaction has 280 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014 been attributed to trade with Ohio Hopewell trav­ ley et al. 2014) and remains of non-mound activ­ elers who were in the area to procure mica (Chap­ ities (31Hw8) (Figure 1). This portion of the site man and Keel 1979), or to aggregations among has been investigated intermittently by antiquar­ local communities and representatives from the ians and archaeologists since the late nineteenth Hopewell core at particular ceremonial encamp­ century. ments (Walthall 1985), or to efforts by southeast­ Two field projects are relevant to the present ern leaders to integrate their communities through study. The first, directed by Bennie Keel, involved feasting and exotic gift giving (Knight 2001). The the systematic documentation of the site as it ex­ possibility that multiple mechanisms of interre­ isted in the mid-1960s and the total excavation gional interaction may have taken place at a single of Garden Creek Mound No. 2, which was threat­ site—either simultaneously or sequentially—is ened by encroaching residential development rarely addressed. (Keel 1976). Keel's efforts revealed multiple The following case study aims to explore pre­ stages of mound construction, an underlying mid­ cisely this issue by establishing detailed di- den, and features and artifacts indicative of com­ achronic narrative of interregional interactions munal ceremonialism and extralocal exchange. experienced at the Garden Creek site. Specifically, Specifically, nonlocal pottery, ceramic figurines, I propose that different forms of monumental ar­ and certain exotic artifacts were cited as evidence chitecture and diverse artifact assemblages rep­ that the people of Garden Creek participated to resent critical shifts in the ways in which interre­ some degree in extralocal Hopewellian interaction gional interactions were brought to bear on the networks (Chapman and Keel 1979; Keel 1976; creation of culture and unfolding of history at Walthall 1985). Based on his seminal ceramic Garden Creek. In this regard, the present study chronology for the Appalachian Summit and a brings the discourse on southeastern Hopewell single radiocarbon date from the top of the mound full circle, being grounded in an important insight (1145 ± 85 B.P. [Eastman 1994:66]), Keel offered early on by Joseph Caldwell [1964:143]: (1976:86) assigned Mound No. 2 to the late Mid­ dle Woodland Connestee phase. He also sug­ The explanation then for a positive correlation gested that the scatter of artifacts that surrounded between interaction and innovation is that the mound represented a contemporaneous village when different cultural traditions are brought occupation, although his efforts (Keel 1976:71) together there becomes available to each a new to characterize this area through surface survey supply of diverse forms upon which new were hampered by houses and established lawns. rearrangements of forms—innovations and The second field project of interest, which I inventions—can be built. directed in 2011-2012, targeted this off-mound component. Nearly 50 years after Keel's field- Middle Woodland Investigations work, the neighborhood had expanded consider­ at the Garden Creek Site ably, eliminating the possibility of surface-level archaeological survey and mapping, and restrict­ The Garden Creek site is located on a small (less ing opportunities for extensive excavation. How­ than 25 ha, approximately) terrace and floodplain ever, by combining shallow archaeological geo­ along the Pigeon River in the Appalachian Sum­ physics with highly targeted excavations, the mit, where the Middle Woodland period is sub­ Garden Creek Archaeological Project (GCAP) divided into the Pigeon phase (300 cal B.C -A.D. managed to maximize data recovery across almost 200) and the Connestee phase (200-600 cal A.D. the entirety of the western half of the landform ), largely according to differences in ceramic as­ with minimal disturbance to the site and its mod­ semblages. The site's most extensive occupations ern day occupants. resulted in two more-or-less spatially discrete To briefly summarize our field methods, we components dating to the Mississippian and Mid­ employed several complementary techniques to dle Woodland periods, respectively. The latter in­ measure subsurface geophysical variability and cludes two or three mounds (31Hw2, 31Hw3, to identify anomalies indicative of past human and the newly discovered Mound No. 4; see Hors- activity (for more details, see Horsley et al. 2014). Wright] HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 281 "7~"

•V Enclosure No. 2

. Mound No. 4 Mound No. 2 Edge of (deflated) (excavated) floodplain Enclosure No. 1

Presumed area of Middle Woodland occupation (31Hw8)

Mound No. 3 (excavated)

A 50 100 150 m

Figure 1. Schematic map of the Middle Woodland component Garden Creek site. Inset shows location of Garden Creek in the greater Appalachian Summit.

Using a Bartington single-axis magnetic field Creek Enclosures No. 1 and No. 2, they resemble gradiometer, we surveyed approximately 7 ha of small geometric enclosures common in midwest- the Middle Woodland portion of the site, includ­ ern Adena and Hopewell contexts, particularly ing large portions of the landform that were not so-called "squircles" (Anderson 2013: 252-253; subsumed by the proposed boundaries of Keel's see also Burks and Cook 2011; Clay 1987,1998). Middle Woodland village. Nearly 1 ha of this From inside edge to inside edge of the linear survey area was also mapped using ground pen­ magnetic signatures, these features measure 18 etrating radar (GPR). These efforts identified nu­ m southwest-to-northeast and 16 m northwest- merous negative features (e.g., pits, pit hearths, to-southeast; the outline of each is broken by an large postholes). Based on the spatial distribution opening or "gateway" that measures about 4 m of these anomalies and the results of limited wide. In plan view the anomalies are slightly off­ ground truthing, we have provisionally attributed set, but their gateways generally face in the di­ much of the off-mound remains to intermittent rection of the opposite enclosure. Moreover, the occupation, perhaps associated with the seasonal enclosures share the same orientation, approxi­ aggregation of Middle Woodland groups whose mately 20 degrees west of magnetic north, sug­ subsistence regimes included hunting, gathering, gesting a purposeful layout that perhaps refer­ and gardening (Ward and Davis 1999:154). ences some presently undetermined geographic Two linear features, each in the shape of a or celestial alignment. As discussed below, these rectangle with rounded corners, stand out among features were further elucidated through targeted these anomalies (Figures 2, 3). Labeled Garden excavation of Enclosure No. 1. 282 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014

Y*

25 50 100 Horsley Archaeological Prospection, LLC

Figure 2. Extract of the Garden Creek magnetometer data revealing anomalies representing Enclosures No. 1 and No. 2. Other archaeological features are also suggested, as well as disturbances associated with modern features (e.g., iron metal, septic systems), and agricultural activities (plow scars). Results are plotted from -12 nT (white) to +12 nT (black). Processing was limited to clipping to within 30 nT, application of zero mean traverse, and interpolation to a resolution of .25 m x .125 m.

In short, old and new field work at Garden cidating the construction history of the site and Creek has generated a record of monumentality revealing significant architectural changes that and material culture that points to similarities may be related to interregional interactions and/or and, by extension, interaction with Hopewellian local developments (Beck et al. 2007). In partic­ peoples in the Midwest. By examining these mon­ ular, architectural information derived from the uments and their associated assemblages in more stratigraphy of these monuments can be compared detail, I argue that it is possible to delineate not to the architecture of broadly contemporaneous simply participation in so-called Middle Wood­ monuments in other regions as a means of iden­ land interaction spheres, but more specifically tifying significant consistencies and divergences the intensity, directionality, and temporality of in monumental practice at different moments local communities' relationships with foreign peo­ throughout the Middle Woodland period. ple and places. Mound No. 2

Stratigraphic Histories of When Keel encountered Mound No. 2 in 1965, it Garden Creek's Monuments was already partially destroyed by a bulldozer. Fortunately, his detailed excavation of the remain­ As important elements of Garden Creek's built ing mound in reverse stratigraphic order allows environment, Mound No. 2 and Enclosures No. for the formulation of a fairly detailed life history 1 and No. 2 have considerable potential for elu­ of the monument. Keel assigned all mound con- Wright] H.STORY, MONUMENTALLY, AND INTERACTS IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 283

gas^^sssesssss structibn stages and associated features to the late ture 1. The posts from Structure 1 were removed Middle Woodland Connestee phase, based largely before the resulting holes were filled with a clean, on pottery associations. However, because the white sand; elsewhere, I (Wright 2013) have sug­ chronological relationships among Woodland ce­ gested that this pattern represents the purposeful ramic series in the Appalachian Summit are poorly dismantling of this building, and perhaps its ritual understood (Ward and Davis 1999), stratigraphic closure. Pre-mound activity continued in this lo­ relationships and new radiocarbon dates (see be­ cation, producing a dark clayey midden that low)—not ceramic affinities—are considered the ranged between about 9 and 21 cm thick (Keel most reliable sources of chronological information 1976:77), in which Keel encountered cobble at Mound No. 2. hearths, basins, and numerous postholes. The The earliest evidence of activity associated thickness of the midden suggests that the activities with Mound No. 2 consists of several features that produced it were either intense or prolonged. that were dug into the subsoil and were subse­ The possibility that it represents the remains of a quently capped by a midden and, later, by the domestic occupation cannot be ruled out (Wright mound itself. Keel also identified and mapped 2013), but certain lines of evidence—such as a nearly 1,300 postholes at the level of the subsoil, mica-lined pit at the base of the pre-mound mid­ including a rectangular outline designated Struc­ den and the possible ritual closure of Structure 284 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014

1 — suggest that ceremonially prescribed activities "centered on world renewal and feasting, empha­ may have occurred there as well. sizing community integration" (see also Hall The first episode of mound construction di­ 1997). While Knight (2001:327; see also Lindauer rectly overlay this midden. Measuring about 12 and Blitz 1997) suggests that these activities op­ by 18 m in horizontal extent and about a half-me­ erated at a relatively local, intraregional scale to ter tall, this low platform was a locus of diverse facilitate and secure intervillage alliances, the activities, as indicated by the presence of many widespread adoption of such architectural prac­ different types of features—shallow basins, deeper tices across what are now several states suggests pits, hearths with and without cobbles, and a pipe- that some form of interregional interaction oc­ stone cache—and numerous postholes. Some of curred among those communities erecting these posts appear to have formed a rectangular Kolomoki pattern mounds. outline around a burned surface (Feature 41) that Keel identified as the floor of a mound-top struc­ Enclosure No. 1 ture. As a result of plowing and bulldozing, the In a 5-x-3-m horizontal excavation block and two absolute summit of the second stage of Middle 1-m-wide profile trenches, the Enclosure No. 1 Woodland construction was not intact, but con­ anomaly was revealed to be a steep-sided ditch siderable information was gleaned from the sur­ with nearly flat walls and a flat bottom that ex­ face represented by the interface of the plowzone tended 1 to 1.2 m below the ground surface.1 In and the secondary mound fill. Like the summit of profile, the shape of the ditch was generally trape­ the primary mound, this surface was characterized zoidal, measuring 1.55 m wide at the top (below by diverse features, though no structures were the plowzone) and 80 cm wide at its base. The identified among the postholes at this level. entirety of the sub-plowzone ditch was originally This record of platform mound architecture excavated into very dense sandy clay subsoil, but and activity at Garden Creek is similar to a num­ it was eventually filled in with three distinct zones ber of other Middle Woodland platform mounds of sediment. No macroscopic evidence of soil in the Southeast. Concentrated in the deeper formation was noted at any of the interfaces be­ South, the so-called Kolomoki pattern of platform tween zones of fill, suggesting there were no ma­ mound architecture is defined by a number of jor gaps in episodes of infilling. characteristics (Knight 1990:170-171): irregular Interestingly, the life history of Enclosure No. scatters of postholes (attributable to scaffolding 1 does not appear to have ended with the infilling behavior; see Knight 2001:319) and pits; a lack of the ditch. Once it was entirely filled in, the of clear summit structures; extraordinarily large outline of the enclosure continued to be marked postholes, some with insertion and/or extraction by a series of posts that ranged from 12 to 23 cm ramps; burned areas and hearths on mound sum­ in diameter and aligned roughly with the middle mits; multi-stage construction using multicolored of the original ditch. Eventually, these posts were fills; and the presence of exotic artifacts and spe­ removed and the resulting postholes were filled cial ceramics. Garden Creek Mound No. 2 ex­ with tightly packed river cobbles and, in some hibits all of these traits, with the possible excep­ cases, a few fragments of pottery, charcoal, and tion of the lack of clear summit structures. Vernon mica. These features were encountered during ex­ J. Knight (2001:221; see also Jefferies 1994) in­ cavation as discrete columns of rock, beginning terprets these patterns as the remains of "inter­ at the base of the plowzone and extending though mittent, repetitive activities involving manipula­ the top, middle, and sometimes bottom zones of tion of exotic artifacts, caching of goods in small ditch fill (Figure 4). In total, 6 rock-filled postholes pits, food preparation and consumption, frequent were identified across a 5-meter-long exposure of scaffolding of objects unknown, and monumental the ditch, spaced at 80-cm intervals; additional display of poles." Because these activities were rock-filled postholes were identified in separate associated with earthen platforms that were peri­ 1-m profile trenches, suggesting that this align­ odically resurfaced with new construction ment continued around the entire enclosure. episodes, Knight (2001:328) further argues that In many regards, the architecture of Enclosure these contexts represent ceremonial practices No. 1 is remarkably similar to a category of mon- Wright] HISTORY, MONUMENTALLY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 285

Ertft •.—ft *ft «t: :^,5? jJ»%-;C*: :

•WfJl Bs&1

Figure 4. Rock-filled postholes at the base of the top zone of ditch fill in Enclosure No. 1.

uments common in the Ohio Valley called small Using published and unpublished literature geometric enclosures. Sometimes referred to as from the Middle Ohio Valley, Richard Jefferies, the Mt. Horeb tradition (Byers 2004; DeBoer George Milner, and Edward Henry (2013:103- 1997) or Adena sacred circles (Webb and Snow 104) recently identified 259 of these small geo­ 1945), these earthworks are much smaller than metric enclosures from which they extrapolated massive ditches and embankments that serve to the general characteristics of this type of monu­ demarcate monumental sites around the Scioto- 3 ment. Ranging from .01 to 1.35 ha in the amount Paint Creek confluence, and as a result, they have of space enclosed, most of the earthworks in likely suffered far greater effects from erosion their sample were circular, though arcs, ovals, and plowing. Nevertheless, they are noted with and rectangles were also present. Openings or some frequency on early maps of the greater Ohio gateways in the earthworks were also common— Valley (e.g., Squier and Davis 1998 [1848]), both usually one opening per enclosure, most often in association with and away from larger earth­ facing east. In a similarly comparative study work complexes. Though historically associated (Wright 2013), I used Byers's (2004) typology with the Early Woodland period and the Adena complex,2 "It is now recognized that the Mt. to classify 88 enclosures according to their ver­ Horeb tradition not only largely defines the later tical characteristics: 38 consisted of an embank­ Early Woodland period but also extends into the ment with no ditch (K-profile); 49 included both Middle Woodland in limited parts of this region and embankment and a ditch (47 with an interior [the Central Ohio Valley], terminating around AD ditch/SR-profile, 2 with ditches inside and out­ 250" (Byers 2004:28-29). side the embankment); and one had a ditch but no embankment. Anecdotally, the few enclosures 286 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014 that have been excavated are characterized by in the region, and none have been recovered from particular stratigraphic patterns, including the Garden Creek. However, the Summit has long use of sediments from specific locations and ex­ been assumed to have been the raw material hibiting specific properties (e.g., color). The care­ source for such objects recovered at Ohio ful selection of building materials has been in­ Hopewell ceremonial sites (e.g., Griffin 1967; terpreted not only as evidence of architectural Holmes 1919). Given these connections, it is pos­ engineering (sensu Sherwood and Kidder 2011), sible that the mica and crystal quartz debris re­ but also in reference both to symbolic design covered from the ditch represents the transfor­ principles apparent in other Hopewell structures mation of local raw materials into partially or and objects (Greber 2006:90) and to "the wholly finished objects for some sort of extralocal Hopewell people's efforts to manage the spirit transaction—in short, the remains of Hopewellian world" (Lynott 2004:6). craft production. This interpretation is strength­ All that said, Garden Creek Enclosure No. 1 ened by the unique and fairly direct association is distinctive from its Ohio Valley counterparts of these materials with Enclosure No. 1 (for de­ (at least those that have been excavated) in at tails, see Wright and Loveland 2013). The pres­ least one important way—it was filled in and ef­ ence of several large but fragile sheets of mica fectively erased from the landscape probably not (8-11 cm on a side) indicates that these scraps long after it was initially constructed. Such ac­ were directly deposited into the ditch as it was tivity has not been observed at any small geo­ being filled with earth and other materials, sug­ metric enclosure in the Ohio Valley, though it gesting in turn that cut mica craft production was seems likely that many of these enclosures remain somehow associated, spatially and/or symboli­ to be discovered through geophysical survey cally, with the enclosure itself. (Burks and Cook 2011); their subsequent study Mound No. 2's assemblage of diagnostically may affect this apparent pattern. At present, how­ Hopewellian material culture is quite different. ever, the ceremonial and social activities that led This relatively small assemblage includes blades to the erasure of the Garden Creek enclosures made of Ohio Flint Ridge chalcedony, ceramic are events that are unique to this locality and in­ figurines, copper beads, sheets, and pins, and cut dicative of specific historical events. animal mandibles. In contrast to the mica and crystal quartz from the enclosure, the provenance Artifact Assemblages in of the blades is not local; it is also possible that Monumental Contexts the copper artifacts originated from distant re­ gions (Ehrhardt 2009). Importantly, these artifacts On the basis of the earthen architecture itself, in­ are finished objects, and there is no on-site evi­ terregional connections to the south (the dence for their production. In this regard, the Kolomoki pattern) and the north (Adena- Hopewellian artifacts associated with Mound No. Hopewell small geometric enclosures) can al­ 2 appear to represent the targeted consumption ready be identified at Garden Creek. The nature of ceremonial material culture, rather than the of these interactions can be clarified by consid­ production of such items as seen at Enclosure ering certain artifact assemblages from these con­ No. 1. Insofar as these objects may have derived texts in greater detail. value from nonlocal origins or associations, it Both monuments yielded artifacts that tradi­ seems plausible that they were used in mound tionally have been associated with Hopewellian related activities to reference or cite foreign peo­ ceremonialism. Compared to other Middle Wood­ ple, places, or practices, and as such, that they land deposits at Garden Creek, the fill of Enclo­ might have been sources of ritual power or au­ sure No. 1 was remarkable for its high density of thority for their possessors (Helms 1988). cut sheet mica fragments and crystal quartz deb- Whereas these Hopewellian assemblages sug­ itage (Wright and Loveland 2013). Although both gest that certain dimensions of sociality at Garden raw materials are locally available in the Ap­ Creek related to long-distance connections, the palachian Summit, finished artifacts such as mica ceramic artifacts from both monuments highlight cutouts or stylized crystal quartz bifaces are rare the active contributions of a local constituency. Wright] HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 287

The vast majority of ceramics from Enclosure at present, be most parsimoniously attributed to No. 1 and Mound No. 2 can be assigned to local incidental long distance or down-the-line ex­ ceramic series, characterized by sand and crushed changes with people not only in the Midwest, quartz tempers and a variety of surface treatments but also further south. To the extent that these (Keel 1976). Even sherds from the mound that distinctive types of material interaction are sepa­ were initially thought to have originated in Ohio rated in space (i.e., between monuments) and on account of their rocker-stamped surface treat­ time (discussed below), they stand to complement ment (Keel 1976:197) are now known to have documented historical shifts in monumentality been made locally (see petrographic analysis by and, in turn, the directionality and nature of Gar­ Stoltman 1999). Only .21 percent of sherds from den Creek's interregional connections. the ditch (n = 475) and .25 percent of sherds re­ cently reanalyzed from the mound (n = 7,420) Dating the Monuments exhibited nonlocal limestone tempers; these arti­ facts probably originated in eastern Tennessee. Having described the architectural and artifactual Meanwhile, in his analysis of nearly 12,000 characteristics of the Garden Creek monuments, sherds from Mound No. 2, Keel (1976) identified it is now necessary to situate these contexts in five "Georgia type" sherds (Early Swift Creek time to create a robust historical narrative of the Complicated Stamped and Napier Stamped) and site. Fortunately, Keel's and my own fieldwork one "Florida type" sherd (Crystal River Plain) in yielded numerous charcoal samples that were ra­ addition to 92 limestone-tempered sherds from diocarbon dated as part of recent investigations Tennessee. In short, while it is demonstrably true at the site. In total, eight new dates were run on that pots are not people, these macroscopic (and wood charcoal from eight different features in thus preliminary) data indicate that the ceramic Mound No. 2. Based on my reanalysis of field assemblage at Garden Creek is overwhelmingly maps, profiles, and sample elevations, some of attributable to production and, presumably, to use these features were assigned to different strati- by local people. graphic contexts (i.e., pre-mound, primary Arguably, despite their extralocal associations, mound, secondary mound) than they were in the Hopewellian assemblages at Garden Creek Keel's (1976) original formulation. Samples were can also be interpreted as the results of activities selected from different stratigraphic contexts carried out by local people. The materials and throughout the mound to evaluate the duration tools necessary for mica and crystal quartz craft­ and tempo of the entire life history of the monu­ ing were locally available, and mica in particular ment, as well as from several features on the sum­ requires little specialized skills for its manufacture mit of the primary mound, to better date activities into cutouts (Spielmann 2009:184). In fact, the occurring across the most intact monumental sur­ wide distribution of mica fragments at Middle face at the site. The individually calibrated dates Woodland sites across the Eastern Woodlands has from these samples are presented in the fourth been cited as evidence that mica crafting was not column of Table 1. They are in approximate strati- limited to ritual specialists and could be carried graphic order from latest (top) to earliest (bottom), out away from Ohio Hopewell ceremonial with the exception of Sample No. GC1966.07, precincts (Spielmann 2009:181-182). We can which was recovered from either midden or also infer local deployment of nonlocal traditions mound slump and could not be confidently as­ from the Hopewell style potsherds made with signed stratigraphic order. Because no strati- Appalachian Summit paste and tempers. While graphic relationships could be inferred from the the introduction of these distinctive surface treat­ features on the primary mound, they are attributed ments to the Garden Creek vicinity plausibly in­ here to a single phase of mound summit use. volved some material or interpersonal exchange, Even in this relatively raw form, the overlap in the vessels themselves were most likely manu­ these dates is considerable, providing a strong in­ factured by local potters, using local techniques. dicator that the construction of Mound No. 2 oc­ In contrast, the few patently Hopewellian artifacts curred over about two hundred years at the Pigeon- and nonlocal ceramics found in the mound can, Connestee phase transition. These results can be 288 AMERICAN ANTIQUITY [Vol.79, No. 2, 2014

Table 1. Radiocarbon Age Estimates from Mound No. 2, Garden Creek, Calibrated Using OxCal Version 4.2.2 and Int.Cal 9 Calibration Curve (Reimer et al. 2009).

Individually Modeled calibrated calibrated Sample No. Feature No., Stratigraphic 14C years 2-sigma range 2-sigma range (Lab No.) Description Context B.P. (cal A.D.) (cal A.D.) GC1966.04 Fea. 40, basin secondary mound 1765 ± 38 137-382 262-387 (AA100830) GC1966.01 Fea. 5, cobble hearth primary mound 1771 ±38 240^10 294-341 (AA100827) GC1966.02 Fea. 18, cobble hearth primary mound 1760 ± 38 139-385 211-334 (AA100828) GC1966.03 Fea. 32, simple hearth primary mound 1799 ± 38 126-337 184-331 (AA100829) GC1966.05 Fea. 41, burned floor primary mound 1700 ±43 240-425 235-360 (AA100831) GC1966.06 Fea. 43, cobble hearth primary mound 1638 ±38 265-537 249-329 (AA100832) GC1966.08 Fea. 45, mica-lined pit pre-mound midden 1839 ±38 76-312 128-317 (AA100834) GC1966.07 Fea. 44, cobble hearth context unclear 1765 ± 37 137-382 not modeled (AA100833) further refined through Bayesian modeling, which the dates. The integrity of this modeled precision incorporates prior knowledge—in this case, is supported by the fact that six of the seven mod­ stratigraphy—in the statistical calculation of the eled dates had good statistical agreement with absolute dates (Bayliss 2009; Bronk Ramsey 2009). the a priori parameters (A-values > 60 percent). For the present model, a three-stage sequence is The only exception was GC 1966.06, with a low postulated, beginning with the pre-mound midden, A-value of 25.4 percent. Both Keel (in the field) followed by the primary mound, and ending with and I (on the basis of maps, notes, etc.) agree the secondary mound. All dates from the primary that stratigraphic position of Feature 43 (on the mound were included in a single phase; although top of the primary mound) is secure, so it may be their intra-context contemporaneity should not be that the sample selected to date this feature is assumed, the stratigraphic capping of the primary somehow out of context. mound features by the secondary mound are viewed These new modeled dates attest to an earlier as sufficient evidence they share broad temporal and more compressed history of Garden Creek associations. That said, it is important to keep in Mound No. 2 than has previously been attainable mind that neither these strata nor their features are through relative dating methods and the single being dated directly, but rather pieces of wood con­ extant date. The single date from the pre-mound tained in those features whose radiocarbon ages midden (individually calibrated or modeled) sup­ may not correspond precisely with the age of the ports the inference that features and artifacts in feature (e.g., the "old wood" problem). Therefore, this context date to the late Pigeon or early these results should be considered a step in an in­ Connestee phase, likely the second or third cen­ creasingly precise reckoning of time at Garden turies A.D. It would appear that both mound Creek, but certainly not the last word. building episodes occurred shortly thereafter. The The results of the model are summarized in earliest and latest modeled dates for the features the fifth column of Table 1. They demonstrate in the primary mound are 184 cal A.D. and 360 how a Bayesian approach that uses a priori, strati­ cal A.D., respectively, indicating a maximum of graphic information can clip sigmas that are often less than 200 years of primary mound summit overestimated by standard calibration. In all cases, use in the early Connestee phase. At a 1-sigma this modeled 2-sigma range is notably smaller error range, this period of summit activity is re­ than that yielded by independent calibration of duced to a mere century between 217 cal A.D. Wright] HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 289 and 315 cal A.D. The general accuracy of these and GC2011.02. The plausibility of this scenario dates is bolstered by the single modeled calibrated is supported by good statistical agreement (A > date from the top of the secondary mound (262- 60 percent) for all samples in the Bayesian model. 387 cal A.D. at 2-sigma, 274-320 cal A.D. at 1- The results are listed in Table 2. sigma) that effectively provides a terminus ante The dates obtained from Garden Creek En­ quern for primary summit use. Additional dates closure No. 1 all cluster in the latter half of the are necessary from the secondary summit to fully early Middle Woodland Pigeon phase. The tight­ understand the duration of its use. However, if it ness of this clustering is especially apparent in was used for the same amount of time as the pri­ the results of the modeled sequence. At the 2- mary summit, then it is likely that it was a locus sigma level, the infilling of the ditch, the erection for activity mostly, if not exclusively, during the and dismantling of the post alignment, and the fourth century A.D. This would place the entirety infilling of the postholes appears to have maxi­ of known platform mound activity at Garden mally taken 130 years; at the 1-sigma level, the Creek squarely in the Connestee phase. duration of these activities is further reduced to Three AMS dates were obtained on wood char­ 80 years, beginning by 29 cal A.D. and ending coal recovered during our recent investigation of by 106 cal AD. Enclosure No. 1: one from the bottom zone ditch These dates also agree well with two dates fill, one from the middle of the ditch fill, and one obtained from features excavated in the area sur­ from a piece of charcoal nestled among the cob­ rounded by the ditch enclosure. Lacking obvious bles of one of the rock filled postholes that in­ stratigraphic relationships with the ditch itself, truded into all zones of ditch fill. These dates were these dates were calibrated independently. Wood calibrated both independently and through charcoal from Feature 8, a small cobble hearth Bayesian modeling using stratigraphic information located just inside the westernmost wall of the as a priori knowledge, though as with the mound, enclosure, was dated to 74-254 cal A.D. (cali­ the possibility of old wood presents some chal­ brated at the 2-sigma level). Wood charcoal from lenge and the resulting chronology awaits further Feature 28, a refuse pit closer to the center of the refinement. As mentioned above, it is presently enclosure, produced an assay of 49-125 cal A.D. impossible to say with certainty whether there (calibrated at the 2-sigma level). Based on these was any time lag between the deposition of dif­ assays, then, it would appear that the activities ferent zones of ditch fill. At present, we have no that took place within the enclosure are roughly reason to conclude that there were temporal gaps contemporaneous with the later period of the between the deposition of one zone of fill and the monument's life history, including its infilling next; in fact, the overlap in the independently cal­ and demarcation by posts. ibrated dates from the fill indicate that such gaps At present, we lack a stratigraphic context that are quite unlikely. As a result, the dates from the corresponds with the time when the ditch was bottom and middle zones of fill are modeled as a initially constructed and left open, so it is not single phase. In contrast, we do know that there possible to say how much time elapsed between was a gap between the deposition of these fills these events and the ditch's infilling. However, and the deposition of the rocks, mica fragments, the absence of evidence for soil formation at the and charcoal in the postholes. Before Sample base of the ditch suggests that its infilling may GC2011.04 entered the record, the ditch was not have followed quickly on the heels of its original only "erased" with a final zone of fill, but a post excavation. We also lack excavation data from had been erected in its place, stood for some un­ Enclosure No. 2, directly across from Enclosure known period of time, and was then removed. No. 1, precluding any chronological assessment The cobbles and artifacts that replaced the post of this feature on the basis of artifacts, stratigra­ were packed in such a way that this cannot be phy, or absolute dates. For the present, their iden­ considered incidental fill. This date, then, is not tical footprints, orientations, and alignments to temporally associated with the preceding phase each other strongly suggest that they were part and is modeled as the last date in the sequence of a single architectural design plan, and are thus following the phase defined by dates GC2011.01 contemporaneous. 290 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014

Table 2. Radiocarbon Age Estimates from Enclosure No. 1, Garden Creek, Calibrated Using OxCal Version 4.2.2 and Int.Cal 9 Calibration Curve (Reimer et al. 2009).

Individually Modeled Modeled calibrated calibrated calibrated Sample No. Context 14C years 2-sigma 2-sigma 1-sigma (Lab No.) Description B.P. range range range GC2011.04 rock-filled posthole 1952 ± 40 41 cal B.C. - calA.D. cal A.D. (AA99141) cal A.D. 128 24-135 60-126 GC1966.02 middle of ditch fill 1911 ±40 cal A.D. cal A.D. cal A.D. (AA99139) 5-215 9-124 29-107 GC1966.01 base of ditch fill 1919 ±46 36 cal B.C. - cal A.D. cal A.D. (AA99138) calA.D. 217 4-125 29-108

Discussion likely during the A.D. 200s. Shortly thereafter, the overlying secondary mound was built and its sum­ When the architectural, artifactual, and chrono­ mit used through the late A.D. 300s. Both episodes logical data from Mound No. 2 and Enclosure of mound construction are associated with a few No. 1 are considered together, the result is a com­ pieces of Hopewellian material culture, but lack plex history of Middle Woodland monumentality evidence of Hopewellian craft production. What and interaction at the Garden Creek site as a happened immediately after the second episode of whole. Following low intensity use of the site mound construction is less certain, though addi­ during the Archaic and Early Woodland periods tional stages cannot be ruled out (Keel 1976:86). (Keel 1976:153), early Middle Woodland peoples In sum, we now have evidence for at least two excavated two ditches to create enclosures for complex episodes of interregional interaction at which there is no local precedent. However, there Garden Creek, based on unique articulations of is a strong resemblance between these features monumental architecture and material culture. and small geometric enclosures in the Ohio Valley The first episode, represented by Enclosure No. that date to the late centuries B.C. or early cen­ 1, appears to have most intensively involved in­ turies A.D. While the initial ditch excavation has teraction with Hopewell peoples to the north, dur­ not been securely dated, it was purposefully filled ing the heyday of Hopewellian expression in in during the first century A.D. This infilling ap­ Ohio. The architectural grammar of the Garden pears to have been rapid, perhaps occurring Creek enclosures suggests stronger ritual ties be­ shortly after the original excavation of the ditch, tween the Appalachian Summit and the Ohio Val­ and was associated with the production of cut ley than have previously been acknowledged. As­ mica and crystal quartz artifacts for Hopewellian suming these enclosures constitute a form of ritual exchange. While these infilling efforts could have architecture, it stands to reason that there was a effectively erased any trace of the ditch, its loca­ formally prescribed method for their design, con­ tion continued to be marked after it was filled in struction, and appropriate use. The remarkable by a line of large posts. By the early A.D. 100s, morphological similarities between the locally however, these had been removed and the holes unprecedented enclosures at Garden Creek and they left behind were deliberately filled with river their counterparts in the Adena-Hopewell core cobbles and small artifacts. At this point, the mon­ suggest that they were built according to the same umental "squircles" at Garden Creek would have specific prescriptions, which presumably required been rendered invisible. dissemination through face-to-face contact, pos­ By then, monumental activity had shifted sibly through ritual specialists. It is difficult to slightly to the west, where fairly intensive occu­ explain this interregional architectural pattern as pation produced a number of features, structures, the result of trickle-down diffusion from the Ohio and a midden deposit around 150-250 cal A.D. Valley to the Appalachian Summit. Initial construction of the mound resulted in a sum­ The remains of cut mica and crystal quartz mit that was in use for less than a century, most craft production associated with this enclosure Wright] HISTORY, MONUMENTALITY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 291 also point to ritual connections with the Hopewell out according to the same ritual architectural pre­ core, though in some ways this assemblage raises scriptions as its initial construction. It is also pos­ more questions than answers. On the one hand, sible that the effective erasure of the ditch repre­ the local availability of these raw materials, the sents resistance to or rejection of Hopewellian relative ease with which they can be manipulated, ceremonialism by the local inhabitants at Garden and the local provenance of the associated ceramic Creek, and in turn, a moment of "culture making" assemblage offer no indication that craft produc­ in the Appalachian Summit. This scenario finds tion at Garden Creek was undertaken by anyone some support in the emplacement of posts in the outside the local community. On the other hand, now filled ditch, their subsequent removal, and the utter lack of finished cut mica effigies or crys­ the filling of the resulting postholes with a unique tal quartz bifaces at Garden Creek suggests that matrix. Similarly distinctive post setting, remov­ these artifacts were produced for purely nonlocal, ing, and posthole filling has been noted at both possibly Hopewellian consumption. Did traveling Structure 1 below Mound No. 2 (mentioned Hopewellian ritual practitioners visit the Ap­ above), and around the Middle Woodland Bilt- palachian Summit, share their architectural pre­ more Mound, located less than 30 km east of rogatives, and encourage the production of craft Garden Creek (Kimball et al. 2010). In these re­ objects for use in Ohio? Or, did ceremonial leaders mains, we may have recovered evidence for a from the Appalachian Summit coordinate the pro­ uniquely Appalachian form of Middle Woodland duction of mica and crystal quartz offerings, to monumentality. If so, its assertion immediately be carried via pilgrimage to massive Ohio following seemingly intensively involvement Hopewell ceremonial centers, where they could with Hopewellian interaction and ceremonialism receive instruction regarding ritual activities and merits further examination. architecture to convey back to the mountains? The interregional connections apparent in the Christopher Carr (2006:579-580) hints at both of archaeological record of Mound No. 2 are quite these and similar possibilities in his contextual different. Architecturally speaking, Garden Creek approach to interregional Hopewell (e.g., vision Mound No. 2 is best understood as a Kolomoki and power questing; pilgrimage to powerful nat­ pattern platform mound, and the associated ce­ ural places or ceremonial centers; long-distance ramic assemblage, though dominated by local pot­ buying and selling of ceremonial rites), and though tery, included some pottery from adjacent south­ a bit removed in time and space, similar pilgrim­ eastern areas. Together, these patterns suggest age scenarios have recently been proposed for the involvement with a sphere of monumental and Late Archaic Poverty Point site (Spivey et al. material practice concentrated in the Southeast. 2014). In either case, what remains to be more On the basis of available data, it is difficult to say thoroughly investigated is how or why individuals if Garden Creek and other Kolomoki pattern in the Appalachian Summit acceded to participa­ mounds shared a ritual architectural grammar as tion in Hopewell ceremonialism at all, as crafters precise as that documented among Hopewellian or as long-distance pilgrims. Drawing on com­ enclosures, so it is not clear if their construction parative ethnography, Carr (2006) has begun this would have required formal interaction between project by emphasizing the importance of shaman­ far-flung ritual practitioners. If, as others have ar­ like ideology and practices within Ohio Hopewell, gued (Knight 2001; Lindauer and Blitz 1997), but additional work is required to accommodate these mounds served as loci for community inte­ such active participation of non-Ohio peoples. gration activities, it may be more likely that this Although it is unclear exactly how Ap­ mode of monumentality and associated practices palachian Summit people contributed to Ohio spread through more social means. For instance, Hopewell and to Hopewell-style ritual at Garden Carr's (2006:587) assessment that "intermarriage Creek, the infilling of Enclosure No. 1 offers a at the scale of neighboring groups could have tantalizing suggestion that such interregional re­ been a significant factor in the down-the-line lationships ended abruptly. It may be that infilling spread of Hopewellian practices and ideas" may of the ditch was simply part of the ceremonial be just as applicable to the Kolomoki pattern in life cycle of the monument, and that it was carried the Southeast. How, then, can we account for the 292 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014 seemingly Hopewellian artifacts recovered from terial indicators for several processes of culture the mound? Given the small quantity of this as­ contact and culture making: relatively intensive semblage as a whole, the lack of associated craft (i.e., face-to-face) interactions between far-flung production debris, and complementary lines of ritual specialists; more community-scale interac­ evidence for small-scale feasting activities at other tions to facilitate integration and intermarriage; Kolomoki pattern mounds, the current interpreta­ and perhaps the assertion of local culture and tion of these objects as exotic tokens signaling identity in resistance to nonlocal influences. We the possession of esoteric knowledge and as gifts should not expect this exact historical trajectory distributed in communal ceremonies remains vi­ to map onto the diachronic records of other Mid­ able at Garden Creek. dle Woodland sites in the Southeast, but we should not be surprised if these records reveal Conclusion similarly diverse forms of interaction through time. The processes identified at Garden Creek In light of this dynamic history of monumentality thus comprise both a locally salient historical nar­ and material culture at Garden Creek, it is clear rative and a set of concepts amenable to intersite that unitary visions of interregional interaction comparison. In this regard, this study can be used are insufficient to explain the diverse ways in as springboard for developing more nuanced which southeastern communities engaged with questions about intercultural interaction during nonlocal objects, ideas, and people throughout the Middle Woodland period and the ways in the Middle Woodland period. As discussed above, which diverse human experiences actively con­ even though archaeologists have previously ac­ stituted the persistently enigmatic phenomena knowledged that such interactions played out in that we have referred to as "interaction spheres." different ways in different places, most extant models for Hopewellian and other intercultural Acknowledgments. This research was possible thanks to Ben- networks implicitly presuppose their overall his­ nie Keel's meticulous documentation of Mound No. 2, and torical stability. In seeking to provide sub-conti­ the University of North Carolina's Research Laboratories of nental scale explanations for these phenomena, Archaeology's curation of resulting materials, to which R. P. Stephen Davis, Vincas Steponaitis, and Meg Kassabaum pro­ these frameworks have sacrificed temporal reso­ vided access. GCAP fieldwork relied on the generosity of the lution, the consequence of which has been a con­ residents of Plott Farm, the expertise and assistance of Tim ceptual divide between the processes of interac­ Horsley and the University of Michigan's Eastern North tion and the people who actually did the America Range contingent, and support from the National Science Foundation (Grant No. BCS-1225872), the Griffin interacting. Fellowship, and Arts of Citizenship at the University of Michi­ Detailed site histories, like that offered here gan. Thank you to Lacey Carpenter for providing the Spanish for Garden Creek, stand to complement such abstract. This article has benefited from comments and con­ broad explanatory frameworks. Certainly, wide­ versation with Robin Beck, Cameron Gokee, Casey Barrier, spread material signatures justify investigation and Thaddeus Bissett, but any mistakes are my own. of Middle Woodland interaction at the macroscale. Just as certainly, however, the people References Cited who contributed to these interactions were dis­ Alt, Susan M. persed in space and time, not to mention distin­ 2006 The Power of Diversity: The Roles of Migration and guished by diverse linguistic, political, commu­ Hybridity in Culture Change. In Leadership and Polity in nal, and individual identities. The idea that a Mississippian Society, edited by Brian M. Butler and Paul D. Welch, pp. 289-308. Center for Archaeological single framework for interaction could have sub­ Investigations, Southern Illinois University, Carbondale. sumed all of this cultural and temporal diversity Anderson, David G. is not only unrealistic, but also at odds with an 1998 Swift Creek in Regional Perspective. In A World En­ graved: Archaeology of the Swift Creek Culture, edited historical processual paradigm that views diverse by Mark W. Williams and Daniel T. Elliott, pp. 274-300. processes of encounter as the driving forces of University of Alabama Press, Tuscaloosa. cultural transformation. 2013 Social Landscapes of the Early and Middle Woodland Southeast. In Early and Middle Woodland Landscapes of At Garden Creek, instead of evidence for an the Southeast, edited by Alice P. Wright and Edward R. all-or-nothing mode of interaction, we have ma­ Henry, pp. 247-262. University Press of Florida, Gainesville. Wright] HISTORY, MONUMENTALLY, AND INTERACTION IN THE APPALACHIAN SUMMIT MIDDLE WOODLAND 293

Anderson, David G., and Robert C. Mainfort, Jr. Eastman, Jane M. 2002 An Introduction to Woodland Archaeology in the 1994 North Carolina Radiocarbon Dates. Southern Indian Southeast. In The Woodland Southeast, edited by David Studies 43:1-83. G. Anderson and Robert C. Mainfort, Jr., pp. 1-19. Uni­ Greber, N'omi B. versity of Alabama Press, Tuscaloosa. 2006 Enclosures and Communities in Ohio Hopewell: An Bayliss.Alex Essay. In Recreating Hopewell, edited by Douglas C. 2009 Rolling Out Revolution: Using Radiocarbon Dating Charles and Jane E. Buikstra, pp. 74-105. University in Archaeology. Radiocarbon 51:123-147 Press of Florida, Gainesville. Beck, Robin A., Jr., Douglas J. Bolender, James A. Brown, Greenman, Emerson F. and Timothy K. Earle 1938 Hopewellian Traits in Florida. American Antiquity 2007 Eventful Archaeology: The Place of Space in Structural 3:327-332. Transformation. Current Anthropology 48: 833-860. Griffin, James B. Braun, David P. 1967 Eastern North American Archaeology: A Summary. 1986 Midwestern Hopewellian Exchange and Supralocal Science 156:175-191. Interaction. In Peer Polity Interaction and Socio-Political Hall, Robert L. Change, edited by Colin Renfrew and John F. Cherry, pp. 1997 An Archaeology of the Soul: North American Indian 117-126. Cambridge University Press, Cambridge. Belief and Ritual. University of Illinois Press, Urbana. Bronk Ramsey, Christopher Helms, Mary W. 2009 Bayesian Analysis of Radiocarbon Dates. Radiocarbon 1988 Ulysses' Sail: An Ethnographic Odyssey of Power, 51:337-360. Knowledge, and Geographical Distance. Princeton Uni­ Brose, David S., and N'omi B. Greber versity Press, Princeton, New Jersey. 1979 Hopewell Archaeology: The Chillicothe Conference. Holmes, William H. Kent State University Press, Ohio. 1919 Handbook of Aboriginal American Antiquities Part Burks, Jarrod, and Robert A. Cook I: Introductory, The Lithic Industries. Smithsonian 2011 Beyond Squier and Davis: Rediscovering Ohio's Institution Bureau of American Ethnology, Washington, Earthworks Using Geophysical Remote Sensing. American DC. Antiquity 76:667-689. Horsley, Timothy J., Alice P. Wright, and Casey R. Barrier Byers, A. Martin 2014 Prospecting for New Questions: Integrating Geophysics 2004 The Ohio Hopewell Episode: Paradigm Lost, Paradigm to Define Anthropological Research Objectives and Inform Gained. University of Akron Press, Ohio. Excavation Strategies at Monumental Sites. Archaeological Caldwell, Joseph R. Prospection 21:75-86. 1964 Interaction Spheres in Prehistory. In Hopewellian Jefferies, Richard W. Studies, edited by Joseph R. Caldwell and Robert L. Hall, 1994 The Swift Creek Site and Woodland Platform Mounds pp. 133-143. Illinois State Museum, Springfield. in the Southeastern United States. In Ocmulgee Archaeology, Carr, Christopher 1936-1986, edited by David Hally, pp. 71-83. University 2006 Rethinking Interregional Hopewellian "Interaction." of Georgia Press, Athens. In Gathering Hopewell: Society, Ritual, and Ritual Inter­ Jefferies, Richard W., George R. Milner, and Edward R. Henry action, edited by Christopher Carr and D. Troy Case, pp. 2013 Winchester Farm: A Small Adena Enclosure in Central 575-623. Springer, New York. Kentucky. In Early and Middle Woodland Landscapes of Chapman, Jefferson, and Bennie C. Keel the Southeast, edited by Alice P. Wright and Edward R. 1979 Candy Creek-Connestee Components in Eatem Tennessee Henry, pp. 91-107. University Press of Florida, Gainesville. and Western North Carolina and their Relationship with Keel, Bennie C. Adena-Hopewell. In Hopewell Archaeology: The Chillicothe 1976 Cherokee Archaeology: A Study of the Appalachian Conference, edited by David S. Brose and N'omi B. Greber, Summit. University of Tennessee Press, Knoxville. pp. 157-161. Kent State University Press, Ohio. Keith, Scot Clay, R. Berle 2013 The Woodland Period Cultural Landscape of the 1987 Circles and Ovals: Two Types of Adena Space. Leake Site Complex. In Early and Middle Woodland Southeastern Archaeology 6:46-56. Landscapes of the Southeast, edited by Alice P. Wright 1998 The Essential Features of Adena Ritual and Their and Edward R. Henry, pp. 138-152. University Press of Implications. Southeastern Archaeology 17:1-21. Florida, Gainesville. 2002 Deconstructing the Woodland Sequence from the Kimball, Larry R., Thomas R. Whyte, and Gary D. Crites Heartland: A Review of Recent Research Directions in 2010 The Biltmore Mound and Hopewellian Mound Use the Upper Ohio Valley. In The Woodland Southeast, in the Southern Appalachians. Southeastern Archaeology edited by David G.Anderson and Robert C. Mainfort. Jr., 29: 44-58. pp. 162-184. University of Alabama Press, Tuscaloosa. Knight, Vernon James, Jr. 2005 Adena: Rest in Peace? In Woodland Period Systematics 1990 Excavation of the Truncated Mound at the Walling in the Middle Ohio Valley, edited by Darlene Applegate Stie: Middle Woodland Culture and Copena in the and Robert C. Mainfort. Jr., pp. 94-110. University of Tennessee Valley. Office of Archaeological Research, Al­ Alabama Press, Tuscaloosa. abama State Museum of Natural History, Huntsville. Cobb, Charles R. 2001 Feasting and the Emergence of Platform Mound 2005 Archaeology and the "Savage Slot": Displacement Ceremonialism in Eastern North America. In Feasts: Ar­ and Emplacement in the Premodern World. American chaeological and Ethnographic Perspectives on Food, Anthropologist 107:563-574. Politics, and Power, edited by Michael Dietler and Brian DeBoer, Warren R. Hayden, pp. 239-254. Smithsonian Institution Press, 1997 Ceremonial Centres from the Cayapas (Esmeraldas, Washington, D.C. Ecuador) to Chillicothe (Ohio, USA). Cambridge Ar­ Lindauer, Owen, and John H. Blitz chaeological Journal 7:225-253. 1997 Higher Ground: The Archaeology of North American 294 AMERICAN ANTIQUITY [Vol. 79, No. 2, 2014]

Platform Mounds. Journal of Archaeological Research Squier, Ephraim G., and Edwin H. Davis 5:169-207. 1988 [1848] Ancient Monuments of the Mississippi Valley, Lynott, Mark J. Comprising the Results of Extensive Original Surveys 2004 Earthwork Construction and the Organization of and Explorations. Contributions to Knowledge. Smithsonian Hopewell Society. Hopewell Archaeology•: The Newsletter Institution, Washington, D.C. of Hopewell Archaeology in the Ohio River Valley 6, No. Stoltman, James B. 1. http://www.nos.gov/history/mwac/hopewell/v6nl/sixjitm, 1999 Icehouse Bottom and the Hopewell Connection. Re­ accessed February 1,2013. search Notes, Frank H. McClung Museum 17. Pacheco, Paul J. Streuver, Stuart, and Gail L. Houart 1996 Introduction. In A View from the Core: A Synthesis 1972 An Analysis of the Hopewell Interaction Sphere. In of Ohio Hopewell Archaeology, edited by Paul J. Pacheco, Social Exchange and Interaction, edited by Edwin N. pp.vi-vii. Ohio Archaeological Council, Columbus. Wilmsen, pp. 47-79. Museum of Anthropology, University Pauketat, Timothy R. of Michigan, Ann Arbor. 2001 Practice and History in Archaeology: An Emerging Walthall, John A. Paradigm. Anthropological Theory 1:73-98. 1985 Early Hopewellian Ceremonial Encampments in the Reimer, P.J., M.G.L. Baillie, E. Bard, A. Bayliss, J.W. Beck, South Applachian Highlands. In Structure and Process in P.G. Blackwell, and C. Bronk Ramsey Southeastern Archaeology, edited by Roy S. Dickens and 2009 IntCal09 and Marine09 Radiocarbon Age Calibration H. Trawick Ward, pp. 243-262. University of Alabama Curves, 0-50,000 years cal BP. Radiocarbon 51:1111- Press, Tuscaloosa. 1150. Ward, H. Trawick, and R. P. Stephen Davis, Jr. Ruby, Bret J., and Christine M. Shriner 1999 Time Before History: The Archaeology of North Car­ 2005 Ceramic Vessel Compositions and Styles as Evidence olina. University of North Carolina Press, Chapel Hill. of the Local and Nonlocal Social Affiliations of Ritual Webb, William S., and Charles E. Snow Participants at the Mann Site, Indiana. In Gathering 1945 The Adena People. Reports in Anthropology and Ar­ Hopewell: Society, Ritual, and Ritual Interaction, edited chaeology 6. University of Kentucky, Lexington. by Christopher Carr and D. Troy Case, pp. 553-572. Wright, Alice P. Springer, New York. 2013 Under the Mound: The Early Life History of the Sassaman, Kenneth E. Creek Mound No. 2 Site. In Social Landscapes of the 2010 The Eastern Archaic, Historiciied. AltaMira, Lanham, Early and Middle Woodland Southeast, edited by Alice P MD. Wright and Edward R Henry, pp. 108-121. University Sears, William H. Press of Florida, Gainesville. 1962 The Hopewellian Affiliations of Certain Sites on the 2013 Monumental Memes: The Architectural Grammar of Gulf Coast of Florida. American Antiquity 28:5-18. Early-Middle Woodland Enclosures in the Southeast. Seeman, Mark F. Manuscript on file, Museum of Anthropology, University 1979 The Hopewell Interaction Sphere: The Evidence for of Michigan, Ann Arbor, Michigan. Interregional Trade and Structural Complexity. Indiana Wright, Alice P., and Erika Loveland Historical Society, Indianapolis. 2013 Ritualized Craft Production at a Hopewell Periphery: 1995 When Words are Not Enough: Hopewell Interre- New Evidence from the Appalachian Summit. Manuscript gionalism and the Use of Material Symbols at the GE on file, Museum of Anthropology, University of Michigan, Mound. In Native American Interactions: Multiscalar Ann Arbor, Michigan. Analyses and Interpretations in the Eastern Woodlands, edited by Michael S. Nassaney and Kenneth E. Sassaman, pp. 122-143. University of Tennessee Press, Knoxville. Notes 2004 Hopewell Art in Hopewell Places. In The Hero, Hawk, and the Open Hand: American Indian Art of the 1. There was no evidence in the geophysical survey results Ancient Midwest and South, edited by Richard Townsend, or in excavation of Enclosure No. 1 that the ditches at Garden pp. 57-72. Art Institute of Chicago, Chicago. Creek were accompanied by an embankment. However, their Sherwood, Sarah C, and Tristram R. Kidder possible existence in prehistory can not be ruled out given the 2011 The DaVincis of Dirt: Geoarchaeological Perspectives site's history of plowing and grading for residential on Native American Mound Building in the Mississippi River Basin. Journal of Anthropological Archaeology devlopment. 30:69-87. 2. Historically, archaeologists have situated Adena as the Shetrone, Henry C. chronological predecessor of Hopewell in the Ohio Valley, 2004 [1930] The Moundbuilders. University of Alabama but recent research has demonstrated that this presumed rela­ Press, Tuscaloosa. tionship is not necessarily accurate (Clay 2002,2005). Spielmann, Katherine A. 3. These summary results are necessarily preliminary, be­ 2009 Ohio Hopewell Ritual Craft Production. In Footprints: cause these databases derive at least in part from nineteenth- In the Footprints ofSquier and Davis, edited by Mark J. century maps. Recent geophysical surveys (e.g., Burks and Lynott, pp. 179-188. Midwest Archaeological Center, Cook 2011) have demonstated that these depictions in these Lincoln, Nebraska. Spivey, S. Margaret, Tristram R. Kidder, Anthony A. Ortmann, maps can differ significantly from archaeological realities. and Lee Arco 2014 Pilgrimage to Poverty Point? In The Enigma of the Event: Moments of Consequence in the Ancient Southeast, edited by Zackary Gilmore and lason O'Donoughue. Submitted August 5, 2013; Revised November 22, 2013; University of Alabama, Tuscaloosa, in press. Accepted December 2, 2013.