kolomoki settlement ceremony and status in the deep south a d 350 to 750 sep 2003

6.5 Contour map of the structure in Block A 1546.6 Densities of ceramics and ®aked stone in Block A 159 6.7 Ground stone pendants from Block A and vicinity 164 6.8 Locations of features

Kolomoki

Publication of this work has been supported in part by the

Dan Josselyn Memorial Fund

Kolomoki Settlement, Ceremony, and Status in the Deep South, a.d 350 to 750

Thomas J Pluckhahn

t h e u n i v e r s i t y o f a l a b a m a pr e s s

Tuscaloosa and London

Copyright © 2003 The University of Alabama Press Tuscaloosa, Alabama 35487-0380 All rights reserved Manufactured in the United States of America Typefaces are AGaramond and Triplex

∞ The paper on which this book is printed meets the minimum requirements of American National Standard for Information Science–Permanence of Paper for Printed Library Materials,

ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Pluckhahn, Thomas J (Thomas John), 1966–

Kolomoki : settlement, ceremony, and status in the Deep South,

a.d 350 to 750 / Thomas J Pluckhahn.

p cm.

Includes bibliographical references and index.

ISBN 0-8173-1299-4 (cloth : alk paper) — ISBN 0-8173-5017-9 (pbk : alk paper)

1 Kolomoki Mounds State Historic Park (Ga.) 2 Woodland culture—Georgia.

3 Excavations (Archaeology)—Georgia I Title.

E78.G3 P68 2003 975.8 ′ 901—dc21 2003001799

British Library Cataloguing-in-Publication Data available

Figures and Tables vii

Acknowledgments xiii

1 An Introduction to Kolomoki 1

2 Putting Kolomoki in Its Place: De¤ning the Temporal,Ecological, and Cultural Contexts 15

3 Preliminary De¤nition of Activity Areas at

Kolomoki: Summary of Previous Research 47

4 De¤ning Activity Areas at Kolomoki: Results

2.3 Large survey and excavation projects within 200 km

3.1 Valliant’s 1937 map of Kolomoki 50

3.2 Locations of con¤rmed and reported mounds at Kolomoki 523.3 Aerial photograph of Kolomoki, 1948 57

3.4 Waring and Sears in Mound D 61

3.5 Excavation of the ceramic cache in Mound D 61

3.6 Excavation of Mound E 65

Figures and Tables

3.7 Excavation of the central burial pit in Mound E 65

3.8 Topographic map of mounds excavated by Fairbanks 703.9 Fairbanks’s plan and pro¤le maps of Mound K 71

3.10 Locations of previous studies in off-mound

areas at Kolomoki 74

4.1 Locations of intensive samples 93

4.2 Density of Woodland ceramics in samples 94

4.3 Density of Swift Creek ceramics in samples 96

4.4 Density of Weeden Island Red ceramics in samples 974.5 Density of Carrabelle Incised/Punctate ceramics in samples 984.6 Density of other Weeden Island ceramics in samples 1004.7 Density of Napier ceramics in samples 101

4.8 Density of ®aked stone in samples 102

4.9 Density of Coastal Plain chert in samples 103

4.10 Density of quartz in samples 105

4.11 Density of quartzite in samples 106

4.12 Density of Ridge and Valley chert in samples 107

4.13 Locations of activity areas 109

4.14 Ceramic density by activity area 116

4.15 Ceramic ubiquity by activity area 117

4.16 Lithic density by activity area 118

4.17 Lithic ubiquity by activity area 119

5.1 Locations of test units 127

5.2 Locations of geophysical prospection grids 137

6.1 Locations of excavation blocks 147

6.2 Locations of features in Block A 149

6.3 Feature 57 during initial stages of excavation 152

6.4 Feature 57 during ¤nal stages of excavation 153

viii

6.5 Contour map of the structure in Block A 154

6.6 Densities of ceramics and ®aked stone in Block A 159

6.7 Ground stone pendants from Block A and vicinity 164

6.8 Locations of features in Block B 167

6.9 Pro¤le of Block B excavation units 170

6.10 Densities of ceramics and ®aked stone in Block B 174

6.11 Locations of features in Block C 176

7.1 Kolomoki I phase settlement and community plan 188

7.2 Kolomoki II phase settlement and community plan 199

7.3 Kolomoki III phase settlement and community plan 209

7.4 Possible higher status ceramics from Kolomoki III and

IV phase contexts 210

7.5 Kolomoki IV phase settlement and community plan 214

Tables

2.1 Proportions of ceramic types in test units by phase 18

2.2 Summary of proposed ceramic phases for Kolomoki 20

2.3 Recent radiocarbon dates from Kolomoki 21

2.4 Ceramic types represented at Fairchild’s Landing and

Hare’s Landing 23

2.5 Rim treatments represented in a sample of plain sherds

from Fairchild’s Landing 25

2.6 Summary of rim treatments and vessel forms represented in

test unit assemblages by phase 26

2.7 Vessel forms represented in a sample of plain sherds with

unmodi¤ed rims from Fairchild’s Landing 27

2.8 Summary of PP/K types in test units by phase 29

3.1 Chronology of events related to the history and archaeology

of Kolomoki 48

3.2 Ceramic totals for mound assemblages 54

ix

3.3 Summary of rim treatments and vessel forms represented inmound assemblages 55

3.4 Ceramic totals for assemblages from Sears’s excavations inthe North Ravines 75

3.5 Summary of rim treatments and vessel forms in assemblagesfrom Sears’s excavations in the North Ravines 76

3.6 Ceramic totals for assemblages from Sears’s excavations inthe Northwest Area and Central Plaza 79

3.7Summary of rim treatments and vessel forms in assemblages fromSears’s excavations in the Northwest Area and Central Plaza 803.8 Ceramic totals for assemblages from investigations in

off-mound areas 84

4.1 Ceramic totals for the pooled assemblages from samples byactivity area 112

4.2 Ceramic density and ubiquity values for the pooled

assemblages from samples by activity area 113

4.3 Lithic totals for the pooled assemblages from samples by

5.1 Summary data for features in test units 129

5.2 Ceramic totals for test units 131

5.3 Lithic totals for test units 133

5.4 Miscellaneous artifacts from test units 134

5.5 Summary of macro-botanical remains from test units 1355.6 Summary of faunal remains from test units 136

5.7Comparison of artifact density and feature density in

test units 138

6.1 Summary data for features in Block A 150

x

6.2 Ceramic totals for Block A 158

6.3 Summary of rim treatments and vessel forms in the assemblage

from Block A 160

6.4 Flaked stone totals for Block A 161

6.5 Miscellaneous artifacts from Block A 162

6.6 Summary of faunal remains from Block A 165

6.7 Summary of macro-botanical remains of edible and medicinal

plants from Block A 166

6.8 Summary data for features in Block B 168

6.9 Ceramic totals for Block B 171

6.10 Summary of rim treatments and vessel forms in the assemblage

from Block B 172

6.11 Flaked stone totals for Block B 173

6.12 Miscellaneous artifacts from Block B 175

6.13 Summary of macro-botanical remains from Block B 175

6.14 Summary data for features in Block C 176

6.15 Ceramic totals for Block C 177

6.16 Flaked stone totals for Block C 177

7.1 Summary of macro-botanical remains of edible and medicinal

plants from recent work at Kolomoki 186

7.2 Population estimates for the Kolomoki I and II phases 191

7.3 Estimated labor requirements for mound construction

by phase 193

7.4 Population estimates for the Kolomoki III and IV phases 210

xi

I am indebted to many people for their contributions to this book First andforemost, I must thank my major professor, Steve Kowalewski, for his unwaveringsupport and enthusiasm We never succeeded in getting Steve to set foot in asquare hole, but he participated in virtually every other stage of the project, fromproposal writing to cutting transit lines and back-¤lling It would be dif¤cult tooverstate Steve’s contribution to this research or, for that matter, to Georgia ar-chaeology in general.

I also extend my thanks to the other members of my committee, includingDavid Hally, Charles Hudson, Ervan Garrison, and Elizabeth Reitz Thanks aswell to the unof¤cial sixth member of the committee, Mark Williams, for hiscontinued reassurance that this was a signi¤cant and worthy endeavor

I am grateful to the National Geographic Society for funding my research.Additional support was provided by the LAMAR Institute and by a Joshua LaermMemorial Award from the Georgia Museum of Natural History Institutionalsupport was provided by the University of Georgia and the State Parks and His-toric Sites Division of the Georgia Department of Natural Resources

It has been many years since academic research was conducted at a GeorgiaState Park, and I am grateful to a number of individuals at the Georgia Depart-ment of Natural Resources and Kolomoki Mounds State Historic Park for theirassistance in bringing the project to fruition In particular, I thank David Crass,Billy Townsend, and Chip Morgan for supporting my research At Kolomoki,

my work was greatly facilitated by park superintendents Eric Bentley and LarryBlankenship, as well as park personnel Billy Adams, Matt Bruner, Judy Moore,and Sid Sewell Permission to investigate the privately owned portions of the sitewas generously granted by Buddy Jenkins and Mike and Meredith Whitehead.Over the course of three ¤eld seasons, I have had the pleasure of working

Acknowledgments

with a number of ¤ne students, and I thank them all for their efforts The 1998crew members, each of whom worked at Kolomoki for a week, included WillChambers, Erica Dougherty, Natalie Faulkner, Heather Hayes, Josuah Hendrick,Donna Howard, Ryan Hurd, Kim Lewis, Carrie McAlister, Katie Price, BenRichardson, Debbie Rose, Sandy Sekman, Ben Sellers, Chris Swindell, BrianTibbles, and Caroline Wardlaw.

The 2000 ¤eld crew members, who stayed with me the entire summer, deservespecial thanks for much of the dirty work of cutting survey lines and excavatingshovel tests This small group included Tiffany Andrews, James Mauldin, SilasMullis, and Dennis Wardlaw

In 2001, we ran a short ¤eld school in shallow geophysics, followed by a longersession in archaeological ¤eld methods The students for the short session in-cluded Laina Davis, Jennifer Hart, and Megan Risse The 2001 ¤eld school par-ticipants, who provided exceptional service in the excavation of the house, con-sisted of D anielle James, D avid Krizan, Iva Lee Lane, Susannah Lee, CarmenLovvorn, and Julie Ordelt

Three of these students—Danielle James, James Mauldin, and Silas Mullis—worked with me as interns after their ¤eld school sessions Their work greatlyfacilitated artifact analysis and data processing

Each ¤eld season, we extended an invitation to the general public to work with

us at Kolomoki The response was overwhelming, and the volunteers who tributed their time are too numerous to mention I thank them all for their effortsand enthusiasm

con-A number of friends and colleagues lent their assistance to the ¤eldwork Mydeepest thanks go to fellow graduate students Verónica Pérez Rodríguez and MattCompton for their help in directing the 2000 and 2001 ¤eld schools (respec-tively) I also thank Nina Šerman and Jill Wesselman for their assistance with the

2001 geophysics ¤eld school I am greatly indebted to those colleagues who unteered their labor, including Nichole Gillis, Bill Jurgelski, Heather Mauldin,Melissa Memory, Betsy Shirk, Keith Stephenson, Victor Thompson, Jamie Wag-goner, and Jared Wood

vol-Mary Theresa Bonhage-Freund conducted the analysis of the macro-plant mains from my excavations Analysis of the faunal assemblage was completedprimarily by Matt Compton, with additional assistance from Tiffany Andrewsand Barnet Pavao-Zuckerman

re-Other colleagues have assisted me by granting access to data I thank DennisBlanton, Ken Johnson, Jerald Ledbetter, Frankie Snow, and Karl Steinen for lend-ing me source material that greatly aided my interpretation of the site I am par-ticularly grateful to Chris Trowell for generously sharing the results of his exten-xiv

sive research on the history of Kolomoki Thanks also to Richard Vernon and theNational Park Service for the loan of collections and documentation relating toFairbanks’s work at Kolomoki.

My compilation of site ¤les data was abetted by the personnel of the agenciesthat oversee the data in Alabama, Florida, and Georgia My thanks to EugeneFutato, Michelle Cremer, and Mark Williams for their assistance

Tom Gresham and Chad Braley offered me gainful employment at crucialpoints throughout my career as a graduate student I thank them for their pa-tience and support, as well as for the occasional loan of equipment for my re-search I must also thank Charlotte Blume for her assistance in guiding methrough the vagaries of academic bureaucracy

The book bene¤tted greatly from comments and suggestions of Tim Pauketatand Frank Schnell

Last, but perhaps most important, I thank my family—especially my parents—for helping me through the occasional hard times

xv

Kolomoki

Not so, however the Mercier Mound Impelled by some strange impulse, ornecessity, its builders have made it a structure destined to stand till the end oftime By whom, when, and for what this massive work was thrown up, arequestions never to be answered by vague conjecture or inde¤nite speculation.

“The Indian Mounds in Early County,”

Early County News, June 27, 1882

Kolomoki—formerly known as the Mercier Mounds—is one of the most sive archaeological sites in the southeastern United States (Figure 1.1) The site,located in the lower Chattahoochee Valley of southwestern Georgia, includes atleast nine mounds (Figure 1.2) The largest of these (Mound A) rises some 56 ft(17 m) from the broad, high terrace on which the site is located From the rec-tangular summit of Mound A, one can observe the string of smaller mounds (B,

impres-H, G, F, and E) that forms an arc to the south Mound C, to the north of Mound

A, frames the opposite side of this open area Another mound (D) stands nently near the center of the site, almost directly west of Mound A Historicalaccounts point to the former existence of several other mounds and a largeearthen wall or enclosure

promi-The number and magnitude of its earthworks make Kolomoki impressive,but these have also been a source of confusion The site was originally thought todate to the Mississippian period (ca a.d 1000–1500), when agricultural chief-doms built tall platform mounds and palisaded villages throughout the South-east It is now generally agreed that the primary occupation at Kolomoki dates

to the Middle and Late Woodland periods, before the era of the Mississippianchiefdoms

1

An Introduction to Kolomoki

1.1 Location of Kolomoki and other sites mentioned in the text

Kolomoki and a few other Middle and Late Woodland mound sites confoundsimple categorizations of Woodland and Mississippian, tribe and chiefdom, egali-tarian and ranked, and simple and complex societies This book presents a his-tory of Kolomoki from its founding at roughly a.d 350 to its eventual abandon-ment at around a.d 750, with particular attention to Woodland period economyand settlement I have contextualized this within broader anthropological con-cerns regarding the nature of middle-range societies, especially the role of ritualand ceremony in the development of status differentiation among such social for-mations However, the most fundamental concern is to illuminate the historicaldevelopment of an important, but long misunderstood, site for both the archaeo-logical community and the lay public.

Before delving into the research, it is important to provide some context, ginning with a description of what has become known, among some archaeolo-gists, as “the Kolomoki problem.” I then consider some of the challenges thatarchaeologists have confronted in attempting to interpret middle-range societiessuch as Kolomoki Finally, I detail the theoretical framework and research strategywith which I intend to address some of the research issues that I have raised

be-“The Kolomoki Problem”

Kolomoki was ¤rst described in print more than a century ago ( Jones 1873;Pickett 1851; White 1854) The site was subsequently investigated by severalarchaeologists and antiquarians loosely af¤liated with the Smithsonian Institution(McKinley 1873; Palmer 1884) These early investigations, despite being poorlycontrolled and inadequately reported, nevertheless described important featuresthat were subsequently lost to agriculture and erosion

Contemporary archaeological work at Kolomoki began in the late 1930s underthe direction of Charles Fairbanks and Robert Wauchope (Fairbanks 1940a,1940b, 1941a, 1941b, 1946) More intensive excavations were initiated about adecade later by William Sears, who excavated six of the mounds and conductedlimited testing in the presumed village area (Sears 1950, 1951a, 1951b, 1953a,1953b, 1956) Mounds D and E proved to contain elaborate mortuary complexesthat included Hopewellian artifacts such as copper ear spools, meteoric iron or-naments, and mica disks Mounds F and H, on the other hand, each consisted ofsmall platforms later covered by conically shaped capping layers Sears admitted

to some confusion regarding the structure and purpose of Mounds B and C, butthe former apparently resulted from the piling of earth (perhaps sweepings fromthe plaza) around large posts Owing to its large size, only minimal investigationwas conducted on Mound A

Sears correctly identi¤ed the dominant ceramic varieties at Kolomoki as theSwift Creek and Weeden Island types, which had been de¤ned by other research-ers working in the region ( Jennings and Fairbanks 1939; Kelly and Smith 1975;Milanich 1994; Willey 1945, 1949; Willey and Woodbury 1942) However,Sears declined to accept the growing consensus that these were Woodland potterytypes He instead inverted the ceramic chronology to force the dominant occu-pation into the Early Mississippian period, around a.d 1200 (Sears 1956) Sears’serror became known to some archaeologists as “the Kolomoki problem” (Trowell1998).

With the bene¤t of hindsight, it is dif¤cult for us to fully comprehend thereasons for Sears’s mistake However, we must bear in mind that his work at Kolo-moki came near the climax of development of the cultural historical paradigm,

a time in which Southeastern archaeologists were working to establish regional chronologies based on broad similarities among sites (Ford and Willey1941; Grif¤n 1952; Willey and Sabloff 1980) Kolomoki, with its large, ®at-topped mound and elaborate mortuary ceramics, clearly did not ¤t the mold thathad been cast for the Woodland period in the Southeast Constrained by thenormative paradigm of his day, Sears found it easier to believe that the recentlyestablished ceramic chronology for the area was faulty than to accept the notionthat the broad cultural historical sequences for the Southeast could be in error.The reluctance to attribute Mound A, as well as some of the elaborate ceramics

macro-in Mounds D and E, to a Woodland period society was not limited to Sears JesseJennings (1938:1–2), while recognizing the Woodland period occupation of thesite as the most substantial, believed that Mound A must have been the product

of a Mississippian culture Ripley Bullen was also reluctant to attribute the site to

a Woodland society (letter of Bullen to Joseph Caldwell, April 15, 1954, sity of Georgia Laboratory of Archaeology) James Grif¤n, Sears’s academic ad-visor at the University of Michigan, confessed to some confusion regarding theproposed chronology and dismay at the contradiction of established ceramic se-quences, but he nevertheless ultimately concurred with Sears (letter of Grif¤n toSears, December 15, 1950, University of Michigan Museum of Anthropology[UMMA]; Grif¤n 1984) Charles Fairbanks was receptive to Sears’s conclusions(Fairbanks 1956:11–12), although he was also one of the ¤rst to question hisinterpretations in print (letter of Fairbanks to James Grif¤n, February 2, 1942,UMMA; Fairbanks 1952, 1955)

Univer-Many other prominent archaeologists of the day were unconvinced by Sears’sinterpretation of Kolomoki as a Mississippian site Gordon Willey, Joseph Cald-well, and John Goggin objected early and strongly to Sears’s chronology, albeitprimarily in personal correspondence (letter of Grif¤n to Sears, November 21,

5

1950, UMMA; letter of Sears to Grif¤n, December 8, 1950, UMMA; Caldwell1958).1 The de¤nitive rebuttal to Sears’s chronology came from Stephen Williams

(1958), who reviewed the ¤nal report for American Antiquity: “The confusion

arising from an equation of Temple Mound to Mississippian is obvious Tothink of Kolomoki as ‘Mature Mississippi’ because of its temple mound was thefalse step that led Sears’s chronology astray.” Sears himself admitted his errorshortly before his death, effectively laying the controversy to rest (Sears 1992).Over the course of the past 50 years, a great deal of corroborating evidencehas ¤rmly established that the dominant ceramic types at Kolomoki date to theMiddle and Late Woodland periods ( Jenkins 1978; Knight and Mistovich 1984;Milanich 2002; Smith 1977; Stephenson et al 2002) In addition, Woodlandplatform mounds, while not common, occur with greater regularity than waspreviously thought by Southeastern archaeologists (Brose 1988; Jefferies 1994;Knight 1990, 2001; Mainfort 1986, 1988a, 1988b; Pluckhahn 1996) Althoughthe identi¤cation of Kolomoki as a Woodland period ceremonial center is todaywidely accepted, the site has remained enigmatic.2 Very basic questions have, untilnow, gone largely unanswered: How big is the site? What was the economy? Howmuch of the occupation was permanent and domestic in function? What wereSwift Creek and Weeden Island houses like? The lack of basic knowledge is due

to the limited scope and poor documentation of previous work in off-moundportions of the site, little follow-up research since the 1950s, and until very re-cently the absence of survey and excavation in the surrounding area

Kolomoki and the Classi¤cation of Middle-Range Societies

Archaeologists in the Midwest and Southeast have long differed in their tation of the level of cultural complexity that is apparent in the material remains

interpre-of the Middle and Late Woodland periods (compare, for example, Anderson1998; Bender 1985a, 1985b; Braun 1979, 1986; Dancey and Pacheco 1997;Milanich et al 1997; Seeman 1979; Smith 1986, 1992; Tainter 1977, 1983).Most now agree that social ranking existed during the Woodland period, butthey disagree widely over whether such differences were achieved or inherited(McElrath et al 2000:4), and thus they argue about whether Woodland societiesshould be considered tribes or chiefdoms

Consistent with his interpretation of the site as a Mississippian period moundcenter, Sears (1956) borrowed heavily from ethnographic accounts of the lateseventeenth-century Natchez to describe the prehistoric inhabitants of Kolomoki

as a chiefdom He envisioned the site as a densely occupied village of around2,000 people, supported by corn agriculture In later accounts, Sears (1968) de-scribed the site as the administrative center of a state-level society

Some archaeologists, while rejecting Sears’s chronology, accept his basic pretation of the site as a chiefdom (Anderson 1998:287; Milanich et al 1997:21;Steinen 1977) or “proto-chiefdom” (Steinen 1995, 1998) Still other researchersargue that while status distinctions may have arisen at sites such as Kolomoki,these differences can be accommodated within the prevailing model of Woodlandsociopolitical organization as a tribal or segmentary system headed by a big man

inter-or big woman (Brose 1979a:143–144; Scarry 1996:233–234)

In the absence of additional research, attempts to classify Kolomoki into one

or another category have relied principally on Sears’s work, which was heavilybiased toward mounds Certainly, Sears’s mound excavations provide an intrigu-ing glimpse of one facet of life at Kolomoki However, civic-ceremonial activitiesand their spectacular archaeological remains must be understood in relation to themundane activities of everyday life

Perhaps more important, attempts at categorizing Kolomoki have relied onconcepts that come from an older, linear cultural evolutionary theory (Sahlins andService 1960; Service 1962) Recent critiques have demonstrated that such con-cepts are frequently unable to capture the degrees and kinds of variation present

in the archaeological record (Kristiansen 1991; Yoffee 1993) The past two cades have witnessed a fundamental rethinking of traditional notions of non-strati¤ed societies, highlighting previously unappreciated distinctions in age, gen-der, and prestige (Feinman 1995:261; Flanagan 1989; McGuire 1983; Paynter1989; Paynter and McGuire 1991) Conversely, recent discussions of variations

de-in strati¤ed societies pode-int to the existence of “group-oriented” or “corporate”chiefdoms that lack archaeological evidence for personal ranking (Earle 2001;Renfrew 1974, 2001)

Ethnographers have also joined in the debate concerning the usefulness ofsociocultural types Salzman (2000), for example, demonstrates that the Baluch

of Iran combine a horizontally oriented segmentary lineage system with a ingly centralized political hierarchy Tuzin (2001), on the other hand, argues thatthe Arapesh of New Guinea display all of the characteristics of a chiefdom, exceptthat they lack a chief; power is instead held by an oligarchy of ritual leaders.Clearly, the distinctions between tribe and chiefdom, egalitarian and ranked,and simple and complex have become blurred While sociocultural types remainuseful as heuristic devices and for cross-cultural comparison, they obviously fallshort in their ability to account for the full range of societal variation More im-portant, the adherence to typologies makes it dif¤cult to explain how sociocultu-ral forms change over time (Marcus and Flannery 1996:236; Tuzin 2001:8).The uncritical application of labels such as tribe, chiefdom, proto-chiefdom,

seem-or state to describe Kolomoki may thus obfuscate mseem-ore than explain tives to evolutionary typologies include constructs such as sequential hierarchy,

Alterna-7

heterarchy, horizontal integration, and corporate and network political-economicstrategies (Arnold 1997; Blanton 1998; Blanton et al 1996; Crumley 1987,1995; Feinman 1995; McGuire 1983; Paynter 1989) Concepts such as these may

be necessary for the study of systems in which authority was only weakly oped, unstable, deliberately hidden, or collective as opposed to individual

devel-Kolomoki and the Role of Ceremony in Middle-Range Societies

Confusion regarding the interpretation of Kolomoki stems not only from the adequacies of previous research at the site and an overreliance on unilineal evolu-tionary typologies It is also the product of our failure to understand the role ofceremony during the Woodland period in the Southeast and its role amongmiddle-range societies in general Because ritual often provides the context forexchange (Dietler and Hayden 2001:9), trade ¤gures prominently in this debate.Traditionally, Woodland ritual has been interpreted as an institution thatfunctioned to maintain corporate group identity and manage subsistence risk.Brose (1979b:7), for example, describes ceremony and the attendant exchange ofexotica during the Middle Woodland period as a sort of “intragroup safety valve”that functioned to maintain networks of trade in subsistence goods He arguesthat Woodland burial ceremonies reinforced membership in corporate structuresthat controlled access to subsistence resources Ceremonial trade in luxury itemsserved to maintain alliances that ensured occasional access to otherwise restrictedexternal areas during periods of resource stress

in-Braun has put forth a similar model (in-Braun 1986, 1987; in-Braun and Plog1982) His thesis is that subsistence intensi¤cation early in the Middle Wood-land period resulted in population concentration, which in turn required the de-velopment of new leadership roles for resource management Hopewell valuablescirculated among these leaders until suf¤cient social structure was in place at

ca a.d 400 to ensure a smooth transition of authority

Seeman (1979) constructs a functionalist version of this theory He argues thatthe exchange of exotics was secondary to a more basic trade in subsistence goodsthat functioned to redistribute local food surpluses to areas of scarcity In his view,the increase in ceremony during the Middle Woodland period re®ects a ritualregulation of subsistence resources

These interpretations of Woodland ritual and its associated exchange re®ectthe processual view of culture as an adaptive system Under this paradigm, culturechange is frequently seen as a response to external (primarily environmental) fac-tors, such as population growth and food scarcity The processual view has beenchallenged on theoretical grounds but perhaps most importantly for failing to

grant suf¤cient weight to nonmaterial forces and to particular historical meanings(Conkey and Spector 1984; Hodder 1986; Nassaney and Sassaman 1995).However, the criticism of processual explanation is not limited to theory Theempirical evidence for processualist explanation has frequently been found to belacking Archaeological evidence from a variety of regions and societies suggests

that population growth and technological innovation frequently follow social

change (Bender 1985a, 1985b; Blanton et al 1993; Brum¤eld 1976; Hayden1998; Wright 1984) In regard to the Middle Woodland period in the Mid-west, Wymer (1993) uses paleobotanical evidence to refute the notion that Hope-wellian exchange functioned to counteract intermittent shortages in horticultural

or collected resources She further argues that the important changes that tookplace in the transition from Middle to Late Woodland were not due to subsis-tence

In response to the criticism of processualism, archaeologists have begun to lookbeyond functionally oriented ecological and economic explanations, turning toconsiderations of ideology, power, and factional competition (Brum¤eld 1989,1994; Clark and Blake 1994; Earle 1990; Feinman 1991, 1995; Pauketat 1992,1994; Pauketat and Emerson 1991, 1997a, 1997b) Many of these works empha-size the control of information and the power-based enforcement of meaning inculture (Paynter 1989:376–377; Paynter and McGuire 1991)

Extending this perspective to the Woodland period, Seeman (1995:124) gues that Hopewellian prestige goods must be understood as symbols, with mul-tiple meanings that were validated through mortuary ritual Following Helms(1988), he stresses the importance of esoteric knowledge in the “power quest-ing” by aspirants to leadership positions in Middle Woodland society (Seeman1995:136) Seeman further suggests that “the ultimate demonstration of distantknowledge and personal power would have been the cajoling of ‘dangerous’ for-eigners themselves to return to a major Hopewell ritual center.”

ar-Anderson (1998:287) makes a similar case for Kolomoki and the other majorWoodland centers of the Deep South, arguing that they would have exercisedpower through the maintenance of an elite hegemony He claims that “by beingmajor players in the Hopewellian world, the principals at these centers could havebeen perceived as having esoteric knowledge and this, plus their control overdesirable wealth items, may have inspired people to their service over wide areasand at the same time led to their sancti¤cation.”

If the processual explanations for Woodland trade and ceremonialism are notentirely satisfying because of their emphasis on external forces and the functionalroles of system-level institutions, then these actor-based perspectives are perhapsequally unsatisfying because of their narrow focus on individual aggrandizement

9

Middle Woodland societies are well known for their accouterments of the wellian Interaction Sphere” (Caldwell 1964), yet there is signi¤cant debate aboutthe ability of prestige-goods exchange to structure whole societies, as well aswhether exotic goods mark individual aggrandizement or the status or of¤ce itself(Kowalewski 1996) Wiessner (2002:234–235) adds that these agent-oriented ap-proaches have all too often treated egalitarianism as the product of organizationalsimplicity—a sort of blank slate that aggrandizers can manipulate at will More-over, as Pauketat (2001b:79) notes, the narrow focus on individual goal seekersreduces historical change to a vitalistic equation, wherein “all populations haveaggrandizers who behave in predictable ways that, in turn, induce political com-plexity.”

“Hope-Further, like the processual explanations cited above, these actor-based theoriesare frequently lacking in empirical evidence Anderson’s (1998) assertion, for ex-ample, hinges on a presumed importance of Kolomoki vis-à-vis a trade network

in marine shell However, with the exception of a few burials, marine shell is quiterare at Kolomoki More generally, the apparently limited volume of trade in theMiddle Woodland period argues against any sort of regular and planned inter-regional movement of people and resources (Grif¤n 1979:274)

The argument about the role of ceremony and trade in Woodland society allels a large debate over whether such institutions—and the leaders who con-trolled them—should be construed as “system-serving” or “self-serving” (Blitz1993; Feinman 1995:262–263; Flannery 1972; Rathje and McGuire 1982; Wasonand Baldia 1998) However, as Feinman (1995:262–263) suggests, the recogni-tion of the existence of opposing interpersonal relations eliminates the need topolarize these positions Institutions such as trade and ceremony cannot be viewedsimply as “societal problem-solving mechanisms” (Feinman 1995:263), but nei-ther can the functional roles of such institutions be entirely dismissed As Dietler(2001:77) has argued, ritual may divide and unite at the same time

par-Kolomoki Revisited: Theoretical Orientation and Research Design

This book presents a fresh appraisal of Kolomoki I proceed not by constructingideas from ethnographic, ethnohistoric, or archaeological analogy but instead bydeveloping concepts derived through the identi¤cation of patterns of variation inthe archaeological record itself This approach emphasizes ¤nding and comparingvariation in activity episodes; that is, it attempts to tease apart sites and compo-nents into events that can be related to one another in time The intent is tofashion a history of Kolomoki from its founding around a.d 350 to its eventualdecline and abandonment some time soon after a.d 750

One reviewer of the proposal that ultimately funded this research described

my approach—in both a positive and negative light—as “unabashedly inductive.”However, in pursuing a history of Kolomoki, I am in agreement with others whohave turned away from the general and comparative explanatory paradigm of the1970s in favor of the analysis of culture-speci¤c developmental trajectories (Hod-der 1986; Stein 1998) Pauketat (2001a, 2001b) has referred to this emergingrapprochement between archaeology and history as “historical processualism.”Regardless of the label that is applied, recent attempts to utilize a historical ap-proach to archaeological interpretation share some core concepts Central among

these are the notions of structure and practice, as originally formulated by social

theorists Bourdieu (1977, 1980) and Giddens (1979, 1984) Structure may bede¤ned simply as the “rules and resources that people draw upon in their dailyinteractions” (Varien 1999:25) As this de¤nition implies, structure is similar towhat anthropologists have traditionally referred to as “culture,” but the use of adifferent term re®ects the reformulation of this concept as a far less static entity.Individuals are in®uenced by structures, but they also manipulate structuresthrough practice—perhaps most simply de¤ned as their actions and representa-tions (Pauketat 2001b:74)

The relationship between structure and practice is complex People are strained by their dispositions and perceived options for behavior, as described by

con-Bourdieu’s (1980) concept of habitus Individuals may also be constrained by

tra-dition, which is broadly de¤ned as practice brought from the past into the present

(Pauketat 2001a:2) However, individuals consciously develop strategies for nipulating traditions and structures, and in the process they both reproduce andtransform them (Varien 1999:26) Thus, as Pauketat (2001a:4–6, 2001b:80)notes, traditions can be both constraints on and mediums for social change Inaddition, because individuals and social groups de¤ne and interpret their relation-ships to traditions differently, the meanings of traditions are plural and contested(Pauketat 2001a:12–13, 2001b:80; Varien 1999:26)

ma-More thorough treatments of this line of social theory may be found in theworks of Bourdieu (1977, 1980) and Giddens (1979, 1984), as well as in thesynthetic reviews of these sources cited above My intention in this study is not

to burden the reader with a disquisition on social theory However, I believe that

a practice-oriented, historically based paradigm is appropriate for the tion of Kolomoki The activity areas and sites that are a focus of this study re®ectthe practice of individuals and groups of individuals (be they households, kingroups, secret societies, or communities at large) Where these practices persistedlong enough to be visible archaeologically as community and regional settlementpatterns, they re®ect traditions and structures The search for patterned variation

interpreta-11

in these archaeological phenomena across space and through time offers a means

of de¤ning, analyzing, and interpreting these traditions (Lightfoot 2001:247),thus enabling me to document the historical development of Kolomoki

It should be clear that historical processualism does not refer to a simple rative of events (Pauketat 2001a:4), which can rarely be reconstructed archaeo-logically Nor can it be based on the documentary evidence we ordinarily associatewith history, such as letters and journals As a result, historians accustomed to theconstruction of a narrative of events through the study of documents may bepuzzled by my characterizing this work as a sort of history However, this ap-proach is consistent with the broader view of history espoused by social historianssuch as Marc Bloch, who observed that the variety of historical evidence is nearlyin¤nite and that the path of historical inquiry “may turn at will toward either theindividual or the social, toward momentary convulsions or the most lasting de-velopments” (Bloch 1953:20) As all archaeologists know, the archaeological rec-ord has little to say about “momentary convulsions” but a great deal to say about

nar-“lasting developments,” and each of these is as much a part of history as the other.The understanding of complex societies as geographically and historicallyunique entities leads naturally to a focus on proximate, rather than external,causes for cultural change and continuity (Pauketat 2001b; Stein 1998:2) Thus,

the focus of explanation shifts from why we observe similarities in structures or traditions between societies to how structures or traditions developed in particular

historical contexts (Pauketat 2001b:87)

This shift in focus avoids the generalizations of an earlier, normative culturehistory of the kind that branded Kolomoki as Mississippian It also eschews theevolutionary approaches that would force Kolomoki into one or another narrowlyprescribed type By avoiding the appeal to synchronic types, I will demonstratehow sociocultural formations developed and changed over the course of the site’shistory Was Kolomoki a tribe or chiefdom? As Salzman (2000:3) responds to aparallel inquiry in his ethnography of the Baluch, the ¤rst answer is that thequestion is too simplistic

Finally, the focus on proximate causation avoids the extremes of both thosewho would seek the impetus for cultural change at the level of the system andthose who would place the onus on an inherent tendency toward individual ag-grandizement Was the ceremony and prestige-goods exchange that is evident atKolomoki and related Woodland period sites “system-serving” or “self-serving”?

If we recognize the existence of opposing interpersonal relations and contestedinterpretations of traditions, this question is likewise overly simplistic

The emphasis on the speci¤c historical trajectory of Kolomoki does not meanthat I am uninterested in comparing the site to others in the region or its relation

to larger trends that have been noted in the archaeology of the Woodland period.Broad similarities in material culture developed across the eastern United Statesduring the period, and the question of why this should be the case begs explana-tion However, I follow Pauketat (2001b:87) in his assertion that the answers

to such ultimate questions will only be found through “painstaking, data-rich,multiscalar studies of proximate causation.” My own research, in combinationwith the work conducted by other archaeologists at Kolomoki through the years,approaches this scale of investigation, and I thus feel justi¤ed in addressing some

of the broader questions alluded to above Nevertheless, the emphasis here will be

in ¤rst documenting how, rather than why, Kolomoki developed as it did

The Structure of the Book

The research presented herein is fundamentally a settlement pattern study ducted at overlapping scales of analysis At the broadest level, it focuses onchanges within the region From here, the focus narrows progressively to the com-munity, activity area, and household The organization of the book re®ects thisresearch strategy

con-I begin at a broad scale of analysis in Chapter 2 by placing Kolomoki withinits temporal, ecological, and cultural context I brie®y review the established ce-ramic chronologies for the lower Chattahoochee Valley before examining the evi-dence for ¤ner temporal divisions at Kolomoki Next comes a review of the envi-ronment of the region and an examination of ecologically based models forKolomoki’s location Finally, I summarize previous archaeological research andsettlement pattern data for the area within a 200-km radius of the site Are therecompelling ecological explanations for Kolomoki’s location? How does Kolomokirelate to contemporaneous sites in its hinterlands? Do the settlement data suggestthat it was at the center of a corporate social group or that it was a shared node

in a wider social network?

In Chapter 3, the focus narrows to the levels of community and activity area

I present a more thorough review of previous research at Kolomoki and thenexamine the archaeology of the major earthworks Also presented is an evaluation

of previous work in the off-mound portions of the site, which allows for the mulation of some initial assumptions regarding activity areas My reanalysis ishindered by the fact that most of the documentation of Sears’s work at the sitehas been lost.3 However, I have reexamined the artifact collections from many

for-of the previous studies at Kolomoki, including more than 60,000 sherds Thisreanalysis, in combination with a rereading of the published descriptions of thework, allows me to address a number of basic questions regarding the settlement

13

of the site When were the mounds constructed and for what purposes? What isthe signi¤cance of their alignments?

Chapter 4 begins a more rigorous analysis of community patterns and activityareas Here, the focus is on the ¤rst phase of my own research at Kolomoki, whichconsisted of intensive sampling of the off-mound portions of the site Over thecourse of three ¤eld seasons, my colleagues and I have taken more than 1,300samples, resulting in a database that allows me to answer basic questions regardingthe internal structure of Kolomoki What are the limits of the site and how areits temporal components distributed? Is there evidence for ceremonial feasting orother ritual activities on discrete areas on the site? Do differences in artifact dis-tribution support the interpretation of status differentiation among households?

In the next phase of my research, 46 m2 of test excavations were undertaken

in selected activity areas These test excavations, which are summarized in ter 5, permitted a closer look at several activity areas to test the initial impressionsthat were formed on the basis of the sampling Activity areas were further char-acterized through a program of geophysical prospection These investigations alsohelped me pinpoint a domestic area for more intensive investigation

Chap-In Chapter 6, the focus narrows to a few presumed domestic activity areas.Here, I describe the recent excavation of a household at Kolomoki The focus ondomestic activities re®ects the need to answer basic questions about the site, spe-ci¤cally those concerning the permanence of occupation, subsistence strategy,and population growth Was Kolomoki occupied year-round, seasonally, or onlyfor short-term stays related to mortuary ceremonies? What was the subsistencestrategy, and does the evidence suggest the possibility of subsistence risk or, alter-natively, relative self-suf¤ciency in production?

Finally, in Chapter 7, I synthesize the research at Kolomoki Consistent withthe research design described in this chapter, the synthesis is presented as a phase-by-phase narrative The discussion takes the form of a “braided narrative” (Fischer1976; Hudson 1995:xvi) that weaves the archaeological data with my own inter-pretation In presenting the development of Kolomoki as a historical process, Ialso return to some of the themes that were raised in this chapter

Before considering my work and that of others at Kolomoki, it is important toplace the site in its proper context This chapter begins with a revised view of thetemporal placement of the site I then turn to a summary of the environmentaland cultural settings.

Temporal Context

As was noted in C hapter 1, it is now commonly accepted that the dominantoccupation at Kolomoki dates to the Middle and Late Woodland periods How-ever, some confusion remains regarding the precise temporal position of the site.Further, although there have been some attempts to re¤ne the chronological se-quence for the Middle and Late Woodland periods in the lower ChattahoocheeValley ( Jenkins 1978; Knight and Mistovich 1984; Mistovich and Knight 1986;Schnell 1998), these have focused mainly on areas well removed from Kolomoki,and the chronological divisions are fairly general, being on the order of a fewcenturies

Here, I present a revised internal chronology for Kolomoki Because the vious collections from the site were not gathered systematically, and because there

pre-is a possibility that some diagnostic types were pulled from the collections, thechronology is based primarily on recent test excavations I do, however, look toearlier collections from the site for corroboration of the chronology The testunits, as well as their respective artifact assemblages and the methods that wereemployed during excavation and analysis, are described in more detail in Chapter

5 The focus of this discussion is on two classes of temporally diagnostic artifactsfrom test excavations: decorated ceramics and formal projectile points/knives

2 Putting Kolomoki in Its Place

De¤ning the Temporal, Ecological, and

Cultural Contexts

Seriation of CeramicsBefore reviewing the ceramic assemblage from test units, some general observa-tions regarding the complete assemblage from the site are in order I have thus farcounted and weighed more than 100,000 sherds from Kolomoki This total in-cludes about 60,000 sherds from previous work at the site and more than 40,000from my own investigations Although a large fraction of the collection is toosmall or eroded to identify with certainty, only about 500 sherds—or less thanone percent of the overall assemblage—bear surface decoration or temper consis-tent with Mississippian types The remaining identi¤able sherds are clearly more

in keeping with Woodland ceramic types

My test unit excavations, described in more detail in Chapter 5, resulted in therecovery of slightly more than 11,000 sherds For the purposes of de¤ning ce-ramic phases for the Woodland period occupation, I limit the seriation to thosesherds that are consistent with Woodland period types on the basis of temper andsurface decoration (only two sherds, or less than one-tenth of one percent of theassemblage, can de¤nitely be identi¤ed as Middle or Late Mississippian Lamartypes) I further limit consideration to the assemblages from the 17 test units thatwere systematically screened Also excluded are six units that produced ceramiccollections that are either very small or show obvious signs of reoccupation Fi-nally, the seriation is limited to sherds that are relatively large (greater than around

2 cm) and well preserved These restrictions reduce the sample for seriation byabout one-half to 4,265 sherds This is a relatively small sample, and thus theseriation that follows must be considered tentative However, the proposed chro-nology builds on previous work in the region, which lends an additional measure

of con¤dence

Seriation of the ceramic assemblage from Kolomoki must be understoodwithin the context of previous work in the region, which has established the gen-eral sequence of Woodland pottery types in southwestern Georgia, southeasternAlabama, and northwestern Florida Broadly speaking, Deptford wares are theearliest in the region, followed by the Swift Creek type, which in turn is graduallyreplaced by the Weeden Island series (Caldwell 1978; Knight and Mistovich1984; Schnell 1998; Willey 1949) Knight and his colleagues have de¤ned thissequence for the Chattahoochee Valley (Knight and Mistovich 1984; Mistovichand Knight 1986; Schnell 1998; Schnell and Knight 1978) They describe theinitial Woodland period occupation in the region, marked by Deptford CheckStamped and Dunlap Fabric Impressed pottery, as the Shorter phase (1000 to 300b.c.) This is followed by the Mandeville I phase (300 b.c to a.d 1), characterized

by assemblages with a mixture of Deptford Check Stamped and Deptford SimpleStamped wares and Swift Creek Complicated Stamped Subtle changes in theproportions of these types mark the Mandeville II phase (a.d 1 to 300).The Kolomoki phase (a.d 350 to 500) is de¤ned by the decline of Deptfordtypes and a predominance of Swift Creek As several authors have noted, there is

a stylistic gap between Mandeville II and Kolomoki, for which several transitionalassemblages have been noted (Mistovich and Knight 1986; Snow and Stephenson1992:48) Of ¤nal relevance to Kolomoki is the Quartermaster phase, broadlydated to a.d 500 to 750 and characterized by the mixture of Swift Creek andWeeden Island wares

As the work by Sears (1956) and myself suggests, Deptford ceramics appear inonly very modest proportions at Kolomoki This indicates that the earliest Wood-land period occupations at the site should be marked by assemblages with SwiftCreek pottery and relatively few, if any, Weeden Island series ceramics The ce-ramic collections from three of my test units (Units 2, 3, and 6) clearly ¤t thisdescription (Table 2.1) These assemblages all contain relatively high (>40) per-centages of Swift Creek Complicated Stamped pottery Judging from these threeunits, the proportion of the Weeden Island types is less than one percent com-bined during this interval Two of these units also produced a few sherds ofBlakely Complicated Stamped, a variation on the Swift Creek type with wider (>4mm) lands and grooves Although Sears (1956:17) believed that this type datedlate in the sequence, recent research indicates that it is a good marker for theearliest Woodland occupation at Kolomoki (Pluckhahn 1998b:112–115; Snowand Stephenson 1992:44), as the correction of Sears’s inverted chronology wouldsuggest

My hypothesis that there is an early horizon at Kolomoki marked by a derance of Swift Creek ceramics, with few or no Weeden Island types, is corrobo-rated by several larger collections from the site Both Sears’s collection from themidden below Mound D (described in the following chapter) and my assemblagefrom Block A (described in Chapter 6) contain similar proportions of these diag-nostic types

prepon-At the presumed late end of the Woodland period occupation at Kolomoki arethree test unit assemblages (Units 7, 8, and 18) that contain reduced proportions

of Swift Creek Complicated Stamped pottery (<25 percent) and a wide variety ofWeeden Island types In addition to Carrabelle Incised and Carrabelle Punctate,these three units also contain small amounts of Indian Pass Incised, Weeden Is-land Incised, and Weeden Island Red (including Weeden Island Zoned Red).Moreover, these units produced a few sherds of two pottery types, Napier Com-

17

plicated Stamped and Wakulla Check Stamped, that are securely associated withthe Late Woodland period (Caldwell 1955b, 1978; Kelly and Smith 1975; Wau-chope 1966; Willey 1949).

With the early and late phases of occupation established, we can next examinethe remaining test unit assemblages for evidence of transition between these twophases Test Units 10 and 15, like the early assemblages from Units 2, 3, and 6,contain high (>40 percent) proportions of Swift Creek Complicated Stamped.However, unlike the earlier collections, these also contain slightly higher relativefrequencies (one to two percent combined) of a restricted number of WeedenIsland types, including Carrabelle Incised, Carrabelle Punctate, Weeden IslandIncised/Punctate, and Weeden Island Red These two units would appear to epito-mize one intermediary stage between the earliest and latest phases at Kolomoki.Another intermediary stage may be signaled by the assemblages from three testunits (Units 1, 9, and 14) that exhibit a reduced proportion of Swift Creek Com-plicated Stamped (25–40 percent) and an increase in the relative frequencies ofthe above-mentioned Weeden Island types (5–15 percent combined) In addition,the assemblages from these units sometimes include a few sherds of other WeedenIsland types, including Tucker Ridge Pinched and Ruskin Dentate Stamped.Dividing the presumed occupation span at Kolomoki (a.d 350 to 750) toaccount for these divisions produces a series of four 100-year phases For ease ofdiscussion, I have chosen to refer to these as Kolomoki I through IV Table 2.2summarizes these ceramic phases

Carbon dates from several contexts support my chronology (Table 2.3) Asample from Unit 3, which produced a Kolomoki I assemblage, yielded a date of

1660 ± 50 b.p., calibrated at 1 sigma to between a.d 350 and 430 (Beta-121909).This range corresponds quite closely with my estimate of a.d 350 to 450 for theKolomoki I phase At the opposite end of my chronology, a sample from Test Unit

18, which I have dated to the Kolomoki IV phase, produced a radiocarbon age

of 1290 ± 60 b.p., with a 1-sigma calibrated range of a.d 670 to 780 161790) This accords well with the estimated range for the Kolomoki IV phase,from a.d 650 to 750

(Beta-This analysis omits several ceramic types de¤ned by Sears (1950, 1951a,1951b, 1953a, 1956), perhaps most obviously his Kolomoki Complicated Stampedand Little Kolomoki Complicated Stamped types Although some researchershave continued to employ these types (e.g., Johnson 1997; Knight and Mistovich1984; Milanich et al 1997), others suggest that they are simply varieties of theSwift Creek type ( Jenkins 1978; Smith 1977; Snow and Stephenson 1992:112)

My previous analysis of sherds from the site supported the latter conclusion(Pluckhahn 1998b:113–114) Additional evidence is provided by paddle matches

19