AN ARCHAEOLOGICAL INVESTIGATION OF
PRECOLUMBIAN POLITICAL STRUCTURE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
For my mother,
Marjorie Hopkins Payne
and in memory of my father,
Richard McDonald Payne
Many people contributed to this study of Mississippian capitals. At the top of
the list is Dr. Jerald Milanich, my dissertation advisor, who provided assistance and
advice in innumerable ways. Among other things, he arranged funding for several
aspects of my investigations at the Lake Jackson site. In addition, he employed me on
several museum projects and recommended me for others, thus making it possible for
me to continue my research and writing. And he worked wonders when university
red tape occasionally snarled. I especially appreciate his unflagging patience as this
project threatened to go on forever. It has been a pleasure and an honor to work with
I am also grateful to the other members of my committee, Drs. William
Keegan, William Marquardt, Michael Moseley, and Graig Shaak, who were unfailingly
supportive and helpful throughout the dissertation process. Two earlier committee
members, Drs. James Henry and Prudence Rice, whose careers took them away from
the University of Florida before I finished, provided valuable advice and direction in
the early stages of research. I also thank Dr. Kathleen Deagan who graciously
stepped in to fill a temporary committee gap.
Although not a committee member, Dr. John Scarry of the University of North
Carolina did much of the work of one. He commented on drafts, answered
questions, acted as a sounding board, gave permission to use two of his own maps,
and shared his wealth of knowledge about the workings of Mississippian chiefdoms.
Many of my ideas about Mississippian society have been formed in conversations
with John over the last 15 years. As always, his generosity of spirit and action has
My field work at the Lake Jackson site was turned into reality largely through
the efforts of three people. Jerry Milanich found funding for a test excavation in
Mound 5, for analysis of some of the material collected, and for radiocarbon dates.
Dr. Rochelle Marrinan of the Florida State University generously "loaned" me her field
school for a month to carry out an auger survey at Lake Jackson. Moreover, during
that time, she managed all the thankless logistical tasks superbly, freeing me to
concentrate on the site. Later that year, she volunteered her personal time during the
excavation of Mound 4. Throughout the project, she offered encouragement and
assistance. Robert Morley, former manager of the Lake Jackson State Archaeological
Site, provided information about the modern history of the site and took on many
non-research tasks such as repairing equipment and backfilling units so that we
archaeologists could devote more time to our investigations. He even, in his off-time,
lent a hand in the excavations. I am truly grateful for Bob's interest in the
archaeology of the site and his dedication to the site's preservation.
Others also facilitated the investigations at Lake Jackson and my study of
earlier collections from the site. Jim Miller, John Scarry, and Louis Tesar of the
Bureau of Archaeological Research arranged permission to excavate on state-owned
land and facilitated the loan of collections from earlier excavations at the site. John
Scafidi of the Department of Natural Resources, which manages the site, helped with
logistical arrangements. Phil Werndli, also of DNR, marshalled volunteers for the test
excavation in Mound 4 and managed the scheduling of the volunteers. I am also
indebted to Bonnie McEwan and Charlie Poe of the San Luis Archaeological and
Historic Site for additional logistical support.
Over the course of the auger survey and the test excavations in Mounds 4 and
5, a host of volunteers and students contributed their time and labor. Participants in
the 1989 field school were teaching assistants Frank Keel, Tom Kempton, and Debbie
Leslie, graduate students Tim Barton, Bridget Beers, and Rebecca Emmans,
undergraduates George Betts, Eve Kappler, Sheridan Laymon, Darren Milman, Joe
Reinhardt, Susan Stamp, and Claudia Tomlinson. Mound 4 volunteers included
Charlie Arant, Tim Barton, Kevin Butler, Aaron Cohen, Gabe Cohen, Matt Cohen,
Walt Crabbe, Melvin Goffinet, Eve Kappler, Frank Keel, Tom Kempton, Debbie Leslie,
Rochelle Marrinan, Bob Morley, Morrie Naggar, Candy and Eric Nelson, Guy Prentice,
Bob Rubanowice, Clara Scarry, John Scarry, Marion Smith, Peter van Tassel, Stephen
Werndli, Steve Wright, and John Worth. Jim Cusick, Jason Jackson, Bob Morley, and
Mark Stevenson contributed their efforts to the Mound 5 excavation.
Back in Gainesville, several loyal volunteers helped me process the artifacts
from these and earlier investigations. My thanks go to Jean Allan, Barbara Hendry,
Jason Jackson, and Marjorie Payne for assisting with that dirty and tedious, but
essential, task. Ann Cordell, Ken Johnson, Elise LeCompte-Baer, Donna Ruhl, John
Scarry, and John Worth provided technical advice and identified unusual artifacts.
Marc Frank and Lee Newsom identified faunal and floral remains from my
investigations at the site.
A number of people provided financial and logistical assistance while 1
worked on this study. Jerry Milanich and Bill Marquardt of the Florida Museum of
Natural History both employed me at various times, as did the Department of
Anthropology. A predoctoral Graduate School fellowship and an "intra-family
fellowship" made it possible for me to devote large blocks of time exclusively to
research and writing, without which I would still be struggling along. For these
opportunities, I thank former Dean of the Graduate School Madelyn Lockhart and
Marjorie H. Payne. The Charles H. Fairbanks Award also offset some of the costs of
the research. Bill Marquardt graciously gave me access to his Southwest Florida
Project computers on which this dissertation and some of its maps were completed.
Over the years, I have profited from many discussions and friendly debates
with my colleagues at the University of Florida and elsewhere. Jean Allan, Nina
Borremans, Charlie Ewen, John Griffin, Bill Johnson, Ken Johnson, Calvin Jones, Jon
Leader, Rochelle Marrinan, Bill Marquardt, Bonnie McEwan, Jerry Milanich, Jeff
Mitchem, Guy Prentice, Donna Ruhl, John Scarry, Margaret Scarry, Gary Shapiro,
Marion Smith, Ruth Troccoli, and John Worth have all, at various times, shared their
thoughts on political anthropology, the Lake Jackson site, precolumbian and colonial
Apalachee, Mississippian archaeology, or Florida archaeology, and thereby
contributed to my thinking. Particularly enlightening and enjoyable have been
recurring discussions about chiefdoms with Jean Allan, Bill Marquardt, and John
The road to finishing a dissertation is a long one, and the emotional support of
friends and relatives is invaluable. In addition to the friends and colleagues noted
above, I am grateful to Nawal Ammar, Tina Bassett, Chuck Blanchard, Ann Cordell,
Susan de France, Robin Denson, Tom Eubanks, Jean Gearing, Barbara Hendry, Leslie
Johnson, Elise LeCompte-Baer, Darcie MacMahon, Emine Incirlioglu, Elizabeth Schmitt
Page, and Richard and Ann Payne for their friendship and encouragement over the
years. My dear friends Nina Borremans, Barbara Hendry, Elise LeCompte-Baer, Bill
Marquardt, Donna Ruhl, and John and Margie Scarry deserve special mention for
encouraging me in innumerable ways, from loaning me a "magic talisman" (it
worked!) and picking up the lunch tab when 1 was writing instead of earning a living
to providing the psychological support that only friends can and throwing
celebrations when I finished at last. And finally, Marjorie Payne has gone beyond the
call of devoted motherhood by providing money and emotional support, washing
artifacts, reading drafts, and patiently listening to long monologues on Mississippian
esoterica. It is to her and to my late father, Richard McDonald Payne (who would,
I'm sure, have delighted in calling me "Doc") that I dedicate this dissertation.
TABLE OF CONTENTS
ACKNOW LEDGMENTS .................................... iv
A BSTRA CT .............................................. xi
1 INTRODUCTION .............................. 1
Chiefdom s ................................... 2
Power and Authority in Chiefdoms ................ 10
C apitals ..................................... 17
2 CAPITALS IN CHIEFDOMS .................. .... 20
Capitals in the Ethnographic and Historic Record ..... 21
Size of Capitals ................................ 29
Structure of Capitals ............................ 34
Sum m ary .................................... 67
3 MISSISSIPPIAN CAPITAL VILLAGES, PART 1:
THE SIZE OF THE CAPITAL ............... 70
Identifying Mississippian Capitals ................. 70
Mississippian Mound Center Survey ............... 75
Measuring the Size of Sites ....................... 78
Number of Mounds ............................ 79
M ound Precinct Area ........................... 89
Height of Main Mound .......................... 97
Volume Index of Main Mound .................... 107
Sum m ary .................................... 114
4 MISSISSIPPIAN CAPITAL VILLAGES, PART 2:
THE STRUCTURE OF THE CAPITAL ......... 130
Layout ...................................... 130
O orientation ................................... 140
Chiefs H house ................................. 147
Chiefs Storage Facilities ......................... 178
Religious Structures ............................ 186
Communal Structures ........................... 200
Public Spaces ................................. 207
W alls ....................................... 211
Sum m ary .................................... 225
5 THE LAKE JACKSON SITE: PORTRAIT OF A
MISSISSIPPIAN CAPITAL VILLAGE ......... 229
The Site and Its Setting .......................... 232
History of the Site .............................. 235
Archaeological Investigations at Lake Jackson ........ 242
The Structure and Development of the Lake
Jackson Site ............................. 260
Lake Jackson as a Mississippian Capital Village ....... 272
Sum m ary .................................... 284
6 MISSISSIPPIAN CAPITALS AND THE ANALYSIS OF
MISSISSIPPIAN POLITICAL STRUCTURE ..... 288
Political Structure and the Built Environment ......... 288
Power, Authority, and Mississippian
Political Structure ........................ 290
Precolumbian Apalachee: The View from
Lake Jackson ............................ 295
Studying Capitals to Analyze Political Structure ....... 316
REFEREN CES ............................................ 319
BIOGRAPHICAL SKETCH .................................. 346
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fullfillment of the
Requirements for the Degree of Doctor of Philosophy
AN ARCHAEOLOGICAL INVESTIGATION OF
PRECOLUMBIAN POLITICAL STRUCTURE
Chair: Jerald T. Milanich
Major Department: Anthropology
Examining the built environment of political capitals in chiefdoms offers
archaeologists one means to explore the dynamics of chiefly political structure. The
investigation of political capitals provides useful insight into the nature of political
structure because a capital's shape and size reflect various aspects of the leader's
power (the ability to control others' actions) and authority (legitimized power). This
is primarily because the interaction between leaders and populace requires that
leaders advertise their access to power and authority in order to maintain them. One
highly effective means of advertisement is through architecture, especially the
architecture of the capital.
Cross-cultural data from 30 chiefdom areas worldwide demonstrate that the
built environment of chiefly capitals reflects aspects of political structure.
Characteristics relating to size and number are shown to express the chief's power
(e.g., the ability of the chief to mobilize labor). Characteristics relating to location
(horizontal or vertical), form, and decoration convey information about the chief's
The Mississippian world (the southeastern and midwestern United States
between about A.D. 900 and 1600) offers a venue for the study of chiefly capitals in
the archaeological record. Data compiled from 536 Mississippian mound sites reveal
the wealth of archaeological information available on the built environment of
Mississippian capitals. This abundance of data makes possible broad comparative
analyses of many architectural features of Mississippian capitals, including layout,
orientation, platform mounds, chiefs' houses, religious structures, earthworks, and
A closer look at one Mississippian capital provides detail on political dynamics
in an individual Mississippian chiefdom. The Lake Jackson site in northwest Florida
was the capital of the Mississippian chiefdom of Apalachee. Based on the results of
archaeological investigations at the site, a ceramic chronology and an analysis of the
structure and development of the site is presented. The considerable data from Lake
Jackson make possible the analysis of transformations in power and authority both
within the larger context of the Mississippian world and in the narrower world of
In the American Southeast between about A.D. 900 and 1600, a group of
aboriginal societies developed that were roughly similar in configuration. Called
Mississippian by archaeologists, these societies comprised a network of independent
polities set in largely similar environments and linked by social and political alliances,
trade, ritual, and a political ideal (Scarry 1993).
Mississippian societies shared (1) an increased population, (2) a dependence on
maize agriculture accompanied by exploitation of land and aquatic fauna, (3) an
extensive exchange network involving marine shell, copper, and other exotic artifacts,
(4) a worldview/belief system which varied geographically in detail but not in
outline, and (5) a ranked (or chiefdom) form of sociopolitical organization (Peebles
and Kus 1977; Griffin 1985).
In the Mississippian world there was diversity as well as similarity (Steponaitis
1986:387-393; Smith 1986:57-63). Individual polities often varied greatly in size and
complexity. Not infrequently, polities deviated in some respect from the overarching
ideal. Those on the periphery traded and allied themselves with foreign polities. In
some places, local ideals overshadowed the regional.
Despite the differences, a fundamental identity marked the polities within the
network and distinguished them from outside polities. This identity is recognizable in
material remains throughout the region. Religious objects, for example, though
varying in detail, clearly represent a shared cosmology (Brown 1985; Knight 1986;
Muller 1989:25). And, particularly appropriate to this study, the shapes of the
political capitals of the polities share the same basic configuration (again, despite
variations in detail).
This study focuses on the examination of political capitals as one means to
understand Mississippian political structure. The investigation of political capitals
provides a useful insight into political structure because a capital's shape and size
reflect various aspects of power and authority. Consequently, this dissertation
explores the nature of chiefly capitals (Chapter 2); the shape of Mississippian capitals,
from the size and layout of the sites to the form and nature of their architectural
constructions and spaces (Chapters 3 and 4); and the characteristics of a typical
Mississippian capital, the Lake Jackson mound group near Tallahassee, Florida
(Chapter 5); with concluding comments on the utility of studying political structure by
analyzing the built environment of the capital (Chapter 6).
Throughout the course of this dissertation, several terms are used which are
basic to the discussion. These terms-chiefdom, power and authority, and capital-all
contain some ambiguity. Because of the ambiguity and the importance of these
concepts to the discussion here, they merit some explication. The following three
sections thus provide background information and define my use of these three
Elman Service presented an early and highly influential definition of chiefdom.
For Service, "chiefdoms are redistributional societies with a permanent central agency of
coordination" (1971:134, emphasis in original); moreover, chiefdoms are characterized
by the "pervasive inequality of persons and groups in the society" (1971:145). In
recent years, the role of redistribution of subsistence products across regions in
maintaining chiefly societies (a central concept in Service's formulation) has been
reconsidered by Timothy Earle (1977) and found to be of less import than Service
believed. Nevertheless, the recognition of centralization as essential in the definition
of chiefdoms is an important one, as is that of society-wide inequality.
Service's concept of chiefdom (part of a wider scheme classifying human
societies in terms of evolutionary stages of organizational complexity) generated a
flood of commentary which continues to this day. Researchers soon produced
additional definitions of chiefdoms (Renfrew 1973; Carneiro 1981; Wright 1984;
Creamer and Haas 1985). Some investigators attempted to subdivide or refine the
category (Renfrew 1973, 1974; Steponaitis 1978; Carneiro 1981). Others discussed the
place of chiefdoms in evolutionary schemes (Webb 1973; Sanders and Webster 1978;
Carneiro 1981). Scholars explored the dynamics of chiefdoms (Friedman 1975;
Petersen 1982; Wright 1984; Anderson 1990) or broke down the variables of political
complexity for research (McGuire 1983; Feinman and Neitzel 1984; M.E. Smith 1985).
While it is impractical to review all the research published on chiefdoms in the
last thirty years, it is useful to consider several works that bear most directly on the
research presented in this dissertation. First are those defining the chiefdom.
Robert Carneiro's (1981:45) definition, which he calls a structural one because
of its emphasis on political forms, focuses on the supralocal nature of chiefdoms. "A
chiefdom," he says, "is an autonomous political unit comprising a number of villages
or communities under the permanent control of a paramount chief." This definition
serves to distinguish chiefdoms from egalitarian societies, but not, if the term "king"
were substituted for "paramount chief," from states.
Henry Wright (1977, 1984) defines chiefdoms largely on the basis of their
decision-making capabilities. In a chiefdom, therefore,
central decision-making activity is differentiated from, although it
ultimately regulates, decision-making regarding local production and
local social process, but is not itself internally differentiated. It is thus
externally but not internally specialized (1977:381)
In chiefdoms, then, an institutionalized and centralized political office exists but is
generalized in its functioning. Each level of political administration is similar in
structure. Subordinate political units simply replicate higher level political units at a
Like Wright, Ronald Cohen notes the generalized nature of political control
(1978:4-5) and adds to his definition of chiefdoms the regular occurrence of fission
(1981). This contrasts with circumscription theories of the development of
centralization (e.g., Carneiro 1971, B. Smith 1978, 1985) which postulate that the
inability to fission in times of societal conflict leads to the rise of chiefs. Support for
Cohen's contention comes from ethnographic descriptions of frequent fission in
chiefdoms (see Mitchell 1956, Petersen 1982, and Leach 1965 for examples).
There are certain commonalities in these definitions. Most important is the
centralization of decision making. Related to the concept of centralization but not
always made explicit is the supralocal aspect: chiefdoms contain at least one decision-
making level above the local level (see also Wright and Johnson 1975). A third theme
is the generalized nature of leadership. The subject of the definition of the chiefdom
will be returned to shortly.
Early on, anthropologists recognized that the category of chiefdom contained a
tremendous amount of variability. In population, for example, chiefdoms varied from
a few hundred people (e.g., the Mandari of east Africa) to tens of thousands or more
(e.g., Tahiti in the south Pacific). This recognition prompted subdivisions of the
category based on size or structure (e.g., Julian Steward and Louis Faron's [1959:177]
theocratic vs. militaristic chiefdoms or Colin Renfrew's [1974:74] group-oriented vs.
In the present study, divisions based on size or levels of development are
pertinent. Vincas Steponaitis (1978:420) described chiefdoms as simple and complex:
simple chiefdoms have one level of political hierarchy, while complex chiefdoms have
two or three. Carneiro (1981:47) found this two-part division too confining and not
representative of the range of variation in chiefdoms. He proposed a three-part
division: minimal, typical, and maximal chiefdoms. For Carneiro,
a minimal chiefdom is one that meets the minimal requirements of a
chiefdom ... but does not go far beyond them. A typical chiefdom is
one that is clearly a chiefdom, with elaborations in many aspects of its
political and social structure, but still well below the level of a state. A
maximal chiefdom is one that has become large and complex enough to
approach the threshold of the state" (Carneiro 1981:47, emphasis in
Critiques of the Chiefdom Concept
A number of criticisms have been lodged against Service's evolutionary stage
scheme since it was first put forth. Some of these criticisms are reviewed briefly in
this section, and responses to them are considered in the next.
Some researchers (e.g., Lewis 1968; Kehoe 1980) are loath to lump societies into
"pigeonholes," arguing with some justice that categorization obscures the great
variability and dynamic nature of human groups. The argument goes as follows.
Typologies, by their nature, consist of static categories (Lewis 1968:101). To divide up
what is essentially a continuum, arbitrary divisions must be made. The resulting
categories are biased at best, obscurationist at worst.
Service has also come in for criticism for the way he defined two of the
original categories in his evolutionary scheme. Herbert Lewis and Morton Fried have
both taken umbrage at the use of the term "tribe," Fried going so far as to designate it
the second "most egregious case of meaninglessness" in the language of anthropology
(Fried 1968:4-5). Fried's strongest criticism is that tribes
may well be the product of processes stimulated by the appearance of
relatively highly organized societies amidst other societies which are
organized much more simply. If this can be demonstrated, tribalism
can be viewed as a reaction to the formation of complex political
structure rather than a necessary preliminary stage in its evolution
Lewis suggests that the category of tribe is too general and too vague to be of use
(1968:101-104). Conversely, Lewis argues that chiefdomm" is too specific, based on
societies in a single geographic area (Polynesia) and not applicable worldwide
(1968:104-105). For both authors the categories have no relation to evolutionary
reality but are constructs based on (1) in the case of "tribe," "an ethnographer's (or
administrator's) abstraction" (Lewis 1968:102) or a modem reaction to colonialism
(Fried 1968:17-18) and (2) in the case of chiefdomm," "the idea that a system with some
principle of hierarchy and overall integration should have succeeded the multi-group
segmentary tribes" (Lewis 1968:105).
Another problem with Service's scheme is the notion of progression inherent in
his use of the term "evolution" (Claessen and van de Velde 1985:5-6). Service implies
that human societies move through increasing levels of complexity and end up at the
highest possible level (Service 1971:5). This view contradicts both the scientific view
of evolution (one which excludes directionality) and social reality (in which polities
decline as well as develop) (Claessen and van de Velde 1985:6). A byproduct of
directionality is the tendency to see the whole process as a unilineal one--bands
become tribes, tribes become chiefdoms, chiefdoms become states-again ignoring the
reality of multiple forms of change (Claessen and van de Velde 1985:11).
Estellie Smith (1985) points out some practical objections to using Service's
scheme. One is the reification of the stages. She finds that "more attention is being
directed toward fitting the data to the model than to understanding what those data
can indicate about the dynamics of social organization" (1985:97). Smith's second
objection to Service's scheme is the difficulty of distinguishing chiefdoms from states
(1985:97). Cohen's (1981:92) statement that "Ankole was included in Claessen and
Skalnik's (1978) work as a state... I see it as a chieftaincy" is merely one example of
A Defense of the Concept of Chiefdom
Some of these criticisms and objections to Service's formulation are reasonable
and merit a response from any researcher using the scheme. Other criticisms do not
hold up under scrutiny.
A reluctance to "pigeonhole" societies is understandable, given the wide
variability in human societies, but this reluctance results in a loss of understanding
and largely prevents cross-cultural comparisons of societies as systems. Classification,
if approached cautiously, provides (1) a means of creating order out of chaos and (2) a
shorthand of terms understandable by large numbers of researchers. No classification
will account for all variability, nor should we expect it to. The goal of classification is
not to create a perfect typology that fits any and every contingency but to determine
what that typology tells about patterns in the data. Treating classification as an end is
likely to lead to the problems Smith pointed out-reification of categories and attempts
to force data into those categories. This is a problem not of classification but of
misuse of a typology by researchers. Such misuse should not lead us to reject
classification itself. Nevertheless the appearance of misuse should lead us to consider
our own use carefully lest we succumb to the same temptation.
The constructs "tribe" and chiefdomm" have been criticized by Fried (1968, 1975)
and Lewis (1968). Their criticisms of "tribe" seem to be justified in some measure, and
Service himself dropped the term later (Service 1975). Lewis's criticism of the
category of chiefdom, however, is not borne out in ethnographic studies. Cohen and
Schlegel (1968), in a cross-cultural examination of societies that are neither bands nor
states, find that "the polities [break] ... clearly into chief and non-chief categories
based on quite a distinctive set of associated qualities" (1968:144). And, to answer
Lewis's (1968:104-105) criticism that the concept is based on a single geographic area
(Polynesia), it should be noted that chiefdoms occur worldwide, except in Australia
(Cohen and Schlegel 1968; Carneiro 1981:48-49).
The fluidity and dynamism of human societies justifies a closer look at
criticisms about the directionality and unilinealism implicit in Service's scheme.
Probably most researchers today would not presume to suggest a unilineal band-to-
tribe-to-chiefdom-to-state trajectory, but would recognize both the lack of
directionality and the multiple trajectories possible (see, for example, Anderson 1990;
Peebles 1986; Sanders and Webster 1978; Claessen, van de Velde and Smith 1985).
Doubts about the utility of the overall scheme as an evolutionary process should not
lead to rejection of the utility of the individual constructs. It is clear that chiefdoms
are recognizable entities in the ethnographic record (Cohen and Schlegel 1968;
Carneiro 1981). Service's scheme, then, is useful in some situations, such as when
comparing societies in different categories and, as in this study, when addressing
questions general to a category. If we accept the reality of chiefdoms, we can ask,
"What is the nature of capitals in chiefdoms?" Conversely, we should be willing to set
aside the classification when it interferes with the study of a particular issue (short-
term changes, for example, which might be studied better by analyzing particular
The Concept of Chiefdom Summarized
In this dissertation, the view that the concept of chiefdom is a viable one is
accepted, with the caveat that some subjects (e.g., the internal dynamics of chiefdoms)
are better served by setting aside the concept for the course of the analysis. The
category of chiefdom is seen as a descriptive one, not as an "evolutionary stage." The
term "evolution" is avoided here because of its unfortunate connotations of
directionality in anthropological literature. When discussing changes from one form
to another, e.g., chiefdom to state, the less value-laden word "transformation" is used.
It is clear that human societies are extraordinarily fluid in composition, but it is
equally clear that recognizable patterns of political organization exist.
The pattern of political organization in chiefdoms exhibits a centralized,
legitimized political power absent in egalitarian groups. Centralized societies show
some tendency toward increasing population density, but the size difference is less
important than differences in the structure of political power and authority. The same
is true of the distinction between chiefdom and state. The structural difference in this
case is the internal specialization of the centralized state government (Wright 1984:42).
To summarize the view of chiefdom followed in this study: A chiefdom is a
multi-local polity with political power vested in a central office. Political power is, in
contrast to that of the state, generalized; each level of administration replicates levels
above or below. Carneiro's terms minimal, typical, and maximal (see above) are used to
describe some size variability within chiefdoms. The term simple or petty chiefdom is
occasionally substituted for minimal chiefdom, and when discussing typical and
maximal chiefdoms, the two are sometimes grouped together as complex chiefdoms.
Power and Authority in Chiefdoms
Central to the study of political structure and to the analysis of capitals in
chiefdoms are the concepts of power and authority. Despite apparent consensus
among students of political structure regarding the importance of these concepts
(Swartz, Turner, and Tuden 1966:7; Lewellen 1983:89), there is disagreement about the
definitions of these terms. Indeed, power and authority are sometimes used
interchangeably (see for example Fogelson and Adams 1977). This fuzziness of
terminology obscures an important distinction, one that is integral to the study of
political structure. The discussion below relies largely on the definitions of political
anthropologists Ronald Cohen (1970), Henri Claessen (1988), and M.G. Smith (1968),
and political scientist David Bell (1975), who are very careful to distinguish between
power and authority.
The definition of power contains two related elements: "the ability to pursue
one's will effectively" (M.G. Smith 1968:193) and "an ability to influence the behavior
of others" and control their actions (Cohen 1970:488). The use of the term "ability" in
these definitions further emphasizes that power is not the right to control people's
actions, but the capacity to control them. Power is control of people-their labor,
services, or products-regardless of the holder's right to exercise power.
Sanctions (both positive and negative) play a large part in the exercise of
power (Bell 1975:21). A positive sanction takes a form like this: "If you do what 1
want you to do, I will give you something you want." A negative sanction, which
might work equally well to gain control of the other's actions, looks like this: "If you
do not do what 1 want you to do, I will harm you in some way."
It is important to separate power itself from the bases of power-the ways in
which individuals gain or maintain power. The simplest, and perhaps most fragile,
power base is persuasion, usually by force of personality. The second power base is
the competent exercise of leadership or the potential for it. In other words, the
individual who handles power wisely gains power. Third, and perhaps most cited as
a source of power, is the control of valued resources (cf. Adams 1975:19; Claessen
1988:23). These resources may be tangible (e.g., exotic trade goods) or intangible (e.g.,
specialized information, access to supernatural forces). The fourth power base is
force. This base requires a certain amount of power to start, that is, the leader must
control the actions of the individuals who make up the force (i.e., the warriors or
soldiers). And, finally, power may be based on authority or the right to exercise
power (see below). These power bases are not always independent of each other, but
often overlap and interlock.
"Authority is legitimate power" (Cohen 1970:488). Legitimacy is "the situation
in which the rulers as well as the ruled share the conviction that the existing division
of power ... is right" (Claessen 1988:23). Legitimacy thus provides the justification
or right by which an individual may exercise power. Authority, then, involves two
main concepts: legitimate right to power and acceptance of (or at least acquiescence
to) that right by the leader's subjects.
These two characteristics make possible commands by those in authority to
gain their wishes. This contrasts with the contractual nature of power acquisition.
An individual in authority no longer needs to promise or threaten ("If you do this for
me, I will do that for you") to gain his will, but may simply say, "Do this" (Bell
1975:37; M. G. Smith 1968:193), with every expectation of being obeyed without
The establishment of authority depends on the successful invocation of one or
more of several sources of societal values. These include (1) tradition (history,
legends, precedent); (2) inheritance or social rank; (3) support by the gods ("divine
right"); and (4) a code or laws. The last type of authority occurs primarily in states;
the first three types appear in chiefdoms as well.
Once authority is established, a leader can maintain it only if the values he
invokes continue to be accepted by his followers. These followers consist of an inner
circle (nobles, the leader's relatives, and others with access to him) and an outer circle
(commoners) (Claessen 1988:25-26). Although a successful leader must maintain a
relatively high degree of legitimacy in both circles, satisfying the inner circle becomes
increasingly important as the complexity of the polity increases (cf. Claessen 1988:30).
The Interaction of Power and Authority
In egalitarian societies, leaders are leaders by virtue of their own personal
power-e.g., their ability to hunt or find food resources, lead raids, settle disputes, or
contact the spirit world. Power in one realm often does not extend into another.
Power is thus individualized and situational. When leaders die or decline, their
power vanishes, it is not passed on to a designated successor; power must be
established anew with each new leader. Thus in egalitarian societies, though the
power of some individuals may be great, authority is minimal (based largely on
tradition). This is the case even for Big-Man societies where considerable generalized
power may be exercised by one person.
In chiefdoms, leaders are leaders by virtue not only of their own power
(control of others' actions through personality, capability, wealth, or force) but by
virtue of their authority (their right to lead). Unlike the segmented power and
minimal authority of egalitarian societies, power and authority in chiefdoms cross-cut
many realms--political, economic, and religious. The invocation of authority results in
a leadership that is no longer individualized and situational but generalized and
heritable. Authority exists independent of individuals and thus may accumulate from
one generation to the next. The possibility exists then of ever-increasing amounts of
power and authority.
Nonetheless, the nature of power and authority means that the success of
chiefs depends entirely on the cooperation or acquiescence of their followers. The
power relationship between chief and followers is reciprocal. The followers give the
chief their support in the form of tribute, corvee labor, and military service. In return,
the chief maintains a prosperous economy, keeps the polity secure from enemies and
internal dissension, and provides for the people in time of need. In addition, to retain
the support of the nobles (the inner circle), the chief must maintain control of wealth
and prestige items and privileges and must distribute them generously to the nobles
(and judiciously, if at all, to non-nobles).
It is also important that the chief and most of the followers (especially the
inner circle) share the same view of the chief's right to rule. Changed legitimacy (the
basis for authority) on the part of either the ruler or the ruled can undermine the
leader's success. For example, reluctance on the part of followers to recognize the
leader's right to rule forces the leader to rely heavily on sanctions (promises or
threats) to gain their cooperation. In times of growth or infrastructural crisis, when
challenges to the leader's power abound, the leader may not be able to compete
adequately through sanctions for the support of his followers.
The Role of Communication in Maintaining Power and Authority
As has been seen, for a leader to acquire and maintain power and authority,
he' must have the support of followers (both inner and outer circle). To get and keep
this support, he must convince followers that he has both the ability to carry out
promises and threats and the legitimate right to rule. Acquisition and maintenance of
power and authority, then, is largely a "public relations job." The leader must "sell"
followers on his ability and right to rule. And, as can be seen every day in the
modem world, success in sales depends as much on the effectiveness of advertising as
on the quality of the product.
1 The use of masculine pronouns in this discussion of leaders (and elsewhere in
this dissertation) comes at the expense of unfortunate implications that all leaders are
male (they are not, of course). Nevertheless, it is nearly impossible to discuss the
nature of leaders without the use of a third-person-singular pronoun. Masculine
pronouns were chosen because the majority of chiefs were and are male.
Advertising power and authority, as with any kind of advertising, is most
effective when it is easily understood, highly visible, and reaches the largest audience
possible. Apart from actual words (persuasion), the best way to communicate an
idea is through visual symbols. The visual representation of power and authority
may be either deliberate or unintentional, but it is as essential to the leader's success
as actual abilities and rights are.
Power is communicated primarily through rituals, activities, and constructions
that demonstrate the size of the leader's following. Large public gatherings (whether
coerced or consensual) show participants and outsiders (e.g., foreign visitors) the
might of the chief who convened the assembly. So, too, does evidence of the ability
to mobilize a large work force. A. C. Milner (1982:27) describes a Malay ruler who
used this principle to communicate his power to a Chinese emperor:
When the Sultan of Melaka wanted to impress the Chinese Emperor he
sent him a ship full of sago. The Emperor was told that the Malay
ruler had ordered each of his subjects to roll out a grain of sago until
there were enough grains to fill a ship: "That will indicate," explained
the Melaka envoy, "how many are the subjects of our Raja." "This Raja
of Melaka," the Chinese ruler is made to reply, "must be great indeed."
Large constructions also provide evidence of large work forces and have the added
advantage of being more permanent and visible than public gatherings or the
collection of tribute.
Symbols of power differ in form but share one characteristic, that of excess.
Constructions are bigger than necessary, crowds overflow public spaces, and feasts
provide more food than anyone can eat. Size and number are thus indicators of
Authority (the legitimate basis of power) is more dependent than power on
symbols for communication (i.e., a certain amount of power is communicated directly
through the continuing prosperity and security of the polity). Because authority exists
solely in the minds of the followers, it must be spelled out, redefined, or recreated
constantly. Authority is advertised by intangible symbols such as ritual ceremonies,
mythology, ritual gestures (e.g., salutes, blessings), and protocol surrounding the
treatment of the leader. Authority is also conveyed by tangible symbols such as the
leader's regalia, the physical surroundings of the leader, and the settings for public
Although authority symbols may, like power symbols, be excessive in nature,
the primary information being conveyed is that the leader is different and special;
ability to mobilize people is less important. Symbols of authority, then, convey
information by decoration or quality of workmanship and by their limited
A successful leader must learn to manipulate these symbols. If he handles
them ineptly or neglects them (e.g., denies nobles access to authority symbols,
converts to a minority religion, discontinues public or elite rituals), he conveys disdain
(whether intentionally or not) for the followers' notion of legitimacy and erodes their
faith in his own right to rule, leaving himself vulnerable to challenges to his power.
It follows then that a leader's continuing success also depends on access to
material symbols of authority and control of them. If, for reasons beyond his physical
control, he loses access to or control of the accepted symbols of authority, the leader's
own legitimacy is undermined in the eyes of the followers. The consequences of this
undermining depend largely on historical circumstances. If the polity is prosperous
and the chief is competent, he (and his successors) may survive by shifting the basis
of legitimacy to one for which he can control the symbols. If the polity is struggling
economically, if there is internal dissension, or if the chief is an incompetent ruler, the
loss of legitimizing symbols may result in a reformation of the power structure
(invariably including deposition of the incumbent).
It is, therefore, vital for a leader to advertise his abilities and rights. As has
been seen, leaders may do this through several media (words, rituals, protocol,
regalia, and structures). In an archaeological study, not all media are available for
analysis. One that is available and relatively accessible is the built environment. The
built environment is a particularly useful medium for study as it can be seen as a
semiotic system (Preziosi 1979:1 and cf. Knight 1981:iii), one which uses architectonic
elements to communicate. If this is so, then the structures and spaces in the leader's
immediate environment (the capital) will communicate information about the political
structure of the polity. This is the premise on which this dissertation is predicated.
Let us now turn to a brief discussion of capitals.
What is a Capital?
The term "capital" is not often used in connection with non-state political
centers. Perhaps this is because of its modem connotations. After all, London, Tokyo,
and Mexico City are capitals. These sprawling cities, hundreds of square kilometers
in area, bear little resemblance to the precolumbian 24-hectare chiefly center at the
Lake Jackson site near Tallahassee, Florida. Is it reasonable to lump such disparate
entities under the same rubric? Yes-with one caveat.
"Capital" is an appropriate word to describe political centers of any scale. It
derives from the Latin for "head," and its dictionary meanings include "seat of
government" and "chief or principal city." If we eliminate any implication of "city"
from the definition (this is the caveat mentioned above), "capital" describes chiefly
religio-politico-administrative centers quite well and has the advantage of being
succinct. Furthermore, "chief" (and through it, chiefdomm") derives from the same
Latin root, making the use of "capital" to describe a chiefly center all the more
Capitals are not found in all societies. The notion of a seat of government or
even a principal settlement presupposes the presence of a centralized government.
We should not expect to find capitals in egalitarian societies. Centralized polities such
as chiefdoms, early states, and industrial states all have capitals, though these may
vary in size and structural complexity.
Capitals as Reflections of Political Structure
The capital of a polity provides a very visible medium for the expression of the
power and authority of the centralized government. Few archaeologists have
explored the symbolic nature of capitals in any detail (but see de Montmollin 1989),
but it is not a new idea to historians and art historians. So Peter Duus (1969:22) says
confidently regarding an early Japanese state,
The central bureaucracy was housed in a permanent capital city, a
visible embodiment of the power of the monarch. Laid out on a symmetrical
grid pattern, it symbolized the orderly and harmonious character of imperial
rule (emphasis added).
Though archaeologists studying chiefdoms have rarely looked to the structure
of capitals for information on political organization, the concept is merely an
extension of a basic tenet of archaeology and geography. This tenet holds that the
spatial arrangement of a polity's settlements reflects aspects of the political and social
organization of the polity (see, for example, Findlow and Goldberg 1983:214;
Steponaitis 1978; Renfrew and Level 1979; Wright and Johnson 1975). The study of
the spatial arrangement of particular settlements (capitals) simply focuses the analysis
In the next chapter, information from 30 ethnographically and historically
known chiefdoms is examined to establish the parameters of size and structural
characteristics of chiefly capitals. Subsequent chapters will focus on Mississippian
capitals and the Lake Jackson site, a precolumbian capital associated with the Fort
Walton culture, a Mississippian manifestation in northwest Florida.
CAPITALS IN CHIEFDOMS
In studying political organization, archaeologists have attributed certain
characteristics to political capitals. Colin Renfrew and Eric Level (1979:146), seeking
to determine polity area from the location and size of the capital, make the
assumption that the capitals "are in general the largest settlement or administrative
sites within the territories of their polities." Similarly, Henry Wright (1984:43)
characterizes paramount capitals as
both larger than and architecturally differentiated from ordinary chiefly
centers, both physically accommodating the paramount's following and
providing a focus for major social rituals.
Rarely are assertions like these questioned in print, but such questions are
often raised verbally. Is it, in fact, reasonable to assign such characteristics to
precolumbian chiefly capitals? Are there other factors that might lead a site to look
like the archaeologist's very general image of a capital? Perhaps a site with a large
population lies in an economically important location, and that accounts for its large
size. Maybe an archaeological site with monumental architecture is a ceremonial
center, not a political one.
With these questions and thoughts in mind, I reviewed the ethnographic and
ethnohistoric record, looking specifically for information on the physical
characteristics of chiefly capitals and other settlements in chiefdoms.
As it turns out, many of the assertions made by archaeologists are generally
accurate. Some surprises do occur, however. The information in this chapter, then, is
presented in support of statements such as those cited above, and additional
information is offered on the shape of chiefly capitals.
Chiefdoms in the Ethnographic and Historic Record
To acquire data on chiefly capitals, 30 areas world-wide were investigated (see
Table 2-1 and Figure 2-1). An effort was made to examine capitals from chiefdoms of
all sizes (e.g., minimal, typical, and maximal chiefdoms).
Most groups discussed here are not single political entities, but congeries of
chiefdoms similar in adaptation and political, social, and religious structure. When
the Alur are mentioned, for example, the term encompasses the Ugandan chiefdoms
of Ukuru, Paidha, Padea, War, and many others. In a few areas, the group name
refers to a single (frequently maximal) chiefdom (e.g., Bemba or Tonga).
The nature and quality of data on chiefdoms varies greatly from area to area.
Before discussing the results of the survey, then, let us examine the database itself to
see how its shape might affect results.
The Geographic Shape of the Database
In this survey of chiefdoms, some geographic areas are better represented than
others. More than half (17) of the cases come from Africa and Oceania (Table 2-1 and
Figure 2-1). Only eight cases come from the entire American landmass. The
Table 2-1. Chiefdoms Discussed in Text.
GROUP DESCRIPTION LOCATION SOURCES
Powhatan (17th North America
S( 1 typical chiefdom h A a Rountree 1989
Natchez (18th 1 minimal or North America Le Page du Pratz
century) typical chiefdom (Mississippi) 1975 
Apalachee (16th North America Hann 1988;
& 17th 1 typical chiefdom (Florida) Varner and Varner
Calusa (16th & North America Solis de MerAs 1964;
17th centuries) typical chiefdom (Florida) Hann 1991
Taino (16th 1 typical chiefdom Central America Wilson 1990
Comogre and 30+ minimal and Central America
neighbors (16th Helms 1979
century)typical chiefdoms (Panama)
Omagua (16th & typical or maximal South America Meggers 1971; Fritz
17th centuries) chiefdom (Brazil) 1922
Tapajos (16th typical or maximal South America Meggers 1971
century) chiefdom (Brazil)
Nabdam several minimal West Africa 71
Nabdam Archer 1971
e ca. 90 chiefdoms West Africa Littlewood 1954;
of all sizes (Cameroon) Guidoni 1975
Mandarin 30+ minimal East Africa Buon
Mandarin Buxton 1963
many minimal East Afca
Alur and typical (Uganda)Southall n.d.
Ovimbuu 22 minimal and Central Africa Mc loh
Ovimbundu McCulloch 1952
typical chiefdoms (Angola)
ca. 80 minimal Central Africa Jaspan 1953; Smith
lla and Dale 1920;
chiefdoms (Zambia) Light 1941
Ll I Light 1941
GROUP DESCRIPTION LOCATION SOURCES
1 maximal Central Africa Whiteley 1950;
Bembachiefdom (Zambia) Richards 1939, 1971,
5 typical and
5 typical and Central Africa
Yao minimal Mitchell 1956
1S i maximal South Africa Kuper 1963;
chiefdom (Swaziland) Pettersson 1953
Zulu (19th 1 maximal South Africa Biermann 1971;
century) chiefdom (South Africa) Pettersson 1953
Basseri 1 typical chiefdom Southwest Asia Barth 1961
and typical Southeast Asia Leach 1965;
Kachin chiefdoms (in Friedman 1975
(Burma) Friedman 1975
close contact with
Gullick 1965, 1987;
Western Malays typical or maximal Southeast Asia Gullic 1972 1987;
McNair 1972 ;
(19th century) chiefdoms (Malaysia) Milner 1982
Loeb 1935; Fraser
Asi 1968; Cole 1945;
South Nias several minimal Southeast Asiaol 19;
Islanders chiefdoms (Indonesia) Heine-Geldern 1935;
Toraja and neighboring Southeast Asia Waterson 1989;
Luwu of South minimal Waterson 1989;
Sulawesi chiefdoms (Indonesia) Kennedy 1953;
Powell 1960; Weiner
Melanesia 1988; Malinowski
Kiriwina minimal chiefdom (Trobriand 1935, 1961 ;
Islands) Fraser 1968; Johnson
and Earle 1987
2(?) minimal Melanesia
Tikopia chiefdoms (Solomon Firth 1957
GROUP DESCRIPTION LOCATION SOURCES
i McKnight 1974;
1 typical chiefdom Micronesia McKnight 1974;
Palau (formerly 2) (Palau) Morgan 1988;
1987; Smith 1983
5 minimal Mic a Hanlon 1984;
Pohei chiefdoms (prehis- Pohnpei Morgan 1988;
torically 1Pope Petersen 1982;
chiefdom) [Ponape]) Riesenberg 1968
Kosrae 1 typical or Micronesia M n
maximal chiefdom (Kosrae)
Tonga (18th 1 typical or Polynesia Beaglehole 1961;
century) maximal chiefdom (Frindl Ferdon 1987
Tahiti (18th 2 typical or Polynesia Beaglehole 1955;
century) chiefdoms (Society Islands) Ferdon 1981
continent of Asia yields five cases (all but one from southeast Asia) and Europe none
at all. This distribution is not a deliberate bias. Rather, it reflects the geographic
availability of documentary data on chiefdoms.
In Africa and Oceania, many chiefdoms survived into modern times, to be
visited and described by anthropologists and travelers in the nineteenth and twentieth
centuries. These sources provide the best and most complete information on capitals.
In other parts of the world chiefdoms rarely survived into recent times.
Documentation of historical chiefdoms (e.g., in Iron Age Europe, sixth century A. D.
Japan, or sixteenth century South America) is scattered and incomplete at best. I
have made an effort to include some historical chiefdoms in this study where
documentation is available. Nevertheless, a substantial geographic gap remains.
Earlier researchers have been criticized for depending so heavily on Africa
and Oceania in defining chiefdoms (Lewis 1968:105). However, my review suggests
that while detailed studies of other chiefdoms around the world would add variations
on the chiefdom pattern, the additional information would do nothing to alter the
general concept of chiefdoms. So, while it would be desirable to fill in the geographic
gaps, the existing database is adequate for studies of chiefdoms.
The Effect of Contact with States
Many, perhaps most, chiefdoms with ethnographic or historic documentation
have been affected directly or indirectly by contact with states in one of two ways:
(1) simple contact with a state-level society or (2) imposition of governmental control
by a state-level society (colonialism). It might reasonably be asked if this influence
introduces bias into a cross-cultural study.
The first form of contact--simple contact with a state--can be eliminated as a
problem. Chiefdoms throughout the ages rarely, if ever, exist in isolation. Whatever
the time period, they are subject to influences from other societies at other levels of
organization. These influences may or may not be incorporated into the structure of a
particular chiefdom. Unless the influences result in a transformation to a different
level of organization, they should be seen simply as part of the ongoing dynamic of
chiefly political organization. Thus, for example, the western Malay chiefdoms
incorporated many features of Indian states but were, in the early nineteenth century,
still clearly organized at the chiefdom level' (see Gullick 1965).
In the second case, that of control over the chiefdom by another polity
(colonial dominion), the individual circumstances must be evaluated to determine the
effect on the traditional society. Some colonial situations leave the traditional political
structure largely intact, merely adding the colonial government as the top political
level (e.g., Apalachee, Alur, Yao, Kiriwina). For these groups, we can use available
ethnographic information comfortably. In other situations, the traditional political
structure is completely disrupted by colonial domination, and the ethnographic
information is too distorted to be of use. Many chiefdoms in the late twentieth
century fall into this category.
It is important therefore to evaluate data carefully, given the possibility that
some features of a particular ethnographically known chiefdom may be a product not
of chiefdom-level organization but of influence from a state-level polity. In particular,
if the goal is to acquire data for use in comparison with precolumbian chiefdoms (as
it is here), we must be careful to filter out distorting colonial influences. Given the
immense changes of the modern era, this task is difficult, but it is not impossible.
Many chiefdoms were recorded ethnographically before major disruptions occurred;
some (such as Kiriwina) even continue relatively unscathed to the present (Weiner
1 By the end of the nineteenth century, however, further influence in the form of
British colonial presence had resulted in a transformation to state-level political
organization (see Gullick 1987).
Some Comments on Terminology
I have identified the groups included here as chiefdoms, based on my
definition (see Chapter 1). As noted in Chapter 1, I use the terms "chief" and
chiefdomm" in very specific ways. The same cannot be said of the many sources I
consulted. In these sources, the entities called chiefdoms here are variously called
tribes, kingdoms, nations, or states. Their leaders are chiefs, kings, rajas, and sultans.
These terms, by their connotations, prejudice the reader regarding the political nature
of the polity and may cause considerable confusion.
In the following discussion of the 30 chiefdom areas, I have standardized the
terminology relating to polities and their leaders except in occasional direct
quotations. The reader should not view terms such as "king" or "sultan" or the
characterization of the polity as a "tribe" or "state" by the original recorder as
indicative of the polity's political complexity.
The Availability of Data on Capitals in Chiefdoms
Ideally, an archaeologist studying chiefly capitals would wish for a
quantitative study of the characteristics of capitals (including such features as number
of houses, population, area, dimensions of chief's house compared to ordinary houses,
and so on). Two factors restrict the ability to conduct a quantitative study of the
characteristics of chiefly capitals. One is the (statistically) small number of
ethnographic chiefdoms in the sample. Although the 30 chiefdom areas surveyed
represent a sizable proportion of those areas for which adequate data are available,
the sample is too small to be able to draw meaningful statistical conclusions.
The second factor concerns the paucity of specific quantitative data on
characteristics of capitals. So, for example, while recorders often remark on the size
of a capital in general terms, the number of structures and even population are noted
only occasionally. One size characteristic particularly useful to archaeologists-area--is
completely absent in accounts of the 30 chiefdom areas.
Although specific quantifiable data are few, careful readings of accounts of the
30 areas reveal considerable descriptive information about chiefly capitals. Many
ethnographers, especially the British social anthropologists working in Africa from the
1930s to the 1950s, recorded data on political structure. In so doing, they provided,
though often indirectly, data on the material characteristics of capitals.
The database thus lends itself to a qualitative rather than a quantitative cross-
cultural analysis. In the following sections, a synthesis of data on the size and shape
of capitals in chiefdoms is presented.
Size of Capitals
Writers describe the chief's village or capital as the largest in the chiefdom in
20 of 30 cases (see Table 2-2). Unfortunately, more specific information about the size
of the capital (e.g., population, number of houses, area) and its size relationship to
other villages is rarely provided. Nonetheless, the assertion that the capital is a
chiefdom's largest settlement is an important one because it provides a measure by
which to identify capitals archaeologically. Let us look more closely at this statement.
Why should the capital be the largest settlement in a chiefdom? First, the
capital is the residence of the ruler, and the ruler's immediate family is often larger
than the average family, especially in polygynous societies. The ruler typically uses
Table 2-2. Size of Chiefly Capitals (based on population or number of houses).
Capital is Largest Capital is Not Largest Insufficient Data on
Settlement Settlement Size of Capital
Direct evidence: Powhatan
Mandari South Nias Islanders
Alur South Sulawesi
marriage to create alliances with other high-ranking families both inside and outside
his chiefdom (Buxton 1963:68; Errington 1989:279; Gullick 1965:86; Leach 1965:205;
Southall n.d.:82-83; Weiner 1988:105). Consequently, the chief frequently has more
wives and children than ordinary men. Often, as in Kiriwina, the privilege of
polygynous marriage is restricted to those of high rank (Weiner 1988:105). Moreover,
distant family members may settle nearby to take advantage of being related to the
ruler (Richards 1972:109; Riesenberg 1968:66). In addition to relatives, the leader's
household often contains servants and slaves which increase the size of the capital
(Gullick 1965:64; Southall n.d.:77).
Second, in chiefdoms, the political, economic, and religious realms are
inseparable. The capital is, therefore, the center of all these realms. People directly
involved in centralized political, economic, or religious activities tend to live nearby.
The chief's councillors, for example, may live close by to facilitate communication
with the chief (Richards 1972:110; Southall n.d.:77). Moreover, people wishing to
benefit politically, economically, or religiously are drawn to the capital. In particular,
the poor and the ambitious are attracted by the prospect of support, wealth, or
advancement (Gullick 1965:64; Southall n.d.:77). Physical nearness to the chief puts
them in position to attain any or all of these.
It should be noted that though the capital is attractive to followers for the
reasons given above, living nearby also has higher costs. Those close to the chief are
called on more often to perform service for and provide tribute to the chief (Southall
n.d.:88). Proximity to the chief means that rivals have less chance to develop
followings and create their own power bases. Nevertheless, the attraction of the
capital is strong enough to offset its contradictory repellent forces, particularly for
those who have little likelihood of developing their own power bases.
The attraction of the political capital is apparent even in modem nations where
political, economic, and religious functions are often separated. In the world today,
the political capital is commonly the largest settlement in a polity. North Americans
may not immediately recognize the validity of such a statement because it is not true
of either the United States or Canada. The North American situation, however, is
highly unusual. Some figures taken from The World Almanac and Book of Facts 1992
(World Almanac 1992:734-821) show just how biased a view based on a North
American perspective is. For example, 15 (75%) of the 20 largest cities in the world
are current or former national capitals. The remaining 5 (25%) are current or former
Viewed another way, 137 (87.2%) of 157 world capitals are the largest cities in
their countries. An additional 11 capitals (7.0%) are the second largest cities. This
means that an astonishing (at least to a North American) 94.3% of world capitals are the
first or second largest cities in their polities. Only nine capitals (5.7%) are not the first or
second largest cities. These are almost all artificial creations (e.g., Brasilia, Canberra,
and Washington), recently moved capitals (e.g., Belmopan, Belize and Islamabad,
Pakistan; the former capitals were the largest cities), or capitals of confederations of
states (e.g., Brasilia, Canberra, Washington, Ottawa, New Delhi, and Bern) or some
combination (cf. Renfrew and Level 1979:146).
As has already been noted, modem national capitals are less likely than chiefly
capitals (because of the separation of political, economic, and religious activities) to be
the largest settlements in their polities. The very high proportion of modem capitals
which are, in fact, the largest settlements (87.2%) suggests that chiefly capitals may be
expected to be the largest settlements in their polities at least 90-95% of the time.
The data from the chiefdom survey, though limited, support this expectation
(Table 2-2). In all areas for which information exists regarding the size of the capital,
the capital is the largest settlement. There are no cases in which the capital is not the
largest settlement. Given this body of evidence, we may safely conclude: In a
chiefdom, the capital is the largest settlement in the polity.
As definite as this seems, applying this information to the interpretation of
archaeological data requires some care. To begin with, all the data above are based
on population figures. The largest settlement is the settlement with the largest
population. Population information, of course, is not immediately available to
archaeologists, but it can be derived indirectly from material remains such as number
The settlement with the largest population is not necessarily the settlement
covering the largest area. Consequently, archaeologists tempted to estimate size
based on the area covered by a site should exercise caution. Factors such as dispersed
or compact settlement and chronologically distinct living areas may distort
interpretations of site size.
Moreover, identifying the largest site in a polity is entirely relative. This can
be done only within the context of a knowledge of the entire range of site sizes and of
the boundaries of the polity. Identification of one seemingly large site as a capital is
thus unreliable in the absence of reasonably thorough settlement pattern data.
Finally, the archaeological record may present a confusing picture, revealing
perhaps more than one site that appears noticeably larger than ordinary sites. Such
unclear data may indicate periodic re-location of the capital (a typical pattern in
chiefdoms) or the presence of more than one level of administrative leadership.
Clearly, a good grasp of the local chronology will help in sorting out problems like
Size, of course, is only one factor in identifying chiefly capitals. Equally
important, especially for archaeologists, is the structure of sites-constructions, spaces,
and their arrangement across the landscape.
Structure of Capitals
Capitals differ from non-capitals not only in size, as seen in the previous
section, but in structure. In this section the structural elements of capitals are
In the first part the disposition of these structural elements across the
landscape is examined, followed by a consideration of the importance of orientation.
Then, individual components are taken up in turn, including the chief's house, the
chief's storage facilities, religious structures, communal structures, public spaces, and
walls. Finally, the issue of monumental architecture is addressed; monumental
architecture is an infrequent feature of ethnographic chiefdoms but one which bears
heavily on the subject of this study.
Several factors affect the form of a settlement (both capital and non-capital).
The first is the type of adaptation to the environment. Many of the areas in this
survey, for example, exhibit a dispersed form of settlement (11 cases). This is
undoubtedly due to reliance on small-scale, family-run agriculture. Conversely, some
agricultural settlements may be clustered because crops grown or the method used
requires cooperation among families or use of large fields for efficient production
(e.g., among the western Malays). In the cattle-herding areas of Africa, the need to
contain livestock often results in a circular settlement plan with structures built
around the livestock pens (e.g., Ila, Swazi, Zulu). If the ocean, a river, or a road play
an important part in the economy, the settlement may be laid out to maximize access
and so may take on a linear form (e.g., Tikopia, Tapajos, some Yao).
The second factor that influences settlement form is the need for defense.
Some settlements, even agricultural ones which might be more efficiently located in a
dispersed pattern, are clustered for protection (e.g., western Malays, South Nias
Island). Others, having retreated to hilltops or mountain ridges for security, may
adjust to the physical configurations of the setting. So, for example, Ovimbundu
villages located in the plains tend to be circular, but those situated on mountain
ledges (due to an earlier need for defense) tend to be linear (McCulloch 1952:25).
Historical factors also affect the form of settlements. A vivid example of this
comes from the Micronesian island of Palau. During the Japanese colonial period
(1914-1944) on Palau, the traditional layout of Ngiual (a capital) gave way to "the
linear pattern of a rural Japanese village" (Parmentier 1987:58). The influence of
colonial rule thus resulted in a restructuring of the settlement's layout.
The plan of a capital differs from that of an ordinary settlement in several
ways. First, as shown in the last section, a capital is larger and may thus have a more
complex form. Second, a capital contains components not found in ordinary
settlements (e.g., chief's house, chief's storage). These unique components, as will be
seen in later sections, often form the focus of a capital and are set off in some way
from the ordinary elements. Third, in areas where the general form of settlement is
dispersed, the capital may be more clustered and take on the characteristics of a
"town" (e.g., Tonga, Bamileke, Apalachee). Fourth, in areas with foreign influence, a
capital may take on a new form (as in Ngiual) while ordinary settlements retain
In addition to differences, there are also similarities in the plan of a capital and
that of an ordinary settlement. Apart from the cases noted above, the plan of a
capital usually takes the same shape as that of other settlements in the polity. If the
typical form is circular, as in Kiriwina, for example, the capital is circular (Malinowski
1935:24). If the ordinary settlement plan is amorphous, as it is among the Nabdam,
the plan of the capital is amorphous (Archer 1971:49-50).
Clearly, many factors influence the layout of settlements. As a result, obvious
differences in layout between sites should be viewed cautiously to avoid interpreting
environmental or defense-related factors as political ones. Moreover, characteristics
relating to the difference between capital and non-capital (higher population, unique
structures) may or may not be readily apparent in archaeological remains. In some
cases identifying a capital may require nothing more than a broad knowledge of sites
in a given area. In other cases, extensive research at several sites may be necessary to
make a determination.
The orientation of a capital (or of any settlement) may be symbolic (e.g., to the
cardinal directions) or practical (e.g., to some feature of topography or weather).
Should it be expected, given the high political and ritual symbolic nature of the
capital in a chiefly society, that the orientation of a capital would have a strong
symbolic element? Apparently not, or at least no more so than for ordinary villages.
This survey of ethnographic chiefly capitals found that orientation of capitals
is generally similar to that of ordinary villages and most frequently depends on
topography. Where information is available, capitals are oriented to features such as
a river (Natchez, Powhatan, western Malays, Omagua, Tapajos), the beach or coastline
(Tikopia, Palau), a ridge or hilltop (Kachin, South Nias, some Ovimbundu), or a road
A symbolic orientation is present in a few settlements, primarily in Indonesia
(at least in this sample). On South Nias Island, for example, the siting of villages on
hilltops had both practical and symbolic significance. The hilltop provided a good
defensive position, but, more importantly, the slope of the hill acted as a physical
manifestation of social rank (Fraser 1968:37). The higher one's house site, the higher
one's rank. The chief's house, reached by a series of ascending terraces, loomed over
all the others.
The equation of physical height with high rank occurs widely in southeast
Asia and throughout Oceania. Errington (1989:66), for example, notes that in South
"high" rank required literal high and central placement on ceremonial
occasions. Even on non-ceremonial occasions,... to be polite, one
must keep one's head lower than the head of the higher-ranking person
in whose presence one finds oneself.
Other examples abound. In Kiriwina, at the turn of the century, the chief often sat on
a raised platform so people would not have to stoop when passing him. If he
happened to be sitting on the ground, passersby dropped to their knees and crawled
by, or the chief stood up (Malinowski 1935:34, 84). In Tahiti, the paramount chief was
carried about on the shoulders of a low-ranking subject (Ferdon 1981:38). In Tonga,
in the eighteenth century, the paramount refused an invitation to dine with Captain
Cook aboard his ship because descending to the captain's cabin meant that
commoners would be above the chief's head (Ferdon 1987:24). Elevation thus carries
a symbolic meaning in these parts of the world and probably in others as well (e.g.,
Presumably, symbolic orientation could also include orientation of a settlement
to the cardinal directions. However, very little evidence was found in the survey for
this type of orientation. The north-south alignment of houses in South Sulawesi
(Kennedy 1953:133) suggests an overall settlement orientation to the directions, but
none of the sources for this area states this explicitly. An apparent orientation to the
west among Ila villages in central Africa (see Figure 2-2) turns out, upon closer
scrutiny, to be the result of sheltering the chief's house from the prevailing east wind
(Smith and Dale 1920:109).
The lack of evidence for settlements (especially capitals) oriented to the
cardinal directions is somewhat surprising. This lack may simply be a product of the
small size of the sample. Or perhaps surprise at the lack reveals a modern or
geographic bias. It is possible, although the evidence for it is not at hand, that state-
level capitals are more likely to be oriented symbolically (including to the directions)
as a result of increased dependence on authority by leaders to maintain their power.
It is also possible that orientation to the cardinal directions is a characteristic typical
of particular geographic areas-Asia, for example-and is less likely to occur in other
areas. The lack of clarity should suggest caution in ascribing a directional (symbolic)
orientation to a site if a practical orientation also exists.
Whatever the orientation, we should not expect an exact placement of
settlements or of houses within the settlements. Sophie Clement-Charpentier
(1989:149), in discussing modern non-capital Thai villages, notes that "...orientation
may vary by a few degrees" from a defined pattern. Villagers, she says, "are
accustomed to following rules, but they feel free to interpret them in their own way."
This attitude very likely holds true in chiefly as well as non-chiefly societies and for
inhabitants of capitals as well as for those of ordinary villages. So, even in capitals
oriented symbolically, some degree of variation from the apparent norm should be
In general, settlement orientation, as with layout, seems to depend on
environmental and cultural features common to the area. The oriehtations of capitals
and of ordinary villages are similar. Typically, settlements are oriented to some
topographic feature. Symbolic orientations are rare.
The chief's house (or compound) is often immediately distinguishable to
observers by its larger size, its elaboration or decoration, or its prominent location.
Raymond Kennedy (1953:36-37), for example, singled out a large, striking house for
attention while traveling in South Sulawesi (Indonesia) in 1949. Upon inquiry, an
informant told him that the house belonged to the chief of Makale. Kennedy had
recognized, consciously or not, exactly what the characteristics of the building were
intended to communicate: this was the dwelling place of an important person,
someone worthy of attention.
Similarly, an observer viewing a capital in plan can often pinpoint the chief's
house quite easily by its size, shape, and/or location. Consider the Ila capital shown
in Figure 2-2. The chief's compound is clearly identifiable by its greater size and
central location. The chief's own house within the compound is further distinguished
by its rectangular shape and large size. The chief's house is also readily identifiable
in a plan of Omarakana, the capital of Kiriwina (Figure 2-3).
Size. In all chiefdom areas for which information was available (20 of 30
areas), the chief's house is larger than ordinary houses. In some cases, the layout of
the house is also more complex (Kachin, western Malays, Palau, Pohnpei, Powhatan,
Nabdam, Alur under colonial influence).
One reason for the larger size of the chief's house or compound is the larger
size of the chief's household. In addition, some chiefs' houses have functions that
ordinary houses do not (e.g., they may serve as the settings for community-wide
rituals [Kachin, South Sulawesi, Taino, Yao], as guest-houses for visitors [Kachin], or
as the locations for the conduct of government [Bemba, Yao, Ila, South Nias, western
Malays]). A large chief's house also demonstrates visually the chief's ability to
command the services of a large work force and therefore reflects the amount of
power the chief holds.
Form. In general, chiefs' houses have the same shape as ordinary houses.
However, in a few cases (Tahiti, Tonga, Taino, possibly Pohnpei), the shapes of
traditional chiefs' houses are described as different. Taino chiefs' houses, for example,
were rectangular while ordinary houses were round (Wilson 1990:57). Conversely,
Tahitian and Tongan commoners' houses were rectangular with rounded ends, but a
handful of chiefs' houses were round or oval2 (Ferdon 1981:72-73, 1987:21). In
Pohnpei, an early visitor remarked that "the dwelling houses vary in size and in
shape according to the taste and rank of the proprietor" (O'Connell 1972:125, cited in
Areas under foreign or modern influence show the most disparity between the
shapes of chiefs' and commoners' houses (Alur, Nabdam, Ila). This disparity results
from the chief's role as mediator between subjects and the outside world. Information
2 Most Tahitian and Tongan chiefs' houses, however, had the same shape as ordinary
o %es 000 0.
4 Dance 1 O
-4 Ground 0
C >. e Chief's Main
0 0\ House
C Payne 1992
Figure 2-3. Omarakana, Capital of Kiriwina, Trobriand Islands, Melanesia (after
or goods going in either direction are funneled through the chief (cf. Parmentier
1987:69). Chiefs, therefore, frequently acquire new products and adopt new styles
before others in the polity. Consequently, rectangular Nabdam, Ila, and Alur chiefs'
houses are built in European style of modern building materials. At the same time,
ordinary houses continue to be built in traditional shapes (round in all three cases) of
traditional materials (Archer 1971:50, 52; Light 1941:74; Southall n.d.:269-270, 303).
Decoration and elaboration. In this survey the chief's house is described as
more elaborate or more highly decorated in ten cases. In five of those ten cases
(Tahiti, Bamileke, western Malays, South Sulawesi, Kachin), such decoration or
elaboration is specifically described as the exclusive privilege of those of noble rank.
Among the western Malays, for example, the shape of a house's roof indicated the
owner's rank. Only the three highest-ranking nobles were allowed a two-tier roof;
houses of lesser nobles, however, could have concave roofs (Gullick 1965:112). In the
Bamileke area, decorated door panels and door frames marked the houses of nobles
(Littlewood 1954:98). In Tahiti,
it was strictly a chiefly prerogative to have the rafters of the home
wrapped with fine matting or the braided fibers from the husk of the
coconut (Ferdon 1981:79).
Perhaps the clearest association of house decoration and rank comes from
South Sulawesi. The number of flap-like decorations (tipe-tipe) hanging from the
house's gable communicated the owner's rank (Kennedy 1953:108-109; Errington
1989:80). In Luwu, for example, three tipe-tipe indicated the house of a ruler. The
number of gable flaps which signified high rank varied from area to area--south of
Luwu in Bone, rulers had up to seven--but the correlation of the highest rank with the
greatest number remained constant.
Shelly Errington (1989:80) describes an even more visible indicator of rank in
South Sulawesi-the form of access to the raised houses of the area. She records five
types of access, each symbolic of a higher status than the one before: a ladder, an
uncovered stairway, a stairway covered with a flat roof, a stairway covered with a
pointed roof, and a stairway covered with a pointed roof and having one or more
landings. Errington's description of an informant's house in Luwu as having "three
tipe-tipe and an elaborate covered stairway with two landings" thus indicates--even
without her confirmation-that the house belongs to "very high nobles."
In South Sulawesi the connection between house decoration or elaboration and
rank is stronger than that between house size and rank. "Large size in itself," says
Errington (1989:80), "does not indicate higher status, merely wealth. Higher status is
indicated instead by elaboration of spaces and of decoration." The visibility and
symbolic nature of these house features may explain why Raymond Kennedy
unerringly singled out a chief's house for consideration in the example cited at the
beginning of this section.
Construction materials. In nine cases, construction materials of chiefs' houses
are described. In eight of these, materials used in building the chief's house are
different (i.e., more desirable or stronger). Mary Helms (1979:9), for example,
describes Comogre chiefs' houses as built of timber with stone walls while ordinary
houses were made of canes plastered with clay. As noted earlier, in instances where
chiefs' houses take a foreign or modern form, building materials are also usually
modern, though traditional construction materials may continue to be used in
commoners' houses. In most chiefdoms the difference in construction materials is not
this great. In Pohnpei, for example, the chief's house is built of breadfruit wood and
hibiscus wood, while ordinary houses are made of mangrove wood (Riesenberg
1968:67). And Powhatan chiefs' houses differed from ordinary ones in having a
covering of bark rather than one of marsh reed mats (Rountree 1989:60-61).
Location. Finally, the position of the chief's house may be marked in some
way. In 12 of the 13 cases where location is determinable, the chief's house stands in
a prominent location. Prominence is expressed in one or more of three ways: (1)
separation from commoners' houses; (2) central location; or (3) elevated placement.
Segregation of the chief's house may be accomplished with fences or walls
(Kachin, Zulu, Swazi). Or it may be created by surrounding the chief's house with
more space than is usual (Kiriwina, Tikopia, Ila). In Kiriwina, for example, public
and ritual spaces surround the chief's house (Malinowski 1935:25, 431). These spaces
include some areas where access is controlled by the chief.
The chief's house may also be distinguished by a central location (Mandari, Ila,
Kiriwina). In the two plans of circular capitals illustrated earlier in this chapter
(Figures 2-2 and 2-3), the chief's house is positioned in the center of the circle rather
than among the rows of houses.
Chiefs' houses may also be marked by elevation (Natchez, South Nias, Palau).
As noted earlier, the equation of physical height with high social/political rank occurs
in several parts of the world. Chiefs' houses may thus be placed higher than
commoners' houses to demonstrate to both subjects and visitors the chief's high rank
and authority. The Natchez chief's house, for example, stood on an artificial earthen
mound (Le Page du Pratz 1975:338). South Nias capitals were laid out to take
advantage of natural hill slopes; the chief's house was situated at the highest point in
the village (Fraser 1968:37).
Comments. The chief's house is thus distinct from ordinary houses in several
material ways: greater size, prominent location, and more and/or better decoration.
The chief's house may also differ in form or construction material, especially in areas
with foreign influence.
This distinctiveness is not only material. In several areas, the chief's house is
distinguished linguistically. At least six areas (Pohnpei, Kachin, Tikopia, South Nias,
Alur, and western Malays) and possibly one other (Taino), have a special word or
phrase for "chief's house." This linguistic distinction emphasizes the unique nature of
the structure. The chief's house is not merely any house that the chief happens to live
in; it is a definable, recognizable structure that falls into a different category
altogether from ordinary houses.
The physical characteristics of the chief's house communicate both the power
and authority of the chief. The large size of the structure provides a visible indicator
of the size of the work force the chief can command, and the use of different
construction materials indicates the chief's control of resources not available to the
general populace. The chief's house is "a concrete representation of aristocratic
political power" (Waterson 1989:485).
Greater decoration or elaboration conveys both power and authority--power in
the form of the ability to command the labor required to create better decoration and
authority in the form of the right to use symbols restricted to use by individuals of
Location also communicates authority because it illustrates the chief's right to
(and the population's acceptance of) a special place in which to build his house.
House decoration and location express an authority primarily based on social rank (in
contrast to authority based on religion, for example).
Recognizing a chief's house in the archaeological record. Ethnographically, the
best indicators of a chief's house are size, location, and decoration. Chiefs' houses are
larger, prominently located, and frequently have elaborations or decorations available
only to nobles. The last of these indicators, decoration, is not often available to
archaeologists for study, so the identification of chiefs' houses in the archaeological
record relies primarily on size and location data. Because these data are relative, the
archaeologist should have a fairly good body of information on structures of all kinds
from a capital before making inferences.
An unusually shaped building or evidence for different construction materials
is not, in itself, adequate or necessary for the identification of a chief's house, but
either may serve as useful support for other forms of evidence.
One last caveat--a chief's house shares some size or locational characteristics
with chiefs' storage facilities, communal houses, and religious structures. The
following sections contain data which will help the archaeologist in distinguishing
among these structures.
Chief's Storage Facilities
Storage in chiefdoms. Over-production of subsistence products in good years
accompanied by storage of the excess for use in bad times often forms a part of a
chiefdom's subsistence strategy. In areas of year-round plenty (e.g., Palau), such
measures may be unnecessary and storage structures non-existent (Barnett 1960:27),
but for most chiefdoms, storage of food products plays an important role in staving
off recurrent shortfalls. Storage structures in these cases may be prominent parts of
the landscape of settlements.
Periodic shortages exist and must be counteracted because of the nature of
subsistence in chiefdoms. As populations become denser, gathering-fishing-hunting
strategies become impractical, and the subsistence base is eventually intensified
(Johnson and Earle 1987:16). This intensification usually takes the form of agriculture,
but may also occur as livestock herding or fishing. Intensification results in reliance
on fewer staples, thus opening up the population to greater threat in times of
Storage of excess produce, at the local level or at the polity level, is one way to
lessen the impact of shortages. Local-level storage acts as a buffer for small groups,
but, for the whole polity to prosper, centralized storage is necessary. Centralized
storage averages out the effects of bad harvests over the entire population,
minimizing the impact on the society as a whole. The chief thus acts as a "banker"
(Johnson and Earle 1987:223), with subjects sending in foodstuffs (tribute) in good
times and expecting the chief to feed them in hungry times. The chief carries out this
duty by holding periodic ceremonial feasts, by feeding corvee workers, and by taking
into his household individuals unable to support themselves (Barth 1961:101; Buxton
1963:73; Errington 1989:113; Gullick 1965:107-108, 1987:51; Hann 1988:209; Helms
1979:14; Leach 1965:112; Parmentier 1987:68; Richards 1939:147, 246; Southall n.d.:78-
Given the importance of the distribution of food by the chief, it might be
expected that chiefly storage facilities will be different from ordinary structures. And,
in fact, this is what happens. Before exploring these differences, let us look at the
types of storage structures we might expect.
Types of storage facilities. Storage facilities take different forms depending on
subsistence requirements. These forms fall into two categories and possibly a third.
The first and most common facility is storage for agricultural produce. Frequently,
these facilities are freestanding structures devoted specifically to storage, e.g., rice
barns (South Sulawesi), yam houses (Kiriwina, Tonga), corn cribs (Natchez), or millet
granaries (Alur, Bemba). However, in two cases, produce is stored directly in the
chief's house (Comogre, Nabdam). In two other areas, goods are stored in a religious
structure (Ovimbundu, Powhatan).
The second type of facility is "storage" for livestock, i.e., pens or corrals. This
type occurs primarily in Africa where there are many chiefdoms based on cattle
herding (e.g., Mandari, Ila, Swazi, Zulu). In south Africa, cattle pens or kraals are
prominent central features of homesteads and villages.
A third possibility (although only two vague references to it were found) is
"storage" for aquatic livestock--fish ponds (South Sulawesi) or turtle pens (Omagua).
Such structures might be expected in chiefdoms with maritime or riverine subsistence
bases. However, few such chiefdoms are documented, so it is difficult to say what
the likelihood is for such structures.
Size. When storage facilities are separate structures, they may be either larger
or smaller than residences. Bemba granaries, for example, are smaller than dwellings
(Richards 1939:Plates 1 and 3), while Kiriwina yam houses are larger (Malinowski
1935:229). Livestock pens, of course, are invariably larger in area than residences (Ila,
Swazi, Zulu) (see Figure 2-2).
In this sample of chiefdoms, only six cases provide data on the size or number
of chief's storage structures compared to those of ordinary storage structures. It is
noteworthy, however, that in each of these six cases, the chief's storage is larger or
more numerous than ordinary storage.
Sometimes the difference is substantial. The household of the chief of
Kiriwina contains 23 yam houses (Malinowski 1935:25). An ordinary Kiriwinan
household has only one, at most (Weiner 1988:91-92, 105). Moreover, the chief's
personal yam house is larger than all the others (see Figure 2-3). This is not
surprising in light of the fact that Kiriwinans view yam houses not only as symbolic
of power but as literal sources of power (Malinowski 1935:229).
Bemba chiefs also maintained large storehouses. Audrey Richards (1939:85)
found that the average capacity of an ordinary Bemba granary is 661 cubic feet.
"Chiefs' granaries," she says, "are bigger, and the four I measured had a capacity more
than four times this size."
An even more striking example comes from the nineteenth century Zulu
(Biermann 1971:99). Although sizes of ordinary cattle kraals are not given, a chief's
kraal is described as being more than one mile in diameter, surely an extraordinary
size. This enormous kraal existed primarily as a demonstration of the chief's wealth
and power. Though nominally the place where the chief's cattle were kept, the kraal
was actually used for gatherings of warriors. The cattle were kept in smaller kraals
around the edges of the big one.
Location. In looking at the location of chiefs' storage facilities, the problem
again arises that only a few cases provide adequate data. But as with size and
number, a clear pattern emerges despite the limited information. In six of the eight
cases where this information is noted, the chief's storage is in or near the chief's
house.3 This finding is entirely predictable because storage facilities in general are
located near their owners' residences (Southall n.d.:41; Smith and Dale 1920:112-114;
That the chief's storage is likely to stand near his residence means that it, like
his house, occupies a prominent place in the settlement. This is vividly illustrated by
the central and isolated positions of the cattle pen of an Ila chief and the personal
yam house of the chief of Kiriwina (Figures 2-2 and 2-3).
Comments. Clearly, considerable variation exists in storage facilities across
chiefdoms. To begin with, several types of storage occur: (1) agricultural produce
storehouses; (2) livestock pens; and, possibly, (3) containment for aquatic resources.
Variation in size of the storage facility compared to the size of residences also exists.
Some stores are smaller than houses (e.g., Bemba); some are larger (e.g., Kiriwina, Ila,
Swazi, Zulu). These variations depend largely on subsistence requirements, including
what types of food are being stored and the general productivity of the area.
Conversely, there are also patterns of similarity: (1) chiefs' storage facilities are
larger or more numerous than ordinary storage, and (2) chiefs' stores are sited in
prominent locations. However, both these patterns are based on very limited
information, so we may legitimately ask whether they will hold true if more data are
acquired. Most likely they will.
3 The two exceptions are Powhatan where goods were stored in a religious structure
outside the capital (Rountree 1989:133) and Natchez where a communal corn crib was
erected near the fields outside the capital (Le Page du Pratz 1975:339).
The chief's stores will probably be larger than ordinary stores (in chiefdoms
where storage is a necessity) due to the worldwide obligation of chiefs to feed people.
This obligation is specifically stated in 10 of the 30 areas and is indirectly expressed in
the fact that chiefs have larger than average households in 12 of the 30 areas.
Chiefs' storage facilities will probably be located in a prominent position in the
capital, near the chief's residence. This siting is partly a matter of practicality and
ease of access and partly a matter of the store symbolizing the chief's wealth and
power. One exception has relevance for this study: the Natchez build a communal
corn crib outside the capital (Le Page du Pratz 1975:339). It is not clear whether other
Natchez storehouses exist.
Interpreting chiefs' storage facilities in the archaeological record. Care must be
taken in using this information to interpret archaeological sites. As with the chief's
house, the data are relative. Storage facilities vary in size from area to area, and
absolute size cannot be used to identify chiefs' stores. Only by recording large
numbers of storehouses and interpreting the resulting pattern can chiefs' storehouses
be identified with any accuracy.
What do religious structures look like in chiefly capitals? It has been shown
that chiefs' houses and often chiefs' storage facilities are larger and more prominently
located than ordinary houses. Should the same be expected for chiefly religious
Certainly this is a reasonable expectation given that leadership in chiefdoms is
generalized rather than specialized and that the chief invariably fills the triple roles of
political, economic, and religious leader. Because the structures (chief's house and
chief's storage) representing the political and economic roles of the chief are larger, it
might also be expected that religious structures, representing the third aspect of the
chief's generalized power, should be large and conspicuous.
This expectation is not borne out. Religious structures in chiefly capitals vary
greatly in their nature, size, and location.
The nature of chiefly religious structures. Sites of ritual occur in several forms
(see Table 2-3). These include open-air sites (either natural or specially constructed),
buildings, shrines (either freestanding or within a larger building), and miscellaneous
types such as burial mounds. In a few cases, rituals are carried out in the chief's
house. Sometimes more than one type of religious structure appears in a chiefdom
Table 2-3. Forms of Religious Structures.
Open-air Building Shrine Other None
Mandari Powhatan Alur burial mound Nabdam?
Kachin Natchez Ovimbundu Kosrae
Kiriwina Apalachee Ila Tonga
Tahiti Calusa Bemba
Bamileke South Nias rituals carried
Ovimbundu Palau out in chief's
Yao Kachin house
These forms are typical of ordinary settlements as well as capitals. There is no
distinction between the nature of capital religious structures and ordinary ones.
Ancestor worship. Inheritance and social rank are fundamental legitimate
bases of chiefly power in chiefdoms of all sizes. It may be supposed then that
reverence for ancestors, especially chiefly ancestors, will play an important part in
rituals in chiefdoms. This supposition is borne out. Nearly half the 30 areas (13
cases) had religious structures devoted to ancestor worship. In ten cases, ancestor
shrines actually held ancestral bones.
In a chiefdom where ancestor worship is practiced, chiefs' ancestors take on
significance beyond significance to the chief's lineage. Chiefly ancestors often
represent the polity's ancestors. Frequently, "a chiefly ancestor shrine . has a
political focus" as it does for the Alur (Southall n.d.:92). At Alur chiefs' shrines, a
new chief is installed, rainmaking and first fruits rituals are carried out, and rituals
expressing the loyalty of clan heads and sub-chiefs to the chief take place (Southall
Size. Sizes of religious structures vary considerably as is indicated by a
comparison of the sizes of religious structures to those of chiefs' houses for each area
(see Table 2-4; excluded from consideration are structures situated within the chief's
house). In five of the thirty cases, the religious structure is smaller than the chief's
house. Shrines, for example are invariably small; Alur chiefs' shrines measure barely
four feet in diameter (Southall n.d.:99), and Ila ancestor shrines are "miniature huts"
(Smith and Dale 1920:113).
In five other cases, we find religious structures that are the same size or larger
than the chief's house. Further, extraordinary effort is evident in the construction of
some religious structures (six cases). The Natchez "temple," for example, stood on top
Table 2-4. Sizes of Religious Structures.
Smaller than Chief's Same Size or Larger than Extraordinary Effort
House Chief's House Expended
Bamileke Powhatan? Powhatan
Alur Natchez Natchez
Ila Calusa? Calusa
Bemba Kosrae Yao
South Nias Tahiti Kosrae
of an eight-foot-high mound (Le Page du Pratz 1975:338), and Tahitian maraes were
constructed of considerable amounts of stone (Ferdon 1981:54-55).
Little correlation exists between size of religious structures and complexity of
chiefdoms. Small religious structures, for example, occur in chiefdoms at varying
levels of complexity. Large religious structures, however, appear only in typical or
maximal chiefdoms. No large religious buildings are recorded for minimal
Location. The location of a capital's religious structure is unpredictable. There
is perhaps a slight tendency for it to be sited in the chief's house (four cases) or near
it (three cases) However, two cases occur where the religious structure is outside the
capital (Powhatan, Bemba). In four other cases, the religious structure is described
only as being in the capital.
Comments. Religious structures may be buildings (e.g., church, temple,
mosque), shrines or monuments (e.g., Tahitian marae), open spaces, or natural features
(e.g., Mandari "meeting tree"). Many are devoted to ancestor worship.
Religious structures range from small and inconspicuous (Alur chiefs' shrines)
to large and prominently located (Tahitian marae). In general, in small chiefdoms, the
religious structure seems to be smaller and less prominently located than the chief's
house. In some larger chiefdoms (e.g., Natchez, Tahiti) the religious structure is a
very visible and impressive part of the capital. However, small religious structures
also occur in larger chiefdoms (e.g., Bemba shrines). In chiefdoms of varying sizes,
the religious structure may be within the chief's house.
A religious structure is often located near the chief's house (e.g., Tikopia
marae); sometimes it is even a part of the chief's residence (e.g., Kachin madai dap).
This prominent location is by no means inevitable, however; some religious structures
even stand outside the capital.
What is to be made of this general lack of patterning? Why should religious
structures be exempt from the general equation of chiefly structures equals large size
and/or prominent location? The lack of pattern makes sense if considered not in
terms of the generalized political, economic, and religious aspects of the chief, but in
terms of power and authority.
To help explain this, it is necessary to consider power and authority and how
they are manifested in the built environment. As noted earlier, power is expressed by
constructions requiring large amounts of labor. The chief thereby demonstrates
ability to mobilize large work forces, thus advertising his large following--and people
are power. The chief's house and storehouse are mainly representative of power--an
ability to call up labor to build the structures and to demand tribute to fill the
4 The chief's house, by its decoration or elaboration, of course, may also express a source
of authority-social rank.
Religious structures, however, are physical manifestations of one source of
legitimate power or authority (divine support). Size or effort expended
(characteristics of power) may not necessarily therefore be factors in their
construction. If the chief's authority derives from sources other than religion (e.g.,
tradition, social rank), the religious structure may be relatively inconspicuous. If,
however, religion plays a large part in the legitimation of the chief's power, as it often
does in more complex chiefdoms, the chief might seek to demonstrate both power
and authority by building magnificent religious structures.
The role religion plays in the legitimation of a chief's power depends partly on
the size of a chiefdom and partly on historical factors. In a small simple chiefdom,
the chief's power may need little legitimate support beyond his own competent
management of that power and his social rank. Religious structures in these cases
may be relatively inconspicuous.
If the chiefdom grows, the chief may find it more difficult to manage
competently because of the greater complexity. Similarly, if stress occurs (several
years of bad crops, for example), the chief's leadership abilities may be called into
question. In either case, challenges to the chief's power increase. These challenges
lead the chief to shore up power with enhanced authoritative claims or actions (cf.
Cohen 1988:19). Claiming the support of deities is an excellent way to do this.
Advertising this support by the construction of conspicuous religious structures
communicates to the followers and potential challengers an increased legitimacy
(divine support). The nature of religious structures thus gives us information about
the dynamics of political power and authority. Large elaborate structures may
suggest some sort of stress within the chiefdom (due, perhaps, to growth or economic
hardships) which is handled by concentration of power in the hands of the leader
(centralization). Small structures suggest a polity in which stress is dissipated by
other factors (out-migration, for example, or periodic fissioning of the polity).
Interpreting religious structures in the archaeological record. The lack of clear-
cut patterning regarding religious structures should lead to caution when interpreting
archaeological sites. Even in capitals, large and elaborate structures may not have
had the "ceremonial" use often attributed to them. In fact, given the data provided by
this survey, a large, ostentatious, prominently located building is far more likely to
have been a chief's house than a religious structure. This is not to say that large
religious structures do not occur. They do, particularly in larger chiefdoms. But their
presence has more to do with the fluctuating interactions of power and authority than
with the complexity of the chiefdom.
Communal structures, particularly men's houses, frequently function as
settings for the conduct of government in ordinary villages in both egalitarian
societies and chiefdoms (see, for example, Barth 1965:53-56; Barnett 1960:32). In a
capital, the chief's house often takes over this function; chiefs' houses also act as social
gathering places for important people of the vicinity. Perhaps for this reason,
communal houses occur infrequently in capitals, though they are not completely
In 14 of 30 cases, communal structures are clearly or apparently absent. They
are clearly present in only eight cases. These eight cases have a worldwide
distribution (Apalachee, Bamileke, Alur, Ovimbundu, Bemba, Palau, Pohnpei, and
Tahiti). In two cases (Palau and Bamileke), and perhaps in others, more than one
communal structure may be found in a capital.
Size. Only three cases provide data regarding the size of communal houses.
In Palau (Morgan 1988:18-20) a communal house is apparently larger than the chief's
house, but in Bamileke (Fraser 1968:Plate 1) and possibly Tahiti (Beaglehole 1955:129)
the chief's house is larger. With mixed results and such a tiny sampling, it is
impossible to draw any conclusion about what to expect in terms of size of communal
structures in capitals.
Location. As with size, very little information exists regarding the location of
communal houses, but this time the pattern is clearer. Where location is noted (four
cases), communal structures are described as near the chief's house (Bamileke,
Pohnpei) or centrally located (Palau, Apalachee).
Comments. Communal structures may or, more frequently, may not be
present in chiefly capitals. When they are present, they are generally found in a
prominent location near the complex of chief's house and chief's storage. Size is not
now predictable; communal houses may be larger or smaller than the chief's house.
Because there are few clear patterns regarding communal houses, archaeologists
should exercise care in identifying them from the archaeological record. If small,
these structures may be taken for residences unless they have a unique character. If
large, they may be confused with chiefs' houses.
Public spaces, like communal houses, are features of non-capital villages as
well as capitals. Almost two-thirds (19) of the 30 areas yielded data about public
spaces. For the remaining 11 cases, there was insufficient information to determine
the presence or absence of public spaces. In three cases (Apalachee, Taino, Bamileke),
more than one public space may be present in the capital.
Public space is used for a variety of activities. Table 2-5 lists these activities
for areas where this information is given. The majority of activities relate to ritual or
government. The most frequently-cited function of public space is as a setting for the
conduct of governmental business; the public space is often where the chief settles
disputes and meets with advisors.
Table 2-5. Uses of Public Spaces.
Dancing Ritual Ritual Rituals in Conduct of Ordinary
(social or Athletic Proces- General Govern- Daily
ritual) Contests sions ment Activities
Powhatan Apalachee Calusa Mandari Mandari Mandari
Ovim- Taino Zulu Ovim- Alur Kiriwina
bundu Kiriwina bundu Ovim-
Kiriwina Kachin bundu
Tikopia Tikopia Bemba
Taino Tonga Swazi
Condition. The condition or upkeep of public space is noted in four cases. In
only one case is there mention of neglect: a Kachin public space, according to E.R.
Leach (1965:117), "presents a desolate appearance" when not being used for
ceremonies. In the other three cases, writers describe positive effort in the upkeep of
public space: Palauans pave the space (Parmentier 1987:56), and the Taino and
Mandari keep the public space swept and weeded (Wilson 1990:67; Buxton 1963:78).
Location. In eleven cases, the location of public space is determinable. In nine
of these, the public space lies near the chief's house; in the other two, the location is
described as "central." This central location is expectable in view of the stated ritual
and political functions of public space.
Comments. The results seen above lead us to expect the presence of one or
more public spaces in capitals and to predict that a public space will be prominently
located, near the chief's compound. The presence of a public space should not,
however, be used to identify a capital in the absence of other data, because many
types of settlements contain public spaces.
Walls are sometimes a feature of chiefly capitals. Under the general term
walls are included fences (ranging from flimsy screens to stockades), earthen
embankments (possibly accompanied by ditches), stone walls of various sizes, and
even mere visual separators (such as the low wall surrounding Tahitian maraes
[Ferdon 1981:55]). Walls may be internal (within the confines of the settlement) or
external (encompassing the settlement).
Walls are put up for a variety of reasons. These reasons may be grouped into
two categories: symbolic/social and practical. Symbolic or social reasons include
segregation of one part of the population from another, marking of social boundaries,
definition of restricted areas, and communication of status. Among the practical
reasons for walls are defense, security from wild animals, and containment of
Walls built for symbolic/social reasons are almost always internal walls (see
Table 2-6), while walls with practical purposes are generally external.
Table 2-6. The Nature of Walls.
Area Symbolic Practical Area Symbolic Practical
Nabdam Comogre yes
Alur yes no Nabdam -
Ovimbundu Ovimbundu yes yes
Ila -yes Ila yes
Yao yes W. Malays yes
Swazi possibly South Nias yes
Zulu possibly yes Kosrae yes
Internal walls. Internal walls occur in 12 areas. In six other areas internal
walls are clearly absent from the capitals. Internal walls are generally flimsy,
incomplete, or otherwise inadequate as practical barriers. It should be no surprise
then that in most cases where purpose was determinable, internal walls have
symbolic/social significance rather than practical uses (see Table 2-6).
Among the Alur, for example,
the chief's homestead is referred to as kal which means "fence." This
refers to the fence of criss-cross elephant grass which only those of
chiefly descent could use .... Neither the chief's nor any other Alur
homesteads ever seem to have been highly fortified, and the kal was of
purely symbolic importance (Southall n.d.:77).
In Pohnpei, a small structure within a two-to-three-foot high wall stood near
the communal house. Behind this wall sat the highest ranking chiefs (Riesenberg
1968:68), separated symbolically from the populace.
In two cases where internal walls (fences) serve a practical purpose, they pen
livestock (Ila, Zulu).
External walls. In contrast to the largely symbolic nature of internal walls,
external walls, when present (seven cases) and the purpose is determinable, all have
practical purposes (see Table 2-6), primarily defense against people or animals and
holding of livestock. For example, stone walls, moats, or cactuses protected various
Panamanian capitals from invaders or wild animals (Helms 1979:9). Similarly, among
the western Malays, capitals usually had a stockade for defense against raiders and
ditches that served both as defensive works and to prevent buffalo from wandering
into the rice fields (Gullick 1965:29).
It is worth noting that the absence of external walls is far more frequent in the
survey than is their presence. Nearly half the areas (14 cases) specifically lacked
external walls. The absence of external walls probably results from the prevalence of
a dispersed form of settlement in the 30 areas.
Comments. The survey indicates that archaeologists should not necessarily
expect to find walls in chiefly capitals. When we do, we can predict that external
walls serve a practical purpose while internal ones generally (though not invariably)
have symbolic or social significance.
Bruce Trigger (1990:122) has suggested that greatly elaborated fortifications
(beyond normal defense needs) displayed the power of the leader not only to his
subjects but to invaders. The presence of fortifications thus gives us information on
the nature of power in the chiefdom as well as on the particular historical situation.
The presence of internal walls may help in defining spaces reserved for nobles
and thus yield data on the nature of authority in the chiefdom.
Monumental architecture encompasses construction in which "scale and
elaboration exceed the requirements of any practical functions that a building is
intended to perform" (Trigger 1990:119).
What is the point of this unnecessary expenditure of effort? Simply stated,
monumental architecture advertises the power of the leader who built it. Power, as
seen in Chapter 1, is the ability of a leader to control the actions of others. The more
people a leader can muster, the more powerful he is. People are power.
The power of a leader then can be directly measured by the number of his
followers. This precept is clearly displayed in the chiefdoms of this survey,
frequently in explicit statements. For example:
[An Alur] chief's power depends ultimately on the number and size of
his subject groups (Southall n.d.:188).
[A Bemba] chief's reputation depends largely on the size of his capital
Political power in the Malay States rested on the control of manpower
[In south Sulawesi,] a large kapolo [or following] formed the substance
and source of influence .... To have "influence" and "power" in
Western terms, in short, one needs someone to influence, and that was
the kapolo (Errington 1989:102).
[In Indonesia,] when a village is oppressed by its radja and wrongfully
treated, the members leave and place themselves under the protection
of a neighboring radja, who always receives them with open arms since
they strengthen his power (Loeb 1935:38-39).
That people are power is recognized explicitly by followers themselves. The
Mandari and Yao peoples both have maxims to that effect. The Mandari say, "He is
no longer chief; a man without people cannot be so" (Buxton 1963:70). For the Yao, "a
chief who has no people is not a chief' (Mitchell 1956:109). A chief who ignores this
precept will soon find trouble. An ancient Pohnpeian chief discovered this too late.
Paul Riesenberg (1968:51) recounted the story:
A chief who acted too unilaterally might come to grief. A [chief] in
precontact times is said to have made numerous decisions without
consulting his subjects, who rose and marched against him. He called
upon his own clansmates to help, but he had alienated them too and
they did not respond to his plea: finally, he was killed.
If a leader wishes to advertise his power (to foreign or internal rivals, for
example), it follows that he will do it through some means that demonstrate his large
following. This can be accomplished literally and directly by amassing large groups
of people in his support (the masked processions of the Calusa or the huge gatherings
of warriors of the Zulu). It can also be accomplished indirectly by acts which
demonstrate the ability to mobilize a large labor force. The construction of
monumental architecture falls into this latter category.
Monumental architecture in chiefdoms. In the survey sample, monumental
architecture is present in eleven cases; it is absent in ten cases. Nine cases provide
insufficient information to determine presence or absence.
Table 2-7 shows the way these data break down. In the survey, the size or
complexity of a chiefdom seems to have little connection with the presence of
monumental architecture, but there may be a slight tendency for monumental
architecture to be absent more frequently in smaller chiefdoms than in large ones.
The survey results regarding the treatment of chiefs' houses (larger, more
decorated, prominently located) and religious structures (sometimes small, sometimes
large and prominent) lead to the expectation that monumental architecture in
chiefdoms will be primarily associated with chiefs' houses. And, indeed, chiefs'
houses are monumentalized most frequently (6 of 11 cases), but religious structures
run a close second (5 of 11 cases) (some areas have more than one type of
monumental architecture). Other monumental structures also occur: communal
houses (2 cases), walls (2 cases), and freestanding stones (3 cases).
Table 2-7. Monumental Architecture in the 30 Chiefdom Areas.
pe Monumental Monumental
Type of Cases with Data Architecture Architecture
Minimal 7 3 4
Typical 8 4 4
Maximal 6 4 2
TOTAL 21 11 10
Comments. Monumental architecture is absent in almost as many chiefdom
areas of the survey as it is present in. The lack of monumental architecture, therefore,
is not an indication of lack of power. Power, after all, can be expressed in other
ways. Within a chiefdom area where monumental architecture does exist, however,
the labor involved in construction provides a useful comparative measure of power,
thus providing an insight into the dynamics of power and authority. Monumental
architecture in itself does not provide a measure of overall societal complexity,
although its presence is probably confined to non-egalitarian societies.
It might be expected, based on the above comments on power and the display
of power, that larger, more populous chiefdoms would exhibit monumental
architecture more frequently than smaller ones. I suspect that this statement would
prove accurate if additional data were available, but the limited data in this survey
make it untenable to draw such a conclusion.
In sum, monumental architecture embodies power. The particular structure
monumentalized adds information about the source of the builder's power. Sources
of power include persuasion, personal ability, control of valued resources, force, and
authority based on tradition, social rank, divine support, or a law code (see Chapter
1). Monumental fortifications thus express the might or force of the leader, whereas a
monumental chief's house reflects power based on persuasion or on social rank, and
monumentalized religious structures express power derived from the authority of
In state-level societies, these general interpretations may not apply. This is
because the nature of leadership in chiefdoms differs from that of leadership in states.
In chiefdoms, the political leader is also the economic and religious leader. In states,
especially post-industrial states, these realms may be segregated, with leadership held
by different individuals. Thus, for example, the monumentalization of a religious
structure in a state may not reflect reliance by the political leader on religious
authority. Rather, it may express the power of a particular subset of the population.
Before turning to a discussion of Mississippian capitals, the findings of this
chapter are presented below in a series of statements about chiefly capitals.
Exceptions to or variations on these statements may, of course, occur, and in some
cases additional or more detailed information on capitals may change the conclusions.
Notwithstanding these qualifications, the statements presented below accurately
reflect the results of this survey.
* The capital is the largest settlement (in population) in a chiefdom.
* The layout of a capital may be similar to or different than that of an ordinary
* Capitals are generally oriented practically, particularly to a feature of
topography, rather than symbolically.
* The chief's house is larger than ordinary houses.
* The chief's house is more decorated or elaborated than commoners' houses; it
may take a different shape or be constructed of different materials.
* The chief's house stands in a prominent location, e.g., in the center of the
settlement or on a high point.
* The chief's storage facilities are larger or more numerous than commoners'
* The chief's storage facilities stand in a prominent place, near the chief's house.
* The chief's storage facilities may be larger or smaller in plan than the chief's
* Religious structures in a capital may vary in size independently of the chief's
* Religious structures are not necessarily located near the chief's house.
* Communal structures appear infrequently in chiefdoms.
* When communal structures are present, they usually stand near the chief's
* Communal structures may be larger or smaller than the chief's house.
* One or more public spaces occur in the capital (and in non-capital settlements).
* A public space is located near the chief's house.
* Walls appear infrequently in capitals.
* When walls are present, external walls serve practical purposes, while internal
walls usually have social or symbolic significance.
* Monumental architecture does not occur in all chiefdoms.
* Where present, monumental architecture occurs in minimal, typical, and
* Chiefs' houses and religious structures are the most frequently
MISSISSIPPIAN CAPITAL VILLAGES, PART 1:
THE SIZE OF THE CAPITAL
In Chapter 2 some material correlates of chiefly capitals were presented, and
the ways in which power and authority are expressed in the size and structure of the
capital were considered. This chapter explores the size of Mississippian capitals.
Chapter 4 will consider issues relating to the structure of Mississippian capitals.
Identifying Mississippian Capitals
The first task is to identify Mississippian capitals. A capital is, by definition,
the residence of a supralocal political leader (see Chapter 1 for discussion of capitals).
The clearest evidence for identifying capitals (Mississippian or otherwise) thus comes
from a direct ethnographic or documentary statement similar to the following:
"Village X is the home of the high chief." To be useful to archaeologists, such a
statement must include a complete enough description to allow for recognition of the
site in today's landscape.
The physical characteristics of capitals also help to identify them. For example,
Christopher Peebles and Susan Kus (1977:432) and Colin Renfrew (1973:543) have
pointed out that a hierarchical settlement pattern invariably accompanies the political
hierarchy of chiefdoms. Indeed, the presence of a settlement hierarchy stands as one
of several archaeological correlates of chiefdoms (Peebles and Kus 1977:432). The sites
uppermost in the settlement hierarchy thus represent capitals. Data from the
chiefdom survey discussed in Chapter 2 of this dissertation support this view. The
capital is the largest site; it contains larger, more prominent, and more elaborate
structures, and it is the location of monumental architecture if any exists in the polity.
Evidence for identifying Mississippian capitals comes from both documentary
sources and archaeological research.
The best information on Southeastern capitals comes from descriptions of the
early eighteenth century Natchez. Moreover, the characteristics of Natchez capitals
are discernible in the modern landscape. Pierre LeMoyne d'Iberville (McWilliams
1981:125), visiting the area in 1700, says,
When 1 got to the landing, I sent a man to notify the chief of my arrival
.. I went to this village, which is 1 league from the edge of the
water .... Halfway there, I met the chief coming to meet me.... We
proceeded to his hut, which is erected on a 10-foot mound of dirt
carried there, 25 feet wide and 45 long. Close by it are eight huts.
Facing the chiefs is the temple. These form a ring somewhat oval-
shaped and enclose a public square about 250 yards wide and 300 long.
Writing some years later, Antoine Le Page du Pratz (1975:338) described the
chiefs quarters in the Natchez capital.
Strangers are then invited to dine with the Great Sun, and in the
evening there is a dance in his hut, which is about thirty feet square,
and twenty feet high, and like the temple is built upon a mount of
earth, about eight feet high, and sixty feet over on the surface.
Because the eighteenth century Natchez lived somewhat later than the
Mississippian period as described here, it is useful to examine the documentary record
for evidence from a time closer to the precolumbian period.
Two chroniclers of the de Soto expedition of 1539-1543, one who was present
(Luis Biedma) and one who interviewed participants (Garcilaso de la Vega), both note
the association of chiefs and physically high places. Biedma (Buckingham Smith
1968:251), in describing the placement of a cross on a mound in the town of Icasqui
(or Casqui), notes that
... it is the custom of the Caciques [chiefs] to have near their houses a
high hill, made by hand, some having the houses placed thereon ....
Garcilaso, who also mentions the mound at Casqui, provides a general description of
contact period Mississippian capitals (Varner and Varner 1951:170-171, 431).
... [T]he Indians of Florida [i.e., the Southeast] always try to dwell on
high places, and at least the houses of the lords and Caciques are so
situated . [T]hey build such sites with the strength of their arms,
piling up very large quantities of earth and stamping on it with great
force until they have formed a mound from twenty-eight to forty-two
feet in height. Then on the top of these places they construct flat
surfaces which are capable of holding the ten, twelve, fifteen or twenty
dwellings of the lord and his family and the people of his service, who
vary according to the power and grandeur of his state.
Although Garcilaso is known to have exaggerated numbers and dimensions,
his general statement rings true (compare it, for example, with the statements for the
later Natchez). Documentary evidence thus suggests that flat-topped mounds with
chiefs' houses on top were integral features of Mississippian capitals.
Archaeological evidence indicates that Mississippian sites fall into several
general categories based on the presence or absence of earthen platform mounds and
the extent of the occupation area, although there are, of course, local variations (see
Steponaitis 1986:390; Smith 1978b, 1985:75,77, 1986:62).
Most Mississippian sites are small occupation areas of less than 1 ha.
Excavations often reveal the presence of one to four houses accompanied by
outbuildings (Smith 1978a, 1978b; Scarry 1989; Green and Munson 1978; Morse and
Morse 1983:238-239). Archaeologists generally interpret these as farmsteads occupied
by one or two families. Some excavated examples include the Gypsy Joint site in
southeast Missouri (Smith 1978a) and the MacArthur site in southeast Arkansas
Other sites contain evidence of higher populations with remains of five to
fifteen structures, sometimes including a larger structure interpreted as a communal
building (Ham 1978:254; Milner et al. 1984:186; Green and Munson 1978:310;
Rolingson 1976:110, 113). This site type is generally termed a hamlet. The Borrow Pit
site in northwest Florida (Jones 1990:83; Payne 1982; Shapiro and McEwan 1992:264-
265), though only partially excavated, presents a good example of a hamlet with its
four houses arranged in an arc flanking a larger communal structure.
A third site type consists of a large occupation area with more than 15
structures. Termed villages or towns, these are often fairly compact and frequently
surrounded by a fence or an embankment. The Turner and Snodgrass sites in
southeast Missouri are excellent examples of this site type (Price 1978:218-219).
A fourth site type contains one or more platform mounds in addition to a
residential area. The size of the residential area varies from farmstead to town size.
Archaeologists sometimes subdivide this category into sites with one platform mound
and sites with two or more mounds (cf. Fowler 1978:468-471; Steponaitis 1978:437).
Platform mounds range in height from 0.5 m to 30 m, and most can reasonably be
described as monumental architecture (i.e., unnecessarily large and elaborate
constructions, the building of which is beyond the scope of a single household [cf.
Trigger 1990:119; Peebles and Kus 1977:432]).
Mississippian sites can thus be arranged in a hierarchical fashion from the
smallest (farmsteads) to the largest (multi-mound centers). Moreover, the largest sites
contain large and prominent constructions that can be described as monumental
On Using Platform Mounds to Identify Mississippian Capitals
To summarize the evidence presented above: (1) direct statements in
documents from the early contact period clearly link Mississippian chiefs and their
capitals with platform mounds; (2) archaeological research reveals a hierarchy of
Mississippian sites-farmsteads, hamlets, villages, single-mound sites, and multi-
mound sites--with mound centers standing at the top of the hierarchy; and (3) most
platform mounds can be classed as monumental architecture.
It seems safe to say, given this evidence, that sites with platform mounds
represent Mississippian capitals, especially if accompanied by a large occupation area.
Such sites embody in their construction the presence of or the control of large
numbers of people.
The presence of platform mounds is a more useful indicator of precolumbian
Mississippian capitals than a characteristic such as size of the residential area for
several reasons. First, platform mounds occur in all areas of the Mississippian world.
Second, population density varies across the Mississippian world. In general,
Mississippian settlement pattern includes both dispersed and nucleated patterns
(Smith 1978b:489-490, 1985:75-77; Muller 1986:173-174). As a result, some chiefly
capitals may not have had large resident populations, while others may have been
highly nucleated. The size of the residential area thus is not comparable across the
whole area. Third, determining the size of residential areas requires extensive
archaeological investigations. Acquiring this information for large numbers of sites is
highly impractical. Conversely, data on mounds are often recorded during even the
most cursory research. The presence of platform mounds is therefore highly
Given the foregoing information, Mississippian capitals then can be identified
with some confidence as sites with one or more platform mounds.
Mississippian Mound Center Survey
Having identified Mississippian capitals, the next step is to collect information
about the physical attributes of mound sites just as was done for capitals in
ethnographically known chiefdoms (see Chapter 2).
Goals of the Survey
The primary goal of the survey is to establish a database that includes basic
information about the physical characteristics of Mississippian mound centers. It is
particularly important to acquire data that the chiefdom survey indicated was useful
in studying political structure.
Contingent on the completion of the first goal, the second is to compare
mound centers in terms of size and structure attributes (see Chapter 4) and to
interpret any patterns discerned.
The third goal is to establish frameworks (or classifications) within which to
discuss Mississippian mound centers. The classifications presented in this chapter and
the next have been designed strictly for heuristic purposes. They are intended not
necessarily to answer questions but to bring to light new or more useful questions
and to suggest directions for continuing research. The frameworks also provide a
common language for discussing various aspects of the size and structure of the
mound centers. These classifications are not intended to confine the discussion of
mound centers to delimited categories.
The goals of this survey then are broad, synchronic ones. They are primarily
descriptive and comparative and encompass the entire Mississippian time period. An
understanding of change through time is not the primary purpose. As a result of this
broad focus, some of the details in the picture of Mississippian capitals may be lost.
This loss is compensated for by the creation of a broad comparative base on which to
build and, most importantly, by the establishment of a Mississippian-worldwide
perspective within which to examine political capitals.
Mound Center Database
The sites included in this survey are Mississippian sites with one or more
earthen platform mounds. The resulting database includes information on 536 mound
centers, although not all information was available for all sites. The database includes
Mississippian sites from all parts of the Southeast and Midwest. The geographic
extent of these sites effectively shows the extent of the Mississippian world (see
Some constraints existed due to the nature of the resource base. For example,
while it was possible in the chiefdom survey to record attributes of the chief's house
directly, such information is not directly accessible in the archaeological record. Data
recorded in the mound center survey, therefore, comprise archaeologically recoverable
characteristics deemed to relate indirectly to features of the capital (e.g., size of the
main mound which relates to the size of the chief's house).
A second constraint occurs because the survey is a very broad one. In a
survey including more than 500 archaeological sites, the data collected necessarily
must be very visible or easily acquired (i.e., without extensive excavations).
Fortunately, because of the public nature of political structure, many informative
features of mound centers are highly visible, most notably the mounds themselves.
The remainder of this chapter deals with features related to the size of the
capital. Several measures of size were readily available, allowing for a detailed
discussion of this subject. Aspects related to the structure of mound centers are
considered in Chapter 4.
Measuring the Size of Sites
Size (population) of chiefly capitals was earlier (Chapter 2) shown to be an
important indicator of power. Recall that power is the ability to control the actions of
people (see Chapter 1). The more people a leader controls, the greater his power.
The size of a capital, then consists of the number of people resident there or,
alternatively, the number of people whose labor is commanded by a leader.
Archaeologists have used several measures to describe the size of mound
centers. Number of mounds is most frequently used, particularly the distinction
between single-mound sites and multi-mound sites (Payne 1981; Bell 1984:227-228;
Brown 1984:242-243; Smith and Kowalewski 1980:3, 5). Area of site has been used
occasionally (Smith 1987:68-72; Brown et al. 1978:190-192; Price 1978:213; Morse
1981:46; Lewis 1990:46) when that information is available from area excavations. In
the discussion below, a variant of area, mound precinct area, is examined. The height
of the main mound at mound centers has been used to describe the size of the site,
sometimes implicitly (Steponaitis 1986:390), sometimes explicitly (Brain 1978:340-341).
The total volume of earth in the constructions at mound centers also has been used
(Steponaitis 1978:446-448; Scarry and Payne 1986:82-83; Blitz 1993:46; Muller 1986:200).
Total volume is difficult to acquire for large numbers of sites, so volume of the main
mound has been substituted in the discussion below. Both the last two criteria and
probably the number of mounds measure not resident population but labor controlled
by the leader. Each of these four measures of size (number of mounds, mound
precinct area, height of main mound, and volume of main mound) will be discussed
and evaluated in more detail in the following sections.
Before turning to a discussion of the first measure, it is necessary to comment
on one difficulty of all these measures and indeed of any undertaking of this kind.
This is the problem of the longevity of sites. Since many Mississippian sites spanned
a hundred or more years (some as much as three or four hundred years), values for
any of the criteria mentioned above may be a product, at least in part, of the duration
of the site. In a broad scale survey like this it is virtually impossible to acquire
adequate data relating to small time spans for large numbers of sites. This problem
must therefore remain unresolved, and the reader, who is urged to keep in mind the
heuristic nature of the study, must remain aware of possible distortions in any
Number of Mounds
The number of mounds per site is a particularly accessible piece of
information. Out of 536 sites recorded, this information was available for 467 sites
(87.1%). All 467 sites have at least one platform mound, but the number of mounds
per site includes other kinds of mounds as well.
The number of mounds per site ranges from 1 to 100, although only 7 sites
(1.5%) have more than 16 mounds. Figure 3-2 illustrates the frequency of sites by
number of mounds per site.
The most frequently occurring number of mounds is 1, constituting 47.1% (220
sites) of the total. The actual figure was probably far higher than the 220 recorded
here. Many small mound sites were undoubtedly destroyed or eroded away without
any recognition of their significance. Indeed, many single-mound centers could still
be extant but unrecorded. The piechart in Figure 3-3 illustrates the predominance of
single-mound sites, as well as the high proportions of sites with two and three
The mean number of mounds per site is 3.2. This figure is quite small
compared to figures for the most familiar mound centers (e.g., Moundville: 20
mounds; Angel: 11 mounds; Spiro: 9 mounds). This suggests that the current picture
of mound centers is skewed by research emphasis on the largest sites.
Even smaller than the mean is the median: 2 mounds per site. This figure
again emphasizes that most mound centers are very small. To get a better handle on
the shape of the distribution, it is useful to construct a box-and-whisker plot.
A box-and-whisker plot is an Exploratory Data Analysis (EDA) technique that
provides a visual summary of the spread of a distribution (Shennan 1988:44-46;
Hartwig with Dearing 1979:23-25). This technique relies on the median and
midspread (or interquartile range), measures regarded as more resistant (i.e., less
sensitive to a few extreme values) than mean and standard deviation (Hartwig with
Dearing 1977:19, 21).
Number of Sites
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19+
Number of Mounds per Site
Figure 3-2. Frequency of Sites by Number of Mounds: Bar Graph.
Number of Mounds per Site
Figure 3-3. Frequency of Sites by Number of Mounds: Pie Chart.
In a box-and-whisker plot (see Figures 3-4 and 3-5), the central vertical line
marks the median (a cross indicates the mean). The sides of the box lie at the lower
and upper quartile values (or hinges). The box therefore encloses 50% of the cases in
the sample. The "whiskers" (the horizontal lines attached to the box) indicate the
values outside the middle 50% but within 1.5 times the midspread (the difference
between the upper and lower hinges). Values outside the range marked by the box
and whiskers are plotted individually and referred to as outliers. Outliers more than
3.0 times the midspread beyond the hinge are designated by stars.
Box-and-whisker plots are valuable in detecting asymmetry in a distribution
and in identifying outliers or extreme values, the nature of which can be informative.
Box-and-whisker plots provide information about the bulk of the distribution (the
values in the middle) as well as providing detail about the tails of distributions with
extreme values (Hartwig with Dearing 1977:21, 23).
The box-and-whisker plot in Figure 3-4 shows that values of 9 or more
mounds are extreme and do not describe the bulk of the data, 50% of which falls
between 1 and 4 mounds per site. However, the outliers, especially the most extreme
value, 100, overwhelm the graph, making the pattern difficult to see. Removing the
highest value temporarily from the database makes it possible to see the details of the
box plot a little more clearly (see Figure 3-5). Note that removing the highest value
does not change the box, the whiskers, or the outliers (except to eliminate "100"); it
just expands the figure, making it more legible.
Sites with five to eight mounds (the area enclosed by the "whisker") have
sometimes been regarded as medium-sized sites (Brain 1978:341; Payne 1989). In the
context of the distribution, however, these sites can be seen to be considerably larger
0 20 40 60 80 100
Number of Mounds per Site
Figure 3-4. Number of Mounds per Site: Box-and-Whisker Plot.
0 10 20 30 40 50
Number of Mounds per Site
Figure 3-5. Number of Mounds per Site (excluding Cahokia). Box-and-Whisker Plot.