Social Theory in Archaeology 0874806410, 0874806429

Citation preview

Social Theory in Archaeology

Edited by Michael Brian Schiffer

THE

UNIVERSITY OF UTAH SALT LAKE CITY

PRESS

FOUNDATIONS OF ARCHAEOLOGICAL INQUIRY James M. Skibo, editor Copyright © 2000 by the University of Utah Press

All rights reserved co Printed on acid-free paper

Manufactured in the United States of America O06

O§

LIBRARY

O04 OF

O03

O2

OL

CONGRESS

OO

.

§

4

3

2 =F

CATALOGING-IN-PUBLICATION

DATA

Social theory in archaeology / edited by Michael Brian Schiffer. p. cm. — (Foundations of archaeological inquiry) Includes bibliographical references and index. ISBN 0-87480-641-0 (acid-free paper) — ISBN 0-87480-642-9 (pbk. : acid-free paper) 1. Social archaeology. 2. Archaeology—Philosophy. 3. Social sciences—Philosophy. I. Schiffer, Michael B. II. Series CC72.4.563

930.1—dc21

2000

99-088528

Foundations of Archaeological Inquiry James M. Skibo, series editor

VOLUMES

IN THIS

SERIES:

Expanding Archaeology

James M. Skibo, William H. Walker, and Axel E. Nielsen

Behavioral Archaeology: First Principles

Michael Brian Schiffer

Evolutionary Archaeology: Theory and Application Edited by Michael J. O’Brien

Unit Issues in Archaeology: Measuring Time, Space, and Material

Edited by Ann FE Ramenofsky and Anastasia Steffen

Pottery Ethnoarchaeology in the Central Maya Highlands Michael Deal

Pottery and People: A Dynamic Interaction Edited by James M. Skibo and Gary M. Feinman Material Meanings: Critical Approaches to the Interpretation of Material Culture Edited by Elizabeth S. Chilton

Social Theory in Archaeology

Contents

Preface and Acknowledgments _ vii . Social Theory in Archaeology: Building Bridges MICHAEL

BRIAN

1

SCHIFFER

. Revisiting Power, Labor Rights, and Kinship: Archaeology and Social Theory JEANNE

E. ARNOLD

. Corporate/Network: New Perspectives on Models of Political Action and the Puebloan Southwest Gary M. FEINMAN . Abandonment: Conceptualization, Representation, and Social Change MARGARET

C. NELSON

. Elements of a Behavioral Ecological Paradigm for the Study of Prehistoric Hunter-Gatherers ROBERT L. KELLY

52

79

. On What People Make of Places: A Behavioral Cartography NIEVES

31

63

. Food, Lies, and Paleoanthropology: Social Theory and the Evolution of Sharing in Humans STEVEN L. KUHN AND CATHERINE SARTHER MARIA

14

ZEDENO

97

. Strange Attractors: Feminist Theory, Nonlinear Systems Theory, and Their Implications for Archaeological Theory r1r2 SUZANNE M. SPENCER-WOOD . Evolutionary Archaeology: Reconstructing and Explaining Historical Lineages MICHAEL J. O’BRIEN

IO.

R. LEE LYMAN

Reconfiguring the Social, Reconfiguring the Material

JULIAN THOMAS Il.

AND

143

Ideas Are Like Burgeoning Grains on a Young Rice Stalk: Some Ideas on Theory in Anthropological Archaeology 156 SUSAN

Kus

126

References Index

173

227

Contributors

237

Preface and Acknowledgments MICHAEL

BRIAN

In the fall of 1996 James M. Skibo, editor of the series Foundations of Archaeological Inquiry (FAI), published by the University of Utah Press, invited me to organize the second FAI “roundtable.” FAI roundtables provide the occasion for a dozen or so scholars to assemble during several days, with few distractions, to discuss timely topics of mutual interest. Discussions focus on each participant’s paper in a seminar-style format involving the

SCHIFFER

believed that it was an opportune time to address some long-standing issues. What strikes me most about social theory in the archaeology of the 1990s is its extreme diversification and fragmentation. Since the advent of processual archaeology in the 1960s, social theory has veritably exploded, with new ideas introduced—almost daily it seems—from sundry disciplines ranging from philosophy to physics. Accompanying this diversification is a near-debilitating fragmentation, which is manifest along two dimensions. The first dimension is theoretical programs, such as processualist, selectionist, and postprocessualist; among theorists of the major programs there is little communication, much less mutual respect. A second dimension of theoretical fragmentation is in the levels of societal complexity about which archaeologists theorize. Although the neo-

give-and-take of civilized scholarly debate.

The first FAI roundtable had just taken place at Illinois State University; organized by James M. Skibo, its subject was “Pottery and People.” I had been invited to that session and had prepared a paper, but a scheduling conflict prevented my participation in the roundtable itself. Aware that the first roundtable had been a great success, I eagerly ac-

cepted Jim’s generous invitation to organize a

second—on any topic in archaeological method or theory. I chose “Social Theory in Archaeology.” Social theory might seem like an odd choice for me, since I am usually associated with other bodies of method and theory. Indeed, my contributions to social theory had been, until recent years, sporadic at best. However, as editor of several serial publications (among these Archaeological Method and Theory and Journal of Archaeological

evolutionary

terms

band,

tribe,

chiefdom,

and state have for many investigators fallen out of fashion, the underlying cleavage planes based on complexity remain real. Those who study simpler societies seem to have little theory in common with those whose interests lie in complex societies. Always lurking in the back of my mind was the fear that archaeologists were traveling down the same road as sociocultural anthropologists. One day we might wake up and discover that there was no longer any common ground, no core set of concepts and

Method and Theory), 1 was somewhat famil-

iar with developments in social theory and

Vu

PREFACE

principles, to integrate researchers having varied theoretical interests and commitments. I hoped that we could, at all costs, avoid that unpleasant outcome, which has in recent decades compromised anthropol-

ogy’s credibility, both inside and outside the

academy. My thought was to bring together practi-

tioners from various theoretical programs— representing as well interests in different so-

Ten of the 12 papers presented at the FAI roundtable, significantly revised on the basis of the discussions and written com-

ments provided by participants and the organizer-editor, appear in the present vol-

ume. I hope that the reader will sense the ex-

citement we shared in jointly exploring the frontiers of social theory. Social Theory in Archaeology is intended,

obviously, to reach archaeologists interested

cietal levels—who might be willing to attempt to “build bridges” across theoretical

in social theory; in practice this means all archaeologists, for everyone employs social

cally sophisticated archaeologists whom I believed might share my vision of the need to

tions of past human phenomena. In addition

gulfs. Thus, I assembled a group of theoreticreate a more

theory, explicitly or implicitly, in explana-

many of the perspectives developed in the chapters that follow have profound implica-

integrated, coherent body of

social theory. In making my selections I also tended to favor colleagues who have ongoing substantive research projects—that is, they are consumers as well as producers of social theory—and those who would not be immediately dismissive of the theoretical work of others. In giving participants their charge I stressed that the aim of the roundtable was not to create artificial syntheses or integration but rather to promote meaty discussions of social theory in a context where one could not ignore the existence of other programs and other levels of societies. In such a context some bridges might be built. Indeed, many draft papers that were circulated prior to the roundtable sought seriously to cultivate previously unrecognized common ground. We assembled at Snowbird, Utah, a ski resort not far from Salt Lake City, on October 24, 1997, during a snowstorm. Sessions were held at a rectangular—not round—table on October 24 and 25. Each person offered a 10-minute précis of his or her paper, and discussions ensued, led by another participant. In my view the sessions were marked by an extraordinary degree of candor, civility, and sincere striving to reach new understandings. I cannot remember ever having enjoyed so much intellectual stimulation.

tions for the development and application of social theory across the social sciences, especially in sociocultural anthropology and sociology. Indeed, a future project surely must be that of finding ways to build bridges across many different disciplines. Jeff Grathwohl, Director of the University of Utah Press, deserves our inestimable gratitude for sponsoring the FAI roundtable on “Social Theory in Archaeology.” He is an editor of exceptional vision. Not only did he conceive and implement the FAI roundtables, but he actually attended most of our sessions. Local arrangements were handled, in exemplary fashion, by Glenda Cotter, Senior Editor of the University of Utah Press. Her preparations freed me and the participants to concentrate on theoretical discourse. I thank the participants of the FAI roundtable, all of whom I now consider friends, for making this experiment in archaeological communication so successful. As referees for the University of Utah Press, Philip G. Duke,

Kathleen D. Morrison, and Timothy R. Pauketat furnished helpful comments on a draft of the book manuscript. Finally, I thank James M. Skibo for inviting me to organize the roundtable and for his contributions to our discussions.

Vill

Social Theory in Archaeology

I

Social Theory in Archaeology Building Bridges MICHAEL

BRIAN

When I was a graduate student in the early 19708, I prided myself on having read nearly everything in archaeology, published in English, that was overtly a theoretical contribution. Although trained originally as a processualist (Schiffer 1995a), I had even digested the theoretical writings of culture historians (e.g., Willey and Phillips 1958). The claim by

one person to have mastered the entirety of

archaeology’s theoretical literature seems like an extraordinary display of hubris, for

doubtless I had missed much. Yet such a

claim would not have seemed outlandish at that time. Archaeology had been a small discipline, and few of its practitioners wrote copiously on theoretical topics. Thus, as recently as a quarter century ago, a person

could have mastered the literature of archae-

ological theory. Tucked into those scattered publications were formulations that today would be called, at least by some, “social theory.” Let me begin by furnishing a working defi-

nition of social theory (adapted from Schiffer

1988). Social theory consists of bodies of general knowledge about sociocultural phenomena expressed in postulates, premises, assumptions, principles, and models. Often having rich empirical implications, social theories ostensibly answer how and why questions about human behavior and societies. Under this definition virtually all theories that archaeologists use to explain behavioral and/or societal variability and change, in-

SCHIFFER

cluding theories of Darwinian evolution and behavioral ecology, qualify as social theory (for other views of social theory in archaeology, see, e.g., Preucel and Hodder 1996a; Shanks and Tilley 1988; Yoffee and Sherratt 1993). It is more accurate to speak of social theories, reserving the term social theory for contexts in which one distinguishes the domain of social theories from domains of theories having other functions, such as “formation process theory,” which contributes to creating proximate explanations of the properties of the archaeological record (Schiffer 1988).

There are countless kinds of specific phenomena for social theories to explain: from small-group interactions to the use of mediated communication in industrial societies, from diet choice among hunter-gatherers to the operation of multinational corporate networks that trade foodstuffs around the globe, and from the design of a flint knife to the functioning of an interconnected power grid on half a continent. It is my contention that to explain these diverse human phenomena, numerous theories are required, varying in level of generality, degree of abstraction, and empirical content. No single social theory can serve all explanatory needs in archaeology. Needless to say, many social theories in archaeology are shared with other disciplines throughout the sciences—from sociology to evolutionary biology—and humanities. Regrettably, archaeologists are far more familiar

MICHAEL

B.

SCHIFFER

with the theories built in other disciplines than the latter’s practitioners are with ours. It is my hope that this volume will begin to change that situation.

were presented at a small conference, sponsored by the University of Utah Press, at Snowbird, Utah, on October 24-27, 1997.

chaeology is large and abstruse, with perhaps hundreds of investigators contributing to the

chose as the subject “Social Theory in Archaeology: Setting the Agenda.” For two days participants vigorously discussed frontiers in social theory, with an eye toward “building bridges” that might begin to span

Today, the social-theoretic literature in ar-

theory-building and -borrowing enterprise.

Those who doubt this number are urged to peruse articles and their bibliographies published, during the past two decades, in journals and periodicals that have been receptive to social-theoretic contributions, including Advances in Archaeological Method and Theory, Archaeological Method and Theory,

Journal of Archaeological Method and Theory, Journal of Archaeological Research,

Cambridge Archaeological Journal, Archaeological Discourses, Norwegian Archaeological Review, Journal of Anthropological Archaeology—even American Antiquity and Antiquity. During the same two decades sundry topics in social theory have been the subject of books and monographs numbering in the many dozens. Although all archaeologists employ social theories, explicitly or implicitly, in their explanations of past phenomena, the recent growth of interest in building and debating social theory is long overdue and should be welcomed. There is, however, a dark side: an information explosion that burdens—indeed, threatens to bury—us all. No one today could read more than a fraction of the primary literature on social theory, and few seemingly attempt to read widely and deeply. That as individuals we conform closely to a popular caricature of the scholar—someone who learns more and more about less and less—is perhaps inevitable in the face of an exponentially growing body of complex and diversified knowledge. Accompanying this expansion of social theory has been a polarization of archaeologists and a proliferation of archaeologies that are seemingly symptomatic of a disintegrating discipline.

The present book is, in part, an effort to counteract the fissioning of social theory and the factioning of archaeology. The papers

As organizer of the second Foundations of Archaeological Inquiry (FAI) roundtable, I

the conceptual chasms that had formed in ar-

chaeology during the past few decades. Although many participants in the FAI roundtable accepted the challenge of building bridges, some rejected my premise that bridge building is desirable and feasible. Even

so,

every paper included in this volume contributes something provocative or significant to the enterprise of constructing social theory in archaeology and to setting the agenda for future social-theoretic research. In this chapter I touch on some of the factors that led me to pose building bridges as the central problem. I also enumerate several promising strategies that archaeologists can adopt for producing more integrated social theory. As is customary in edited volumes, I include in this introduction a brief précis of each of the remaining chapters. Finally, I offer a few concluding remarks on the advantages enjoyed by archaeologists in their attempts to build social theories that are truly general. THE

STRATEGY AND IDEOLOGY ECLECTICISM

OF

For some archaeologists, having to deal at all with social theory is a necessary evil, for they are most keenly interested in illuminating the past of a particular people or place. Thus, they tend to be impatient with debates about theory (social and otherwise), insisting that

any products of theoretical discussions above all be useful in the context of their own substantive, areal research projects. What is more, although aware that their explanations are infused with social theory (this had been made abundantly clear by Binford [1968a, 1968b] and Clarke [1973], among others), such archaeological practitioners tend not to

SOCIAL

THEORY

identify with particular theoretical programs. Instead, they maintain that the best strategy is judiciously to select, cafeteria style, from among potentially useful formulations offered by different theoretical camps or programs. The eclectic strategy is justified, I suggest, by an appealing ideology. The ideology of eclecticism appears to be based on two major tenets. First is the belief that human phenomena, being very complex, cannot be adequately comprehended by just one theory or theoretical program. Thus, like the proverbial three blind men and elephant, archaeological theorists can each describe only one part of a massive and complex whole. This much most of us admit. But it does not follow that comprehensive understanding can emerge only if we are selective, taking the “best” of each theorist’s or program’s offerings. Second, in view of the faddism that seems to afflict archaeological theory in general and social theory in particular, commitment to one theoretical program is seen as an inadvisable, short-term strategy that might doom one eventually to obsolescence as new theories and programs emerge. No one wants to be closely identified with a theoretical program that might soon fall out of fashion. Eclecticism is a compelling and sensible strategy for archaeologists who have no overriding interest in social theory per se. They can pick and choose among the products of theorists, cobbling together diverse ideas that seem to help in the solution of specific research problems. Thus, eclecticism makes it possible to do archaeology without having much familiarity with the primary literature of social theory, for guidance on what theories might be useful can be acquired from secondary and tertiary sources. Clearly, this is an attractive and efficient strategy for the prehistorian or historical archaeologist who wishes to avoid being overwhelmed by the explosion of social-theoretic writings. But it does not by itself lead to better and more integrated social theory or even to an improved understanding of the relationships among different social theories.

IN ARCHAEOLOGY

REDLINING

Regrettably,

THE SOCIAL-THEORETIC LITERATURE

eclecticism is not a strategy

open to archaeologists with a passion for theory building, for they are far too aware of the contradictions and incompatibilities that can arise from juxtaposing fragments from different theoretical programs (see Harris 1979 on the shortcomings of eclecticism). Yet archaeological theorists are not without strategies for dealing with a burgeoning social-theoretic literature. Many archaeologists cope with the information explosion by simply “redlining” large areas of the social-theoretic literature, judging which works can be ignored.‘ There are, I submit, two main criteria for redlining: theoretical program and level of societal complexity. | now examine each of these in turn and caution the reader that I overgeneralize somewhat in order to highlight major patterns. (Redlining, of course, also provides guidance for archaeologists employing the eclectic strategy.)

As everyone is painfully aware, archaeology in the 1990s is divided into programs, camps, or schools, each of which is tied implicitly and explicitly to specific bodies of social theory. The most prominent and well-defined theoretical programs are processual (a program that includes evolutionary/ ecological and cognitive varieties), behavioral, selectionist (also called evolutionary), and postprocessual (in varieties too numerous to list). It is easy enough for a principal theorist associated with one program to dismiss in a single utterance the theoretical writings of others. In this way followers of the first program have been given guidance for redlining the rest, freeing them from the burden of reading much of the _ primary social-theoretic literature. Thus, Binford’s ex-

plicit denunciations of behavioral (e.g., Binford 1981a) and postprocessual archaeologies (e.g., Binford 1987) make it easy for Binfordians to ignore the offending books and articles. Likewise, when Dunnell (e.g., 1978, 1980) rejects the reconstruction of past

behavior and organization and claims that

MICHAEL

B.

SCHIFFER

the only theory needed for explaining change is Darwin’s, other adherents of selectionism are apt to redline processual and behavioral

evant articles, books, and monographs. One cannot become familiar with the entire corpus of social theory, so focusing on what is

tique of the “extreme relativists” (Schiffer 1988) might be taken as a cue by behavioralists to pass up postprocessual writings. More

sensible. Redlining helps archaeologists carve out a much smaller segment of social theory

archaeologies. Similarly, my uncharitable cri-

subtle cues to redlining are provided by a the-

orist’s bibliographic citations and the invita-

tions he or she issues when assembling symposia, conferences, and edited volumes.

most closely related to one’s interests seems

over which they can claim expertise. It is but one more manifestation of specialization brought on by the expansion and differentia-

tion of knowledge, no different from tradi-

Preucel and Hodder’s (eds. 1996) reader on

tional redlining in archaeology based on areal specializations.

uously excludes behavioral and selectionist

adaptive response to the information explo-

archaeological theory, for example, conspic-

In support of what, I suggest, is mainly an

contributions, and even processual ones are rare. Such patterns of inclusion and exclusion

sion, archaeologists have adopted ideology that rationalizes these practices. Dismissing

which theorists and which theoretical writ-

justified by reference to Thomas Kuhn’s in-

encourage readers to draw conclusions about

ings should be taken seriously. Large segments of the social-theory literature are also redlined on the basis of level of societal complexity. Three principal levels of complexity are generally recognized, although the labels archaeologists apply to them differ somewhat:

(1) bands, which are

usually mobile hunter-gatherers, (2) interme-

diate

societies,

which

include

tribes

and

chiefdoms (Arnold, ed. 1996; Gregg 1991), and (3) complex societies—states and empires. Because most archaeologists work predominantly with societies at only one level of complexity, theoretical literature pertaining to the other two levels is easily redlined. For example, I suspect that archaeologists of complex societies do not avidly follow research on the technological organization of mobile hunter-gatherers (e.g., Nelson 1991). Nor are hunter-gatherer specialists likely to take a keen interest in the latest debates on the ideology of bureaucratic organizations. Clearly, level of complexity furnishes a criterion by which archaeologists can redline appreciable portions of the social-theoretic literature.” Redlining on the basis of theoretical program and societal complexity is, it seems to me, a rational response to the problem of information overload. It helps archaeologists to prioritize their reading of the primary literature in the face of a torrent of potentially rel-

the social theory of other programs is easily

fluential work on scientific paradigms. In The Structure of Scientific Revolutions, Kuhn (1970) suggests that competitive theories or paradigms can be so different at the level of fundamental assumptions that adherents have different worldviews and speak different languages. An implication of the Kuhnian thesis is that adherents of incompatible “paradigms,” who attempt to solve only the puzzles their paradigms pose, are incapable of engaging in reasoned debate with holders of different paradigms. Kuhn’s assertions of paradigm incompatibility permeate much of the archaeological, social science, and philosophical literatures, and doubtless all archaeologists have been exposed to the basic ideas. That being the case, It is easy enough for us to equate major theoretical programs with paradigms (e.g., for an early effort, see Martin 1971) and to relentlessly apply Kuhn’s apparently incontrovertible lessons. Kuhn’s thesis implies that an individual trained in one theoretical program can learn nothing useful at a theoretical level from advocates of another program. Thus, a selectionist easily comes to believe that postprocessual theory is devoid of value and vice versa; after all, the metaphysical, ontological, and phenomenological assumptions of these two programs seem worlds apart. By building on Kuhnian notions of paradigm incompatibility, then, archaeologi-

SOCIAL

THEORY

cal theorists can wave off the literature—and thus social theory—of other programs. Redlining, it would appear, is a reasonable response to the alleged conceptual incommensurability of theoretical programs. Ideology also justifies redlining with respect to levels of societal complexity. Culture historians had a body of theory—diffusionism—that could be applied to societies at all levels of complexity. However, faced with the disappointing results of attempting to employ systems theory and neoevolutionary formulations in archaeology, many processualists concluded, probably by the late 1970s, that no single body of theory could any longer handle societies at every level of complexity. And so began to emerge the doctrine that each societal level would require its own social theories. Thus, investigators of smallscale hunter-gatherer societies are properly concerned with theories, like optimal foraging, that obviously pertain to such societies and can justifiably overlook theories more appropriate for intermediate and complex societies, such as those treating redistributive economies or state religions. The ideology of level-specific theory helps archaeologists to rationalize the redlining of theoretical literature on societies at the “wrong” levels of complexity. Although redlining of social theory on the basis of theoretical program and societal level makes sense from the standpoint of an individual archaeologist striving to maintain mastery of narrow specializations, these practices are not likely to generate the best possible social theory. It is time, I suggest, to follow the past two decades of specialization with efforts at integration. Let us look at long-standing trends in areal specialization and synthesis for guidance. Since the advent of the discipline, archaeologists have carved out areal specializations. Over time the territories of individual archaeologists have grown smaller, doubtless in response to the greater number of practitioners, as well as the consequent accumulation of knowledge. Although extreme areal specialization has been the norm in archaeology throughout most of the twentieth century,3

IN ARCHAEOLOGY

from time to time there have been efforts to build conceptual bridges across the balkanized landscape. In the American Southwest, for example, Alfred Kidder’s pioneering synthesis, A Study of Southwestern Archaeology, tied together the findings of areal investi-

gations from the previous half century and helped to crystallize the modern concept of the Southwest as an archaeological culture area. This effort prioritized research questions, indicating fruitful directions that were pursued during subsequent decades of intensive specialized areal research. Similarly, Di Peso’s bold use of the Pochteca concept to bridge Mesoamerica and the American Southwest (e.g., Di Peso 1968, 1974), although not without problems (e.g., McGuire 1980), brought to the forefront questions about interregional interaction that are still being intensively addressed (e.g., Ericson and Baugh 1993). Clearly, a hypothesis worth investigating in more detail is that synthetic, crosscutting formulations have had a salutary effect on areal research. Although I have previously applauded specialization in archaeological theory (e.g., Schiffer 1978, 1988), the process may be going too far. My fear, expressed elsewhere (Schiffer 1996), is that we will become permanently organized into theory-based enclaves having largely impermeable boundaries. In this divided state, I suggest, archaeology would be hampered in achieving its historical and nomothetic goals. There is another way. BUILDING

BRIDGES

By analogy with areal specializations and syntheses, I suggest that social theory in archaeology is ripe for efforts at bridge building. Indeed, the time has come to counterbalance the seemingly inexorable fissioning of social theory brought on by the information explosion and consequent redlining. “Building bridges” is a strongly evocative, attractive metaphor, but what does it really mean? By bridge building I mean, in general, the construction of any conceptual integument that can connect previously disparate theoretical formulations or relate, theoreti-

MICHAEL

cally, formerly discrete phenomena. Needless

to say, this definition emphasizes that theo-

ries are invented, not discovered; thus, the archaeological theorist assumes the responsi-

bility for exercising creative thought in attempts at bridge building. He or she must de-

vise new concepts and principles for forging unexpected connections and relationships. Bridging Theoretical Programs

At its most ambitious, bridge building rejects the Kuhnian doctrine (perhaps dogma) that

different paradigms are conceptually incom-

mensurable and seeks to create connections between two or more theoretical programs.

As I have discovered in my effort to explore

relationships between behavioral and evolutionary archaeologies (Schiffer 1996), bridge building at this level is fraught with peril, even between programs

that seem to share

important tenets. Despite the selectionists’ less than lukewarm reaction to my formulations (Lyman and O’Brien 1998; O’Brien et al. 1998), I remain convinced that bridging theoretical programs can succeed and that continuing efforts are justified. Admittedly, I am willing to define success quite modestly. Aiming for the coalescence of two programs, for example, strikes me as unrealistic. More feasible is crafting a few conceptual tools potentially useful in both programs, such as the concept of “stimulated variation,” which I suggested could get selectionists and behavioralists on the same page when it comes to explaining spurts of behavioral variation (Schiffer 1996). In her chapter of this volume Jeanne Arnold offers the useful distinction between “small-scale” and “intermediate hierarchical” societies based on how leaders call on the labor of others. This distinction and the discussions and evidence that justify it seem well suited to bridge, albeit somewhat narrowly, processual and postprocessual approaches to organization. Another realistic goal is to identify shared high-level assumptions and develop their implications. For example, both selectionists and postprocessualists emphasize historical contingency and focus on individual agents. These seemingly superficial similarities might

B.

SCHIFFER

be fortuitous, but their existence is nonetheless tantalizing; diligent effort might forge deeper connections between the programs. There are some obvious places to begin additional bridge-building operations across

theoretical programs.

Archaeologists

inter-

ested in cultural evolution are now divided

into processual and selectionist camps, and the rhetoric between their advocates has re-

cently heated up (Lyman and O’Brien 1998; Boone and Smith 1998; Spencer 1997; pa-

pers

in Maschner

1996

and

Barton

and

Clark 1997; see also Kelly, and O’Brien and

Lyman, this volume). Processual evolutionists emphasize that adaptations arise from

problem

solving,

whereas

selectionists

lay

stress on undirected processes of natural selection that winnow cultural variability. Some bridge building is clearly called for here; I see no grounds for believing that these varieties of evolutionary archaeology must remain at odds. Similarly, both cognitiveprocessualists (e.g., Mellars and Gibson 1996; Mithen 1996a; Noble and Davidson 1996; Renfrew and Zubrow 1994) and postprocessualists (e.g., Hodder et al. 1995; Leone and Potter 1989) acknowledge the importance of knowledge and meaningful phenomena; the former is aligned with psychology and cognitive science, the latter with humanistic, interpretive approaches. Are these views on cognitive phenomena utterly incompatible? Perhaps not, but we will need concerted efforts at bridge building to find out.

Another bridging strategy might be to construct a metatheory that subsumes some formulations from two or more programs. In so doing one begins not with an existing set of shared assumptions but a common problem. The investigator then creates a new theory—a metatheory more general than existing theories in any one program— to solve that problem. This metatheory, however, must be capable in principle of subsuming existing theories pertaining to that problem. For example, virtually all theoretical programs countenance interest in ritual and religious phenomena, and some specific theories have been advanced. I sug-

SOCIAL

THEORY

gest that effort might be profitably directed now at building an archaeological metatheory of ritual and religion. A successful metatheory would strive to integrate, for example,

the

ideational,

symbolic,

affective,

behavioral, material, and spatial aspects of religious phenomena. Needless to say, such a theory could potentially have a large impact both within and beyond the borders of archaeology. In her chapter of this volume Margaret Nelson raises the possibility that abandonment processes can also serve as a fillip to building metatheories. Her own formulations clearly reach across to practitioners of diverse archaeological programs. At the very least, attempts to bridge theoretical programs would force a few archaeologists—bridge builders, their detractors, and curious onlookers—to venture into theoretical literatures that they had previously redlined. This might have the effect of fostering greater respect, if not understanding, among practitioners of different programs, perhaps elevating the civility of theoretical debate. For many archaeological theorists, even the boldest, the challenge of bridging major programs can seem overwhelming. What is more, because principal spokespersons of the programs are unlikely to recognize, much less acknowledge, successful bridge-building efforts, there will be but sparse rewards for erecting the giant spans.4 Thus, I reluctantly conclude that bridging theoretical programs will proceed at a plodding pace. But bridge builders also have less grandiose options. Bridging Levels of Societal Complexity Synthesis and integration of social theory can also occur across levels of societal complexity; crossing these chasms should not be nearly as difficult, and several strategies and successful examples are available. One strategy is to extend the reach of a theoretical program to societies at different levels of complexity. For example, one might conclude on the basis of their literatures that evolutionary formulations (processual as well as selectionist) are best suited, even lim-

ited, to small-scale hunter-gatherer and intermediate societies. Similarly, to date there

IN ARCHAEOLOGY

have been precious few applications of postprocessualism to hunter-gatherers. Here, then, are some opportunities for bridge building. A second strategy is to formulate theories, models, and laws that apply to specific behavioral and social processes that themselves crosscut societal levels. A recent synthesis by Hayden (1998) exemplifies the strategy of constructing process-specific models. Drawing on much previous work, he argues that every human society having in excess of 200 to 300 members will contain small numbers of “aggrandizers,” individuals who are much more aggressive and acquisitive than others. In resource-poor societies aggrandizers are held in check by social mechanisms such as witchcraft accusations. However,

where re-

sources are concentrated and abundant, permitting the accumulation of “surpluses,” aggrandizers are able to engage in a variety of projects that exploit labor and resources. These projects can result in the construction of monumental architecture, the development of new technologies (e.g., metallurgy, ceramics, and oceangoing vessels), and the es-

tablishment of far-flung exchange networks. Ultimately, through the actions of aggrandizers operating in favorable environments, some societies become more differentiated and stratified. Thus, Hayden’s provocative formulations apply to societies at diverse levels of complexity, from hunter-gatherers to states, so long as the aggrandizing process is unconstrained. It appears that the aggrandizer model can contribute to explaining some increases in societal complexity.

Archaeologists who believe that our discipline’s highest calling is the explanation of change (behavioral, social, cultural, technological, etc.) should be eager to fashion process-specific models, for they identify factors that can generate societal variability and change. What is more, by identifying processes that crosscut diverse societies, process-specific models help to bridge levels of societal complexity. Precisely these kinds of bridge-building efforts might be the most fruitful in the decades ahead. Indeed, much

progress can be made as we continue to move

MICHAEL

away from omnibus typologies of societal complexity in favor of types of processes (Schiffer 1995a; Walker et al. 1995). The chapters included here by Jeanne Arnold and

Gary Feinman focus on specific political processes and are noteworthy examples. Other Bridging Strategies Fortunately, the examples furnished above do not exhaust the possibilities for building bridges across the variegated terrain of social theory in archaeology. Indeed, other ap-

proaches to bridge building, less easily clas-

sified, are available to the creative archaeological theorist; I mention only two. A strategy that shows great promise, if we can judge by the papers employing it in the present volume, is that of conjoining two or more specific theories of diverse intellectual parentage to illuminate specific phenomena. Superficially, the conjunction strategy resembles the eclectic strategy discussed above, but the two approaches are worlds apart. Conjoiners, being theoretically sophisticated, are able to recognize and make explicit the specific strengths and weaknesses of two or more theories for solving particular problems. Indeed, they are able to show how the different theories, when conjoined, furnish new insights (see, e.g., the chapters by Kuhn and Sarther, and Spencer-Wood herein). Perhaps employment of the conjunction strategy can lead eventually to the development of new theories different from their progenitors. A second strategy is to formulate new concepts or variables that call attention to heretofore unappreciated dimensions of societal variability (see Rathje and Schiffer 1982 on “behavioral properties”). These concepts often have the power to reveal resemblances across societies at different levels of complexity and to promote the reformulation of explanatory questions. Potentially, new concepts could also exhibit utility in more than one theoretical program.

Cases in point are

provided in this volume by Gary Feinman and Jeanne Arnold. In addition, the concepts of “power with” and “power under,” offered by Spencer-Wood, appear to have widespread applicability.

B.

SCHIFFER

A PREVIEW

OF THE PAPERS

The papers that follow could have been orga-

nized in any number of ways. For example, I could have placed them in a random order.

That move is too postmodern for my tastes,

and it has already been done. Another possibility would have been to group the papers on the basis of theoretical program or level of societal complexity. Such organization, however, would contradict the theme of the FAI

roundtable, as well as do violence to the papers that have succeeded in building bridges

across these theoretical divides. As someone

whose surname is near the end of the alphabet, I was tempted to organize the papers in reverse alphabetical order (based on the senior author); this scheme at least invites read-

ers to peruse each contribution carefully before moving on. However, the book’s actual organization is based on inexpressible aesthetic criteria: I simply put the papers in an order that I found satisfying and hoped would also appeal to most readers. Brief introductions to the chapters now follow. Building on a commitment to the use of ethnographic/historic data and a concern with human agency, Jeanne Arnold examines changes in rights to labor in the evolution of social power. She defines power as someone’s capacity to call on—benignly or coercively— the labor of others on an immediate or delayed basis. The specific focus of her contribution is on whether societies at different evolutionary levels exhibit patterned differences in power and labor relations. Arnold’s working hypothesis is that, indeed, there is a rupture or great divide between “smallscale” societies—those led by elders, great men, and “big men”—and “intermediate hierarchical” societies, which are led by chiefs. In small-scale societies leaders have rights only to the labor of kin, whereas in intermediate hierarchical societies, chiefs have permanent rights to call on the labor of large numbers of nonkin. In furnishing support for this hypothesis she draws on ethnographic data from Melanesia and other South Pacific societies, building on her previous syntheses of California and Northwest Coast ethnographies. Arnold also suggests that the great di-

SOCIAL

THEORY

vide should be detectable archaeologically in the rapid transition of one societal type to another. Hoping to stimulate research along these lines, she supplies a number of archaeological implications for inferring changes in

leadership and labor organization. Not only

does Arnold provide formulations applicable

IN ARCHAEOLOGY

complexity. He provides a nuanced model of the political organization of late-prehistoric pueblo societies that explains why such societies have only some of the traditional trappings of complexity and also reinterprets the pithouse-to-pueblo transition in_ political

terms.

The

network/corporate

dimension

to all nonstate societies, but she also strives to bridge processual, postprocessual, and even selectionist programs. She is also concerned to demonstrate the complementary roles of archaeological and ethnographic data in the testing of evolutionary hypotheses. In his chapter Gary Feinman calls into question traditional formulations of political

clearly crosscuts levels of societal complexity, thereby serving as a bridge between societies whose similarities have been obscured by previous typologies. Moreover, although Feinman is working firmly within the processual program, his formulations may also resonate with postprocessualists.

cal complexity, wealth stratification, and centralization of power. He argues instead that societies can follow different paths to political complexity; thus, the political complexity of any society should be assessed in terms of several independent dimensions, of which hierarchical complexity is but one. A second important dimension is said to be a continuum of political organization marked at its extremes by the “network” and “corporate” modes, which pertain to strategies for acquiring and maintaining power. The network mode focuses power on individuals and their personal networks, such as exchange partners and kin groups. Archaeologically, societies dominated by network strategies are characterized by depictions of individual rulers, princely burials, palaces and other elite residences, and large wealth inequalities. At the other extreme is the corporate strategy, in which power is more apt to be shared, a consequence of group membership. Manifestations of the corporate strategy commonly include large-scale public works projects, “insignias of rank, monumental ritual spaces, central and uneven food storage, shared corporate codes.” Armed with the network/corporate dimension of political complexity, Feinman approaches an enduring controversy in the prehistory of the American Southwest, showing that the traditional dichotomy of egalitarian versus complex societies, in terms of which debates have raged, obscures important variation in political

behaviors, and contexts for understanding the social dynamics of middle-range societies. She addresses the variability of abandonment contexts and practices within nonforaging/ nonstate societies, examining how these create different migratory options for individuals and households. Abandonment is viewed as a transformation rather than a failure. Four aspects of abandonment are emphasized: process, scale, strategy, and place use. Nelson discusses how abandonment research has progressed and can progress further through the efforts of archaeologists approaching such research from varied perspectives. She discusses the contributions of agency-based approaches, behavioral approaches, and those that emphasize process. In demystifying abandonments Nelson addresses not only the value of abandonment research within the archaeological community but also the importance of speaking to other communities. Robert Kelly makes explicit his commitment to behavioral ecology as a “very good learning strategy” for understanding human behavioral variation. He demonstrates the effectiveness of this learning strategy by showing how applications of behavioral ecology in anthropology and archaeology during the past two decades have caused the ideas originally borrowed from ecologists to be modified in a more cultural-anthropological direction. For example, models now being built allow the analyst to represent the diverse

changes that assume covariation in hierarchi-

Margaret Nelson’s chapter explores the value of focusing on abandonment processes,

MICHAEL

B. SCHIFFER

goals of human agents, such as maximizing prestige or mating opportunities, that go well beyond optimizing the return rates of foraging activities. He also shows that behavioralecological models can be altered to take into account

variation

in

resource

scientists. Kuhn and Sarther illustrate the potential of conjoining formal models and social theory by presenting a scenario that focuses on food sharing as a stimulus for developing new technologies. In so doing they underscore the point that although specific social relations are unexcavatable, their effects on behavior and on the material record often are; thus, modeling social relationships as they evolved over long spans of time can be productive. Maria Nieves Zedeno’s contribution hinges on the relationship between the construction of social theory and the political arena in which modern archaeology has been practiced in North America. A survey of historical relationships between the social and the political in Americanist anthropology af-

processing,

division of labor, intrapopulation behavioral

variation, and cultural transmission of infor-

mation. Moreover, he points out that behavioral ecologists in archaeology must come to grips with stone tools, as well as the largescale, long-term “conglomerated” data that the archaeological record provides. Somehow, Kelly suggests, we must find ways to build models that are neither too sophisticated (because their data demands are too great) nor too simple (because they are unrealistic). Clearly, the directions in which Kelly is taking behavioral ecology establish connections to both behavioral and _ post-

fords a long-term perspective for Zedeno’s current research concern: reconstructing the

processual archaeologies. Focusing on early hominids, Steven L. Kuhn and Catherine Sarther examine the kinds of social theory that might be relevant to understanding the evolution of behaviors, such as food sharing and economic cooperation, that tend to be infrequent in other animals but highly developed among modern humans. The building of relevant theories, they argue, must rest on several desiderata:

social dimensions of landscapes among mobile groups. Her discussion of human/

land interactions draws strongly from prin-

ciples of behavioral archaeology but is also

grounded in contemporary American Indian

cultural-resource assessments. Zedeno integrates both behavioral and _ traditional knowledge on human/land interactions to develop a materialist framework for reconstructing landscapes, suggesting that archae-

(1) paleoanthropologists cannot simply project the cognitive capabilities of modern

ology,

social

theory,

and

politics

are

not

altogether incommensurable. Using the conjunctive strategy, Suzanne M. Spencer-Wood undertakes the ambitious project of building a bridge between an inclusive feminist theory and nonlinear systems theory (sometimes called “chaos” theory). Acknowledging that her project juxtaposes seemingly opposed and irreconcilable theoretical stances, Spencer-Wood nonetheless argues that this kind of bridge-building effort can, by focusing on similarities and differences between the two bodies of theory, identify their complementarities and usefulness for archaeological research. She develops in some detail the implications of conjoined feminist-nonlinear system theory for the major theoretical programs in archaeology. Michael J. O’Brien and R. Lee Lyman ar-

humans to Plio-Pleistocene hominids, and (2) the archaeological record does not provide clear “snapshots” of social phenomena such as pair bonding, kin bonding, provision-

ing, cooperation, and social roles. Thus, evo-

lutionary explanations must address the cognitive underpinnings of the behavioral trends that the archaeological record does record well, such as technological change. These requirements of theory building can best be met by conjoining aspects of evolutionaryecological theory and social science theory. That is, paleoanthropologists, while necessarily maintaining an evolutionary stance informed by formal behavioral-ecological theory and models, must also find ways to integrate understandings of the cognitive correlates of various behaviors developed by social IO

SOCIAL

THEORY

IN ARCHAEOLOGY

gue that evolution is not simply a different theory from others used to describe and ex-

reconstructed

of theory. What sets it apart from other ways of explaining nature is the focus on continu-

tive, Thomas argues that many takenfor-granted constructs in humanistic archae-

ous change rather than on transformation. Darwinian evolutionism, they point out, is a

ology are, in fact, of modern Western origin, including ambition, individuality, emotional states, even agency. He goes on to emphasize the relational basis of human identity and

materialist rather than an essentialist strategy, by which they mean that both the theory

and the methods that extend logically from it are geared to producing historical explana-

personhood and the embeddedness of power in material relations among people and

tions for things being the way they are. Given that evolutionism is a historical approach, its interests lie in configurational properties,

among people and artifacts. His approach to social theory requires that investigators situate people into networks of historically constituted relationships of many kinds. The worthy goals of a humanistic archaeology, particularly that of putting people into the past, are not abandoned by Thomas; rather, he argues that their realization will require amore ambitious research program based on a different—and less ethnocentric—metaphysics. This metaphysics, it should be noted, crosscuts levels of societal complexity and so can be viewed as an effort at bridge building. Almost as a counterpoint to Julian Thomas’s chapter, the essay by Susan Kus begins with a brief autobiographical journey in which she traces her successive intellectual transformations from Binfordian processualist (where ideas and meaning were epiphenomenal and unrecoverable), through Francophile Marxist and structuralist symbol seekers, to critical theory, the Annales School, and semiotics. Her present condition is that of a “feminist-postmodern-anthropologist” who is concerned with sensuous lived experience. Kus advocates that archaeologists pay attention to the ideal as well as the material and that in doing so we should grapple seriously with three themes/issues: agency, material culture, and reflexivity. In setting an agenda for an archaeology of the “mind, body, and soul,” Kus prescribes a dose of ethnography, focusing on the particulars of the lives of others. She also argues strongly

those properties existing in particular timespace positions. O’Brien and Lyman note

that within the last two decades increased attention has been given to extending Darwinlan evolutionism to the archaeological record. Acknowledging that the Darwinian umbrella can be extended over novel apthe

authors

caution

that

some

so-called evolutionary explanations are missing several key components, one of which is history. Because more such problematic explanations are likely to be encountered in the twenty-first century, O’Brien and Lyman underscore the need to lay out in precise terms why history is integral to any Darwinian evolutionary explanation. Although O’Brien and Lyman do not explicitly buy into bridge building, it is noteworthy that from time to time they continue dialogues with behavioral archaeologists and with their processualevolutionary detractors. In his chapter Julian Thomas points approvingly to a trend, discernible during the past 15 years, in which archaeologists grapple with new kinds of social theories and new formulations of archaeological problems.

However,

he also

In contrast,

Thomas advocates a perspective that is resolutely antihumanist. Outlining this perspec-

plain the natural world; it is a different kind

proaches,

empathetically.

sees in this trend,

especially in the writings of postprocessualism (broadly conceived), the emergence of a problematic humanistic archaeology. Not only does a humanist archaeology perpetuate Cartesian dualisms between mind and body, but it also privileges “the individual” over social phenomena, marginalizes the role of power, and avers that because human emotions are universal those of the past can be

that we should fill the theoretical void in the

area of “the person” (as gendered, as agent, as symbol user, as believer). Finally, Kus recommends that we develop/adopt language more appropriate for evoking “the incomIl

MICHAEL

B. SCHIFFER

answer our countless questions about variability and change in human societies, past and present. The essays themselves show that we have progressed far beyond the stage of finding in other disciplines theories of potential utility and applying them uncritically to archaeological phenomena. Indeed, a theme that runs throughout this volume is that social theories, regardless of source, need to be modified, tailored to the unique characteristics of the archaeological record. This means not just that the social theories of paleoanthropologists and hunter-gatherer specialists must have implications for understanding variability in chipped stone tools (see the pa-

mensurableness of culture and the inchoateness of lived experience.” It almost goes without saying that her ideas crosscut levels of societal complexity. SOME

AFTERTHOUGHTS

I have argued that archaeologists have adopted rational strategies for redlining large segments of the theoretical literature. This move allows an individual to ignore articles, books, and monographs aligned with other theoretical programs, as well as writings on societies at levels of complexity that differ from his or her own interests. However rational from the perspective of the individual, redlining creates and reinforces cleavages

pers by Kelly, and Kuhn and Sarther) but

that work to the detriment of theoretical inte-

also that our theories must be sensitively crafted to handle the wide range of societal

gration in the discipline as a whole. We cultivate our tiny plots but lose sight of the larger social-theoretic landscape. Thus, my simple

variability that the archaeological record dis-

closes. For example, the vast majority of hunter-gatherer societies, especially “com-

plea: let us invest at least some of our theoret-

ical energy in the risky enterprise of building bridges. I hope that the reader will agree that

plex” hunter-gatherers, are known only from

their archaeological records. Thus, it should

the participants in the FAI roundtable on so-

be no surprise that archaeologists such as Kelly, Arnold, and Kuhn and Sarther are ex-

cial theory, whose papers appear herein, responded creatively and constructively to this

ercising a leadership role in building appro-

plea.

priate social theories. Similarly, most of the

Virtually all the papers were revised significantly on the basis of discussions that

intermediate-level societies (the old “tribes”

took place at Snowbird, and each now reflects the author’s enhanced sensitivity to competing and complementary social theories found in archaeology, as represented by the other participants. I would be the first to admit, however, that the bridges that have been built are at best tiny and fragile structures. Even so, they represent the beginnings of a process that must move forward swiftly if the fragmentation of social theory in archaeology is to be checked before the discipline develops the seemingly permanent rifts that afflict other social and behavioral sciences. My sense is that everyone took from our brief conference a new respect for differing social theories, perhaps appreciating that in many cases the theories do not so much conflict as answer different questions or even different kinds of questions. This in turn fostered an appreciation for the magnitude of the task we face in forging the social theories needed to

and

“chiefdoms”)

that ever thrived on this

planet left only archaeological remains. That is why, beginning with Colin Renfrew and continuing through the efforts of Feinman, Nelson, and others, we are beginning to forge new theories that seem more attuned to the extraordinary variability in these societies. And as Spencer-Wood’s chapter demonstrates, even in the investigation of complex societies we ask a wider range of questions about how such societies come into being, are maintained, and collapse than do investigators outside archaeology. Clearly, access to the diachronic and synchronic variability that the archaeological record affords makes it possible to winnow, alter, and create social theories that have a robustness and fidelity difficult for practitioners of other disciplines, even cultural anthropologists, to achieve. Investigators

outside ar-

chaeology are hampered, I suggest, because I2

SOCIAL

THEORY

IN ARCHAEOLOGY

Acknowledgments. For helpful comments

they focus narrowly on an arbitrary timespace segment of the human experience and

on earlier drafts of this chapter, I thank R. Lee Lyman and Brian R. McKee.

because they do so, usually, on the basis of limited lines of evidence. The archaeological

record is far from

Notes Redlining is the practice by banks and other

a faithful reflection of

everything that went on in the past, but as

lending institutions of delineating areas—usu-

O’Brien and Lyman intimate, it is unquestionably the best record scientists have for

ally in minority neighborhoods of large cities—where they prefer not to make loans.

eties. The papers of Arnold, Feinman, Kelly,

that their theories are applicable at all societal levels. Be that as it may, I hypothesize that if

Both selectionists and postprocessualists might take exception to this claim, contending

studying the processes that create variability, throughout time and space, in human soci-

Zedefio,

Spencer-Wood,

and Kus

demon-

one examines citation patterns, he or she would find the same sort of complexity-based

strate unequivocally that when the archaeological record is coupled with other lines of evidence—e.g.,

from

history,

redlining in the social-theoretic literatures of these programs. I certainly do not claim that every archaeologist worked only in one area. A considerable number of culture historians plied their trade first in one area, then in another. I would be surprised, however, if many archaeologists carried out sustained cutting-edge work simuItaneously in more than one major area. In commenting on drafts of my paper “Some Relationships between Behavioral and Evolutionary Archaeologies,” several other behavioral theorists attempted to dissuade me from proceeding with the project. In view of how harshly selectionists had treated behavioral archaeology, in and out of the literature, fellow behavioralists argued that selectionism did not deserve my serious attention. One even suggested that I change the paper’s title to “Sleeping with the Enemy.” Clearly, there are strong social as well as intellectual barriers to bridging major theoretical programs.

ethnography,

and other sciences—archaeologists are in an advantageous position to build and evaluate

social theories.

I have long been an optimist that archaeol-

ogy has unique theoretical contributions to make to the social sciences. The chapters in

this volume indicate, to me at least, that this

optimism is not misplaced. What is more, the

contributions of Kelly, Kuhn and Sarther, and O’Brien and Lyman raise the possibility that archaeologists will help to fashion evolutionary and ecological theories that are more general than those that have been devised in the natural sciences to handle nonhuman species. Thus, as social theories in archaeology continue to mature and be refined, I expect that scientists in other disciplines across the academy will begin to find our formulations of interest.

3

2

Revisiting Power, Labor Rights, and Kinship Archaeology and Social Theory JEANNE E. ARNOLD

Societal changes of a rather profound nature,

drawing on case studies from Melanesia and

such as the development of ascribed hierarchies or the birth of political economies, are

neighboring areas. I will argue that centralized power crystallizes at the moment when

archaeologically detectable and are appropriate subjects for building theory. Theories encompassing concepts from multiple subdisciplines in anthropology have the potential to

rights to labor

become

firmly

dissociated

from close kinship relationships. I do not intend for the term /abor to generate in the reader’s mind

images

solely of large-scale

projects such as the construction of monuments (see also Bourdieu 1977; Roscoe

build powerful bridges in archaeology but may also collapse under a burden of excessive

1993). On the contrary, labor encompasses

eclecticism if taken to extremes (see Schiffer,

many kinds of activities, including a reckoning of people’s participation in ritual events,

this volume). Here I suggest two of the anthropological domains central to the future of archaeological theory regarding societal

(but judicious) use of

raids, warfare, feasts, harvests, hunts, ceremonies, and the like. In this sense the social

making processes (reliant in part on bioan-

cations are accorded equal standing with

change:

(1) extensive

properties of labor relations and their important ideological, kinship, and political ramif-

ethnographic and historical data and (2) attention to agency, motivations, and decisionthropological modeling). A critical question

their more obvious economic properties.

to explore using these approaches, and the issue on which I focus in this essay, is the role of

DEFINING

changing rights to labor in the evolution of centralized social power. I discuss several means by which changing power relations may be monitored archaeologically. The key question is how power, rooted in labor relations, emerges and evolves. Social historians and archaeologists have gathered substantial evidence over the past few decades regarding the close relationship between power and labor. This has been accomplished through the analysis of a considerable body of primary and comparative sociological, ethnographic, and historical material. Some of this evidence is discussed below,

POWER

At the most general level persons who exer-

cise power

are able to routinely call on

specific other people to do their bidding. Power refers to the capacity of persons to use methods ranging from simple persuasion or divine appeal to physical force in order to have other people carry out their directives. As we shall see, it is, more than any other factor, who these “other people” are that marks profound differences in the nature of leadership and societal complexity. Giddens (1979) has referred, in much the same sense, to the

“power over” and “power to” create action, reflecting the broad range of kinds of author-

14

POWER,

LABOR

RIGHTS,

AND

KINSHIP

ity and relationships with others exhibited by actors to achieve various results. Powerful agents may create starkly oppressive sets of relationships, or power may be exercised ona

tent with the way I define chiefly roles and labor organization. Chiefs’ control over the labor of commoners, transients, and slaves in their numaym—and their “ownership” of

relations are present in all societies, but the scope or scale of power exercised by leaders in any given society arises in large measure, I argue, from the degree of articulation between kinship and labor relations. Persons may of course embody many different kinds of power simultaneously and may interact in multiple ways with people of different social positions, including having power over some and being subordinate to others. These complex individual contexts of power are likely to be beyond the view of most archaeological investigations. Nonetheless, the aggregate of such relations should be evident in the ground. That is, past societies across a range of levels of integration can be characterized by at least some types of material evidence that reflect forms of power relationships. In this analysis I explore such variations and their consequences for a diachronic and evolutionary understanding of power and labor in small-scale (tribal to big-

Neither Wolf nor I suggest that all social acts

cieties (cf. Yoffee’s [1993] analysis of power at a much higher plane of complexity).

tion, harvests, production of goods for trade,

shamans to mobilize sorcery against enemies (1999:91—-92)—constituted their power base.

grand scale through gentle secular or religious speech, to note a few extremes. Power

between leaders with power and their follow-

ers are overtly economic acts with immediate physical labor consequences (again, the image of the labor manager overseeing the masses toiling to build a mound may come to mind, but this is only one kind of labor). On the contrary, in the labor relationships I refer to here leaders regularly involve followers in ritual events, alliance formation negotiations, political rallies for the purpose of spreading rhetoric, and similar activities that may appear completely benign with respect to direct physical labor. Yet these relationships are rarely benign. The followers accumulate obligations and debts daily that must be repaid to leaders through the provision of fealty, work, or goods. Even generalized contacts between leaders and constituents are interwoven with requirements for followers to contribute, typically on a delayed basis, to leader-sponsored projects, including organized raids, construc-

man groups) and intermediate (chiefdom) so-

RIGHTS

and the like. In these multiple ways followers express their allegiance to leaders—or express their resistance through noncooperation. UItimately, even the most symbolically and ritually charged relationships between leaders and others involve the execution of leaders’ ideas through the actions of their constituents; power, then, is the ability to engage others in those actions. Ina similar vein Julian Thomas (this volume) suggests that human agency is inextricably linked to labor, power, and technology. Agency is relational—reflecting connections among people and between people and technologies—and it is, most fundamentally, the capacity to bring about

TO LABOR

The capacity to have others do one’s bidding fundamentally means that one is routinely able to call on their labor, immediately or ona

delayed basis. That is, power is manifest in ongoing and renewable control over people’s productive capacities. Some readers may find this assertion controversial, narrow, or (bet-

ter yet) self-evident, but ethnographic support for this point is strong. Eric Wolf’s (1999) analysis of power strongly affirms in many ways this contention, although we disagree on when to apply the term kin-ordered mode, a point that I do not have the space to explore here. Power clearly centers on deploying and allocating social labor (and associated ideology, according to Wolf). His analysis of the Kwakiutl is quite consis-

action.

The definition of a successful leader is one who succeeds in persuading followers, first, to believe in an agenda and, second, to make themselves available to execute it. Similarly, a TS

JEANNE

E.

loss of leadership is defined by the defection of followers (and, alas, their labor) to competitors. Such an approach to the nature of power and social change focuses on both leaders and followers, what Giddens (1979: 145-150) labels the dialectic of labor control. This refers to the capacity of all humans to contribute

power/labor organization differ both qualitatively and quantitatively in small-scale and markedly hierarchical societies. I do place emphasis, however, on the suggestion that, in both a classificatory sense and an historical sense, there is a profound disjuncture rather than a continuum in the nature of power relations and who holds rights to others’ labor in these kinds of societies. I also highlight the nature of the changing kinship ties between leaders and followers as institutionalized power emerges. For this analysis I have reviewed ethnographic accounts of about 30 societies in the South Pacific and, earlier, a comparable number in Pacific North Amer-

(or not) to the requests and de-

mands presented to them. This approach also encourages us to think of the ways these agendas might be carried out or thwarted

within societies at many different levels of complexity. These concepts of power building overarch a number of more narrowly drawn models emphasizing specific economic phenomena, such as how leaders manipulate prestige goods or long-distance exchange, how they control local production or build monuments and mounds, or how leadership

ica (Arnold 1996b), extracting details about

their labor and kin organization. Although I drew on good-sized samples for these two regions, I recognize that such surveys are inevitably imperfect and that there are much larger literatures than can be cited here.

of war or ritual behavior may result in politi-

cal centralization.

KINSHIP AND

ARNOLD

My goal is to produce what Foucault and others might call a genealogy (a historical reconstruction) of the relationship between power and labor. My working hypothesis is that a fundamental divide separates the mode of leadership practiced by elders, great men, and big men, who have renewable rights only to the labor of close kin and some distant and fictive kin, and the form of leadership practiced by chiefs, who have permanent rights to the labor of significant proportions of nonkin. My purpose is to determine with whom rights to labor reside within a sample of such societies and to assess whether there is a clear separation that cleaves the many cases into two or more groups. If there is such a distinction, then I suggest that, historically, we could expect that the transition between small-scale and hierarchical societies may be characterized by punctuated change (Arnold 1993).

LABOR

Analysis of the dialectic of labor control permits exploration of power and labor relations through time and across landscapes and cultures. Archaeology is uniquely suited to address the problem of how power is built and lost, and ethnography is necessary to provide insights into the process in the short term. I hope to tease out linkages in power, rights to labor, and kin relations that vary across sets of societies at different levels of complexity/ integration. In a significant segment of the ethnographic literature, systematic data are unavailable regarding the form of power held by leaders, that is, whom they could induce or compel to carry out their wishes. Yet this kind of information is essential to characterize the range of variation in power and labor crossculturally: we need to establish who had renewable, enforceable control over the labor of whom, whereas we are much less concerned with what particular tasks were involved. Although our ethnographic and historical database is thus incomplete in this respect, sufficient data can be distilled from certain regions to highlight important relationships among these variables. I am hardly the first to assert that forms of

This

ARCHAEOLOGICAL

essay

travels

two

GOALS

related

exploratory

paths: (1) the changing rationale for social typologies used by archaeologists and (2) the central role of the relationships among leadership, labor, and kinship in reconstructing past trajectories of sociopolitical evolution.

16

POWER,

LABOR

RIGHTS,

of the broader objectives of archaeology. Ar-

sia and elsewhere in the South Pacific, building on earlier work with archaeological and ethnographic data from nonagricultural areas such as the American Northwest Coast and California (Arnold 1996a, 1996b).

chaeological data are ultimately used to explain how people at specific times and places

were organized and how society evolved. We seek to comprehend the ways that people of

Power and labor relations and kinship reck-

ancient societies behaved: how they carried

oning are weighed in light of observations from these regions. The possible ways that power, labor, and kinship organization may intersect and vary among societies can only be understood through analysis of contemporary and historically observed peoples across a range of levels of complexity. If we can identify meaningful variability and patterning in these relations from multiple areas of the world, and assess their material correlates, then we will have a tangible basis for identifying changing relations of labor and power in the past. Ethnographies (and historical data) are the only sources that can provide the richly textured details of these social phenomena, and archaeological data can provide a complementary picture of longterm dynamic processes. The future of social theory in archaeology, particularly for this type of research, may well rest on this storied interdependency.

out political successions, conducted ceremonies, and organized work projects and burial rites; the ways they used kin networks and other relations to structure daily interactions; the ways they articulated with their natural environments; and so on. In short, we want to know about variable organization and its changing forms. To link the past and present we must often rely on experimental data or ethnographic data and models, as I do here. Current moderately “eclectic” forms of processualism (Arnold 1999) encourage an interlayering of agency and action theory with processual analyses, especially as these topics pertain to power and _ hierarchical emergence. The relationship between agency and power has attracted considerable attenin

recent

years

(e.g.,

Roscoe

KINSHIP

Here I use ethnographic cases from Melane-

First, let us briefly examine these goals in light

tion

AND

1999;

Thomas, this volume). My own approach is to draw from Marxist/materialist perspectives and to refer to agents and their actions, while remaining partial to the overall “science” framework provided by processual and evolutionary explanations. I use agency here in the sense of its focus on actors and their motivations and decisions rather than professing to use an unmitigated agency approach. Ultimately, my objective is to reach a greater understanding of power and the organization of labor, however that can be done most effectively. The changing role of kinship and labor organization in the evolution of powerful leaders is a compelling problem. A web of related questions has long challenged social scientists, and Marxist approaches have of course encouraged exploration of this theme. Rarely, however, have labor rights and leadership been examined by archaeologists (from any perspective) in prestate and/or nonagricultural societies (see Saitta [1997] for a Marxist analysis of a prestate case).

CLASSIFYING SOCIETIES: A FOCUS ON LABOR RIGHTS AND LEADERSHIP

The emergence of centralized leadership and ascribed inequality are not simple consequences of particular subsistence regimes. That is, hereditary power and cultural complexity are neither universally nor necessarily associated with farming economies, and complex hunter-gatherers should not be excluded from discussions of sociopolitical complexity (Arnold 1996a, 1996b; Hayden 19954), as they so often have been in the past. Once we recognize that agriculture neither corresponds to nor explains formalized hier-

archical development, we can turn more pro-

ductively to the study of the key sources of so-

cial differentiation in societies, such as their

extraction, exchange, production, and labor organization systems. In this essay my goal is to expand on a topic largely unexplored in my earlier work

17

JEANNE

E. ARNOLD

Societal “type”

Leadership structure: Where rights to labor reside

Small-scale

Kin elders, great men, or big men: hold rights to labor or household, close kin, fictive kin

Intermediate (hierarchical)

Kin elders and chief(s): both hold rights to domestic labor (for chiefs, a large proportion of labor is nonkin)

Complex (hierarchical)

Kin elders, chiefs or kings, and other bureaucratic layers: all hold rights to domestic labor

Figure 2.1. Labor Rights and Leadership

on

complexity

and

organization

(Arnold

of leadership and labor control. So, on the basis of my working hypothesis I suggest that

and descent reckoning that provides structure to the process of claiming rights to vari-

guishable from hierarchical societies by examining where renewable rights to the labor of commoners/household members reside (see Figure 2.1). Understanding how the

1993, 1996b): in each society there is a changeable social logic of kinship relations

small-scale societies may prove to be distin-

ous people’s labor. The particulars of this logic may determine a great deal about the

structure of the society. It is thus appropriate to examine such relationships in our efforts

rights to draw on domestic labor are structured may provide us insight into how many

to understand sociopolitical variation.

leadership, or administrative, layers there are in a society. I propose that labor rights can be used as a shorthand for overall sociopolitical organization, eliminating the need for

Recognizing that cultural classification is,

in and of itself, an empty exercise, I nonetheless advocate a fundamental recasting of our typologies—rather than outright rejection of all of them—because every archaeologist continues to classify groups. I use hierarchy, heredity, and labor organization to distinguish societies, eliminating subsistence as a primary

criterion.

original

organizing

Given

lengthy trait lists enumerating diet, population size, community size, roles of leaders, in-

tensity of exchange, and the like. (Attempts to catalog such traits and assign ranges of value to them have confounded many efforts to classify societies in the past.) Where rights to draw on the labor of households are singly held by family elders, or male agnatic kin work under the direction of a great man or big man, and no other more highly placed entity calls on that labor, I would classify the society as small-scale (as shown in Figure 2.1). (There are of course gender- and age-based inequalities in all societies, including small-scale groups [Flanagan 1989]). Elders/leaders can only tap the labor of people closely related to them. This I contrast to labor relationships in formal hierarchical societies. In these cases there is domestic labor managed by household or descent group leaders, but there is also a second level in which leaders such as chiefs or bureaucrats positioned above elders consistently hold renewable rights to call on the labor of persons (across many households) who are unrelated

the pervasiveness

and influence of typologies, those that we select should not convey (to students, colleagues, or the public) concepts that are not empirically supported. A case in point: the structure

for the

1997

Snowbird Conference papers, the source of this volume, was a familiar sequence of three implicitly mutually exclusive categories, “hunter-gatherers,

intermediate

societies,

and complex societies.” This structure implied that hunter-gatherer societies could not have been intermediate or complex. That is, agricultural “complexity” was counterposed to “hunter-gatherer simplicity.” This same dichotomy is found in most contemporary articles and monographs citing a typological scheme, but such classifications are factually incorrect and counterproductive. I find it more illuminating to distinguish societies on the basis of broad characteristics

to 18

them,

and

there

are

costs

associated

POWER,

LABOR

RIGHTS, AND

with noncompliance (see below and Arnold

KINSHIP

nected with kinship, and consequently the kin-labor-power relationship is a crucial one

1996b). These distinctions should also correspond in most cases to achieved versus as-

to explore as we examine basic social evolu-

cribed leadership. Without belaboring the obvious, intermediate and complex hierarchical societies where central leaders regularly call on household labor (below the dashed line in Figure

tion. If changing rights to labor did indeed contribute fundamentally to social change, then archaeologists can focus productively on diachronic analyses of group size (population buildups and siphoning), specializations, feasting activities, subsistence intensification, and related processes in an effort to examine

2.1) are different kinds of societies from those

that organize labor solely through kin ties. I suggest that the labor rights/power dialectic

the relative size and composition of labor forces potentially available to leaders.

is the factor that drives and creates an array

of other social and political distinctions be-

tween small-scale and hierarchical societies,

POWER,

but this question requires extended investigation, beyond what can be fully resolved in the present essay. Later in this discussion I touch on some possible explanations for the transi-

Many social ers come to commoners, the power to

tion from one level to the other, which must involve manipulations of labor rights. Some

LABOR RIGHTS, AND LEADERSHIP

scientists agree that when leaddominate lower-level leaders or it means that they have gained secure the labor of these people

at will (Bourdieu 1977; Collier 1988; Roscoe 1993; Wolf 1999). I now consider in greater

support for these ideas may be found if broad divisions in labor rights consistently correlate with single versus multiple layers of leader-

depth whose labor leaders can marshall and what the relationships are among the various

ship (Johnson 1982) and if the historical tran-

participants.

sitions from small-scale to hierarchical societies are punctuated. Punctuated change in power and leadership further implies that we should find few examples of “transitional” types (between small-scale and hierarchical) in the ethnographic record. Changes of great magnitude in power and labor organization—no matter how rapid— ought to be quite visible over time in the archaeological record. These transitions have many potential material consequences, as discussed later in this chapter. Rapid political and economic evolution in the early second millennium A.D. in the Channel Islands region of California appears to be an example of such a transition in the archaeological record (see Arnold 1992, 1993, 1996b), evident primarily through the appearance of intensive, specialized craft production, control over key resources, accelerated and intensified exchange, and major technological innovations. Labor rights touch most dimensions of social life, including daily work choices, interaction, time allocation deci-

This

discussion

sets

up

pro-

posed relationships among variables that will then be evaluated against ethnographic details from Melanesia in the following section. Let us begin with several observations about kin terminology in big-man and chiefdom societies. As is widely acknowledged in the ethnographic literature, people create or redefine kin and fictive kin terms to suit particular purposes (Bourdieu 1990:166-167; Feil 1984:27, 33; Lemmonier 1991:27; Swartz 1960), including occasionally for the purpose of claiming labor. Relatively flexible or expansive definitions of kin are recorded for quite a few great-man and big-man societies in Melanesia, although there is tremendous local variability and relatively low numbers of people are involved (Feil 1987: 130-131).

In these groups

competition

for

kin labor may be exceptionally fierce, and

this competition may encourage creative ap-

plications of kin terms. Evidence compiled by Douglas (1979) and others in the broader South Pacific region further verifies this expectation. As people are needed by big men for labor support, more inclusive terms are often employed, or more outside people are

sions, exchange, and access to resources, to

note a few. Each of these is intimately con-

19

JEANNE

E. ARNOLD

assigned close kin terms (see also Collier 1988; Kelly 1993). Fairly inclusive and flex-

fresh blood into small labor pools. It thus appears that a flexible mechanism is in place in

kin, affinal kin, or nonkin to be assigned close agnatic kin terms in parts of Highland New

terns at some length, emphasizing that extremely flexible descent reckoning in the Eastern and Western Highlands sometimes results in the rewriting of genealogies in just two generations as big men and despots re-

would seem to propel them toward chieftainship—but it seemingly rarely or never does, and this tool turns out to be of little importance in chiefdoms (it is likely too contradictory to the rhetoric of conical descent reckoning). How can we best sort out the roles of kinship and labor, then, in the political transition to chieftainship? Roscoe (1993:114) indicates that formal political hierarchies emerge through the capacity of leaders to accomplish a range of

the data from the range of societies reported by Feil and others indicates that the more dynamic and frequently negotiated the “bigmanship,” the more flexible was the recruitment and kin-term conversion system, as they sought to bring in other people and make them part of the group. As Feil (1987:148) suggests for the Eastern Highlands, “To get people to behave like kinsmen should, one had to give them kinship status,” which implicates labor, rights, and obligations of all

trated in the hands of a few. I argue more specifically that the appearance of political hierarchy should prove to be a function of the ability of increasingly powerful leaders to call on a broadening sweep of the labor force— the vast majority of whom are not kin or even fictive kin. As more powerful leaders appear, the proportion of the total population whose labor is available to them increases markedly, and the nonkin labor force comes to dramatically outnumber the close kin involved. In

In chiefdoms, on the other hand, I believe that there is no need for this particular form of flexibility in recruiting and relabeling nonkin as close kin. Chiefs can and often do drop pretenses of close relationships with

Percentages on the vertical axis provide a sense of relative rather than absolute proportions of nonkin labor available to leaders of different types. Giddens (1979) has called the power to draw on people “relational” power, which is the capacity to secure outcomes that depend

ible kin-labeling systems allow some distant

many

Guinea, for instance, putting them in the cat-

egory of people who can then be called on for

labor and assistance. Feil (1987:130-136) discusses these pat-

cruit labor into their groups. My analysis of

means

to gain

economic

or

the labor ritual

societies, which

Figure 2.2 I present this model schematically.

all group members because they have more authoritative

big-man

tasks as power over others becomes concen-

kinds.

nonkin—through

Highland

of

con-

on actions of others. Giddens, Bourdieu, and

trols, for example. Of course, a general kin term may be widely applied in order to link

others have primarily sought to link “action” and “agency” with the development of institutionalized hierarchy, but I wish to recast this by drawing attention to the fact that what leaders of hierarchies have in common is their renewable right to the labor of an abruptly larger segment of the population— not only their kin but large proportions of nonkin. This transition is signified by the major inflection point in Figure 2.2. Power-

many people, putatively, to the chief. Typi-

cally, in such cases the people are labeled his “children,” for example. Clearly, this is quite different from the phenomenon of selective kin-term expansion by big men. In any case actual descent reckoning in chiefdoms may be aimed explicitly at setting apical lineage(s)—with their repeatedly validated ties to the original chiefly ancestors—sharply apart from the lesser lineages and common people. In short, central descent-reckoning practices in chiefdoms are more typically exclusionary rather than designed to recruit

ful leaders of course accumulate

(and often

redistribute) specific types of valuables add to their prestige in the course of process (e.g., exotics, prestige foods, or mented beverages; see Hayden 1998), 20

that this ferbut

PowER,

LABOR

RIGHTS,

AND

KINSHIP

Figure 2. Labor Rights Held by Leaders 100

100

90 |

/

80 +

aw

S 3

=z

60

PARAMOUNT CHIEF

4 80

70 +

g$

|

ao

/

+

504

/

+ 70 +

60

50.

40 +

+

30 +

+ 30

20 +

+ 20

10 +

+ 10

0 ELDERS

eee

4 BIG MEN

GREAT MEN

CHIEF

40

—+0 CHIEFS

Figure 2.2. Labor rights held by leaders.

the truly prized commodity that allows them to retain and extend power beyond specific events is permanent control over the labor of nonkin. Labor is the great generalizer—the manipulator of labor can accomplish many different things. In short, elders, great men, and big men have little or no access to the labor of nonkin,

whereas

(Wolf 1981:45), so what is worth examining

in order to understand major upheavals in past sociopolitical and economic life is change in these basic labor relations. Moreover, because the labor process requires extensive coordination and planning, the execution of projects is and always has been laden with social information and meaning, and the study of labor organization thus allows us to search for intentions, ideation, and webs of social relations in the past. Actors play a significant role in creating the processes by which power builds; it is not a

chiefs of various

status have significant access to—or virtual control over—such labor. An agency approach suggests that the potential for centralized power to develop exists wherever the agency of others is needed for aspiring leaders to satisfy

certain

goals

(Roscoe

passive process. That is, actors create politi-

1993:

cal systems and labor relations (Roscoe 1993), although the final form such relations take may be unforeseen because of historical contingencies and selective forces.

114). In the archaeological record, then, we can search for material outcomes that took the form of construction projects, contribu-

tions of food, joining in battle, and other visible expressions of loyalty or tribute to

The modes of production that characterize human economies (communistic, tributary,

leaders by abruptly increasing proportions of the society. Wolf (1981) analyzed the labor process as a fundamentally social phenomenon. His analysis underlines several of my main points. Wolf was interested in labor and society in the broadest senses of these terms. Labor is of course carried out by people “linked to one another through social relationships”

feudal, capitalist, kin-ordered, etc.) are, of course, sets of social relations, each deploying social labor in a unique way to transform nature by means of tools, skills, organization, and knowledge (Wolf 1981:46). Labor orga-

nization determines a great deal about the overall organization of society, structuring not only how products are made but also 21

JEANNE

E.

how resources are used and goods and services are distributed. In Wolf’s (1981:48) words, “the way in which the mode commits social labor...governs the way...resources used and obtained are distributed among producers and non-producers.” The kin-ordered mode of production serves to illustrate what kinship systems do in terms of structuring labor relations in smallscale societies (cf. Bourdieu 1990). Kinship is, among other things, a way of committing social labor to various tasks through appeal to

ARNOLD

chiefdoms. As Thomas (1989) points out, there are problems with aspects of historically unsituated Pacific ethnographies, and I depended on Feil’s (1987) archaeological summaries and other scholars’ use of misslonary accounts, as well as ethnographic records, to trace historical depth in some of the patterns discussed below. I am naturally sympathetic

to Thomas’s

(1989:121)

point

that history and archaeology are essential to understand snapshots of social structure, although not every source that I cite here would likely meet with Thomas’s approval. Maori leaders of New Zealand had sub-

marriage and blood ties. The social con-

structs created by kinship systems situate human actors in particular relations with one another. These relationships, and these kin terms, permit persons to call on “shares” of

stantial duties and responsibilities on behalf

1981:53). The concept of available shares of

of their followers but had “very little social power” (Bowden 1979:56). They could neither call on common people to contribute labor for projects nor compel obedience. In war

ganization and is explored further below.

gained leadership through achievement and

social

labor

for

important

tasks

(Wolf

labor and who has the right to “cash them in” is heuristically useful for tracing changing or-

SUPPORTING

DATA FROM

the people were led but were not ordered to participate. Political leaders had mana and

personal skills; ritual leaders had authority in

ritual matters only, although their power was ascribed and inalienable. Maori leaders fit a classic pattern for big-man-level societies in which, I suggest, it could as easily be said that the people had rights to the hard work of the big man as that the big man could expect various kin to do his bidding. These leaders often toiled tirelessly for their constituents—for instance giving gifts, making speeches, and bro-

MELANESIA

The degree to which tribal leaders, great men, and big men everywhere (see Godelier and Strathern

1991)

can

call

on

labor

seems

firmly limited by the scope of their kin ties. That is, these kinds of leaders rarely or never have the capacity to truly transcend kinbased restrictions (even when

descent reck-

oning is quite flexible and in-group recruit-

ment is common;

Feil 1987) and to ask for

kering peace (see also Clastres 1988:38, re-

labor or to reprimand uncooperative people beyond their kin group. Ethnographic and archaeological data from many places around the world appear to verify the general rule that chiefs, on the other hand, can routinely call on the labor of individuals who are distant kin and nonkin and can make noncompliance costly or even lethal (Arnold 1996b). This major distinction between smallscale (great-man and big-man) and formal

garding South American tribal leaders). Such behavior is characteristic of many tribal, great-man, and big-man societies and differs qualitatively from the expectations and relations between chiefs and their followers. In formal hierarchical societies chiefs (and assis-

tants/henchmen) may work hard to maintain their authority, but followers are in no position to presume that the chief will deliver something to them or will toil on their behalf. On the contrary, chiefs expect to receive regular prestations of labor and tribute from their

hierarchical (chiefdom) societies (Figures 2.1, 2.2) can be demonstrated by more ex-

tended reference to cases in Melanesia and elsewhere in the South Pacific, where the bigman (and great-man) concepts were developed and where there were also traditional

supporters (Liep 1991:34).

In the Sepik region of the northern New Guinea coast, there were a number of greatman and big-man groups and other societies 22

POWER,

LABOR

RIGHTS, AND

representing varying degrees of centralized

characterize as “big-man” leadership and work in the Tierras Largas phase at San Jose Mogote, Oaxaca.)

leadership, four of which were compared by

Lutkehaus (1990). In the Boroi region, on the mainland, ritually charged boar’s tusk valuables give clan leaders the right to “tell people when to fight... [and] demand labor of fellow clansmen” (Lutkehaus 1990:188), but we see that it is only the labor of kin (the leader’s

Manam political authority is linked overtly to economic relations, and the leader can engage villagers in production and exchange on his behalf. This simple chiefdom exhibits a qualitatively different kind of leadership from that exhibited by the Wogeo,

own subclan or clan) that the leader can call

on. Even when several clan leaders bestow on one of their rank the right occasionally to lead a village, he is seen as merely transmitting the power imbued in the boar’s tusks and, importantly, only in the short term. That is, the rights to nonkin labor are temporary rather than renewable. Among one of the New Guinean island groups, the Wogeo, the village headmen (kokwol) are moiety leaders with limited authority. They only can persuade kin to do their bidding (Allen 1984; Lutkehaus 1990:190), although Allen also mentions that some may briefly extend authority across two or more villages through great successes in feasts or sorceries. However, among the Manam of this region, traditionally classified as a chiefdom, there are two hereditary classes of people, with the top-ranked clan in power. The leader of this clan can “oversee and coordinate” the magical resources of other clans (Lutkehaus 1990), meaning that they can draw on their powers and their labor and can compel these other clans to provide services. Lutkehaus (1990:193) is quite clear on this point: “Unlike leaders in most other Sepik societies, whose authority does not usually extend beyond their own kin group, a Manam tanepoa has the right to command labor and resources from members of any clan within his village.” In short, big men do not have authority beyond their immediate kin group; leaders of hierarchical “chiefdoms,” such as the Manam and Trobrianders, do. Chowning (1979:81)

agrees,

stating

that

only

KINSHIP

Boroi, and Maori, even though, as Chowning

(1979), Douglas (1979), Rubel and Rosman (1978), and others point out (contra Sahlins 1972), succession to leadership throughout Melanesia is characterized by every possible combination of principles of achievement and ascription. Big men always work very hard and require considerable cooperation to succeed, but often they inherit the right to succeed predecessors (fathers, uncles) if they are reasonably fit for the job (Gustafsson 1992). Their hard work and their battles for

leadership are often directed toward tapping the labor of younger kinsmen who are used to support ritual and economic activities (Gustafsson 1992:87; Rubel and Rosman 1978). The classic and more powerful Poly-

nesian chiefdoms in Hawaii, Tonga, the Soci-

ety Islands, and elsewhere had chiefs who monopolized resources, labor, and sacred knowledge and clearly had the power to im-

pose

onerous

sanctions

for

uncooperative

behaviors. Specific paths to power in Melanesia varied considerably and may have changed significantly as colonial rule attempted to force the cessation of warfare (Chowning 1979; Feil 1987; Roscoe, personal communication 1997). Some hereditary leaders were war leaders before contact; hence, more recent in-

terpretations of processes of succession may reflect only postcontact conditions (when hereditary succession likely weakened). Support for this argument may be found in data

these

that suggest that warfare and head-hunting

chiefs could truly “command obedience” and had the power to enforce their will. (For a useful comparative archaeological case, see Marcus and Flannery [1996:88—89], who describe the likely organization of what they

appeared important to the power of big men in several very recently contacted Highland New Guinea societies, although in other societies war leaders and civil or ritual leaders shared power. (Supernatural knowledge 23

JEANNE

E. ARNOLD

applied to gardening, fertility, weather, and health was important for big-man ritual power [Harrison 1987a].) Both Feil (1984) and Liep (1991:34) discuss other ways that European contacts pro-

claim community bragging rights, but they did not indebt or obligate nonkin labor by dispensing surpluses within their communities (except sometimes within their narrow

chiefdoms in the Massim were at one time

man and great-man systems as foreign goods disrupted economies. Feil traces the rapid

Tombema Enga; Lederman 1991; Strathern 1971). Customarily, meat or live pigs were exchanged. Chiefs involved in similar systems, in con-

gen (Melpa) as missionaries moved into the Highlands and brought massive quantities of

many kinds of goods, including rare valuables, and parceled some of them out to fol-

foundly

affected

leadership.

Liep

kin groups), nor did they typically durable valuables (Feil 1984 on _

believes

more widespread but “decomposed” to bigtransformation of the later postcontact Ha-

trast,

shells to buy food and labor from the people.

lowers in exchange for local labor and tribute

Although

Strathern

(1971)

disagrees,

or wealth

1991; Liep 1991). Big men

(but rarely

great men, who were more focused on warfare; see Feil 1987) underwrote their positions through exchanges of pigs and shells. The evidence for these exchanges is unambiguous and widespread, but the pattern seems difficult to reconcile with the general rule

that

power

resides

fundamentally

to

used the prestige goods to get renewable access to the labor (or tribute) of many people in their spheres. In the end it was the strategy of converting valuables to control over local nonkin labor that made a chief who rose to power through external finance a successful one. It appears that the political transition to chieftainship may be a process by which a broad spectrum (and high proportion) of nonkin labor comes to be controlled by chiefs who have seized specific opportunities to solidify their holds on that labor. Gathering and dispensing wealth is described by many Melanesian scholars as the essential basis for big-man prestige, influence, and power. On the other hand, specialists studying leadership in this region recognize the several very real limits to the power of big men: they could manipulate debt and credit, seek cooperation if they distributed products in return, threaten to withhold economic support, and attempt to

finance, is cited as a primary means of achieving centralized power in many parts of the world. Prestige-goods models of elite emergence focus on how leaders finance their positions (Clarke 1986; Hayden 1998). In Melanesia extensive external exchange activities sometimes appeared in the virtual absence of control over nonkin labor (Lederman

access

Liep 1991:34). Power in the latter cases was clearly linked with labor: hierarchical leaders

gen may have been an unstable chiefdom with “serfs” and clear labor control wrought by leaders’ ability to control limited supplies of prized shells. Subsequently, this system collapsed and democratized rapidly as inflationary pressures and a much wider availability of shell valuables—resulting from largerscale European intrusions—undermined local leaders’ power (Feil 1987). over external wealth,

commanded

(Feil 1984:83-97 on the postcontact Hagen;

Feil

(1984:91) believes the initial postcontact Ha-

Control

increasingly

use the

thwart favorable marriages, for example, but

they could not effectively force their constituents to do work for them. Often there were other kinds of leaders with whom they had to share power, as in New Caledonia, where “masters of the soil” controlled the land through their ties to ancestral spirits and kept political leaders from holding unitary power (Douglas 1979). Typically, great men and minor big men used their wife’s and children’s labor to raise pigs and produce other

in

leaders’ rights to the labor of local nonkin. However, when we consider the uses to which external goods were put in those Melanesian societies that relied on heavy circulations of prestige goods, we see a clear dichotomy between charismatic leaders without institutionalized power (big men) and chiefs. The former used such goods to build prestige and

24

POWER,

LABOR

RIGHTS,

foods for feasts or exchange partners. Polyg-

AND

KINSHIP

alized as members of a patrilineage) was un-

yny was a frequently used tool to maximize the output of this family labor pool and pave

scathed and resolved within a generation by

labeling all of the new additions as kin (Rubel

the way for greater big-man prestige. Where

and Rosman 1978:294). Feil (1987:136) similarly acknowledges that kin terms were ex-

big men were able to also pull in the labor of

their younger brothers—junior agnatic kin— their labor pool expanded slightly (e.g.,

tended and manipulated frequently in the Western Highlands of New Guinea for pur-

group [Rubel and Rosman 1978:56]). Among the Chimbu the more powerful big

and production support. In parts of New Guinea, then, an impor-

among the Northern Abelam, a Sepik River

poses of widening exchange relationships

men tried to persuade sisters’ husbands and

tant rule of social logic appears to have been

wives’ brothers (people who were not agnatic kin) to support them, although such attempts could be costly in terms of payments of land or other goods (Rubel and Rosman 1978: 165-166). Those societies in which big men appeared to hold greater degrees of prestige correspond well with their ability to tap an expanded labor force, but big men typically could not cross over to cash in on the labor of other clans or even subclans, according to

that kin ties governing labor rights followed agnatic-patrilineal lines. Closely tied to this

behavior was of course the set of rules gov-

erning marriage. The more complex Highland New Guinea societies, such as the Enga and Melpa, tended to have had the most local prohibitions regarding marriage partners, the broadest marriage webs, and therefore the most outward-looking, inclusive kin systems.

However, as marriage partners came from in-

creasingly distant sources, either spatially or in terms of relatedness, there was clearly less potential for drawing on the labor of the matri-kin (e.g., brother-in-law or his sons). This may be one of several structural problems that (patrilineal/patrilocal) big men must transcend to develop chiefdom-level power and access to the labor of large numbers of nonkin.

Lemmonier’s (1991) review of some 20 to 25

Melanesian groups. Moreover, clans rarely acted as corporate groups with pooled labor and resources (Feil 1984). What is equally striking is that in the larger and more complex New Guinean societies, people appear to have had increasing flexibility in shifting residence, choosing with whom they were affiliated and with whom they exchanged. The evidence suggests (with a few exceptions such as the Arapesh) that there may have been, in the more powerful bigman groups, especially in the Western Highlands of New Guinea, progressively shallow genealogical recollections, one-to-two generation kin reckoning when it was convenient, and occasional recruitments of nonagnates who were quickly incorporated into the agnatic descent system (Feil 1987:130—131; Rubel and Rosman 1978:245-246). The

DISCUSSION

I next place these diverse observations in perspective. The degree of flexibility in the application of kinship and descent criteria to recruitment of outsiders seems to have played a major role in the relative power of leaders in all Melanesian groups. Where descent was more strictly defined, “the ability of an aspiring leader to recruit outsiders in his support” was very limited, according to Douglas (1979:6). I take this to mean that where descent reckoning was especially strict, important, and narrowly applied, aspiring big men did not have the tools to gain supporters outside their descent groups or subclans, and it is implicit in this statement that these big men could not call on the services of those defined as nonkin or even wife’s kin. In societies where descent was more flexibly defined, a

Kuma, for instance, found it “expedient...to

be uninterested in genealogies since the lack of strict reckoning facilitates assimilation into agnatic descent groups of nonagnatic kin” (Rubel and Rosman 1978:247). Thus, people recruited by big men may have in-

cluded nonagnatic kin (such as sisters’ sons)

and nonkin, but the ideology of the arrangement (that big-man followers were conceptu-

25

JEANNE

E.

external partners successfully gained wealth through goods and labor that flowed in

big man could call on a broader range of agnatic kin or could redefine the status of some of his affinal kin (as above, typically by en-

(Allen 1984; Brunton 1975; Feil 1987: 120-122). On the surface the ceremonial

compassing fictive kin into new agnatic kin categories), which enhanced his ability to col-

mok ink exchanges of the Mendi (Lederman 1991), the tee exchanges of the Tombema Enga (Feil 1984), and the moka of the pre-

lect wealth and influence. He was still limited

to accessing the labor of persons ultimately defined as “kin” (in the end, simply a somewhat broader group). the

contact Melpa (Strathern 1971) appear to have involved similar external manipulations

of wealth. Indeed, these three big-man—based systems were amazingly intricate and broadly linked more than 200,000 people by ceremonial exchange (Feil 1984). Yet patrilineal big

Allen (1984) makes the same point about

varying

flexibility

in kin

recruitment

among many other Melanesian patrilineal so-

cieties, where men could be added with lesser or greater difficulty to kin groups as agnates. Feil (1978:27) points out that for the

men in these societies did not traditionally control durable valuables and did not use the

extra prestige gained by participating in the most impressive pig exchanges to procure la-

Tombema Enga “fluidity is basic to clan com-

position,”

and

individuals

can shift alle-

bor or tribute within their own communities.

giance relatively easily. Thus, kin-designation systems are sometimes a flexible and power-

All extra pigs gained by the one to a dozen (or more) big men in any given community were

ful tool, but leaders and followers can cer-

committed to external exchange, and there was no corporate production of pigs (according to Feil 1984; Lederman 1991:225). Among the Mendi and Tombema every

tainly only manipulate them within limits. A North Abelam (Sepik River area) young man can, for example, withdraw his allegiance from a big man, but he can only transfer his fealty to another big man within the same subclan (Rubel and Rosman

ARNOLD

man prepared pigs and garden produce on his own behalf, and he neither contributed to

1978:56). This

leaders’ efforts nor consulted with leaders about his exchange plans. (A few groups may have had occasional corporate labor efforts in other arenas, such as Kuma land-clearing

inhibition is important because it prevented major shifts in male support of leaders outside the subclan. In my view rules regarding kin-based rights to labor remained the primary structural constraint to the expansion of tribal or big-man power everywhere. It appears that the fundamental structure of kinbased labor relationships had to rupture for person(s) with chiefly powers to emerge. In light of these patterns the effects of matrilineal organization are important to consider. Certain Pacific matrilineal groups

[Rubel and Rosman 1978:142] or Mae Enga

corporate decision making in scheduling or prioritizing occasional exchange transactions [Meggitt 1977].) Perhaps most important of all, the resources needed to produce pigs were available to all families in most areas

of

New

Guinea

(Feil

1978:77-79).

Ultimately, pig production could not be controlled per se; thus, a monopoly in pig production was unthinkable and economically impossible, and hereditary inequality could

(such as the Trobriand Islanders) developed

broad, outward-looking political relations in precontact times, cemented by ceremonial exchanges involving durable valuables, such as the kula ring. Adult males developed special crosscutting societies to strengthen and broaden male-controlled political ties and to

not emerge

on this basis. Moreover,

Tom-

bema big men, for example, never received pigs from their own village members or followers (this was considered akin to incest).

The movement of pigs was always into and away from the community. If it is not unreasonable to speculate briefly about sociopolitical evolution in the Trobriand Islands (I do not have access to requi-

overcome the narrowness (for them) of ma-

trilineal descent reckoning. Ambitious males in some cases developed extensive webs of relatives and contacts and, thus, allies in dis-

tant communities. Those who manipulated

26

PowER,

LABOR

RIGHTS, AND

site archaeological data), the process of lead-

the presence of a tier of leadership with the power and capacity to insist on allegiance in

ers using external network ties to command inflows of controllable, durable, prized prod-

the form of labor (or tribute). That is, there is

ucts—which were then allocated locally to secure local labor or tribute—likely contributed directly to the development of the Trobriand chiefdom. Once this system was in place, only certain people were in a position to effectively compete for leadership, and kula leaders could make it virtually impossible for anyone to “launch a new competitive idiom” (Allen 1984:36). Comparable kinship structures and a similar emphasis

at least one administrative level—beyond the household or descent group—that regularly exercises renewable rights to shares of labor. I suggest that a societal rupture must occur for leaders to begin to formalize rights to manage the labor of nonkin. Big men do not simply become chiefs one day by doing more of the same. The “recruit and relabel”

method of slightly expanding the pool of

people defined as kin, which is favored by great men and big men in Melanesia, has its limits and does not produce a chief and a nonkin labor pool. Through other mechanisms entirely, elders and lineage leaders must be compelled to relinquish their exclusive hold on kin labor (although household-based

on intensive exchanges may have character-

ized the final millennium of prehistory of the matrilocal, exchange-oriented, maritime Chumash of southern California (Arnold 1992). Matrilineality and/or matrilocality, access to local nonkin labor, and a strategy based on large-scale, external network

KINSHIP

ma-

work certainly continues). The major change

nipulations may be common threads in the process of these emergent chiefdoms. Such a pattern merits much closer, cross-cultural examination. Research by Feil (1987:120—122)

that occurs is that people begin also to respond to calls for food (e.g., for centrally

sponsored feasts), products (e.g., specialized

craft goods), or participation in general tasks

and Liep (1991) seems to confirm that matri-

(e.g., building structures) from leaders unre-

lineal kin reckoning in the Trobriands and elsewhere is with some frequency associated with chiefdom organization.

lated to them. How and why does this happen? Many theories have been advanced to account for this important transition, the emergence of centralized power, and I have discussed these elsewhere (Arnold 1993, 1996b). Theories based on internal factors include (but are not limited to) the appearance of charismatic and generous leaders whom commoners wish to emulate (Boyd and Richerson 1985; Spencer 1993); the appearance of ritual leaders who assert special ties to gods and ancestors (Aldenderfer 1993); and the emergence of differential access to unique, important resources that allow economic or technological dominance to develop (Arnold 1992, 1993, 1995; Earle 1991; Hayden 1998). External factors contributing to the emergence of centralized power may include warfare (Keeley 1996), environmental changes (Arnold et al. 1997), and other kinds of circumscription. Any of the following, then, have been argued to account for political transitions and thus transitions in labor rights and labor organization from kin-ordered to centralized in

TIES TO THEORY

Wolf’s (1981) concept of shares of labor is ef-

fective for envisioning how these rights to labor change. Labor can be conceptualized in terms of who has the right to redeem “stored” or potential labor held by each member of a society and how those members and all potential elders/leaders (those who can redeem) are situated relative to one an-

other. In small-scale societies elders or big men can call on labor shares of very close kin, usually on their or their wife’s side only, and they do not have the power to take shares from persons in other families or lineages.

There are many variations on this theme in

myriad small-scale societies around the globe, but typically the available labor pool is firmly defined by actual and fictive kinship relations. However, if both elders of households and lineages and unrelated leaders can routinely call on people’s labor, this signals 27

JEANNE

E. ARNOLD

various world areas: fear of or respect for those with special ties to ancestors and deities, outright conquest by oppressive opponents, widespread emulation of successful and aggrandizing leaders, changes wrought by environmental stresses, successful manipulation of economic or technological innovations, and the like. The merits of many of

some societies maintain an overt egalitarian

ethos and negative attitudes toward anyone accumulating wealth or influence, Hayden (1998) argues that a small percentage of every human

these theories have been debated at some

length, and there is insufficient space to renew this discussion here (see Arnold, ed. 1996; Hayden 1998; Price and Feinman 1995; many others). However, Feil (1984)

From

changed.

a transition from big-man to chiefly leadership might occur. Feil’s interpretation of recent Melpa history identifies how big men came to control the supply of pearlshells from the coast and began to bestow shells locally in return for “political allegiance, patronage, labour, or... prestige” (Feil 1984:83). This process profoundly altered Melpa social relaleaders

could

control

historical

and

archaeological

edly in the past as power relations have

Guinea, worth briefly citing to illustrate how

Rising

the

records, it is clear that significant transitions in labor organization have occurred repeat-

provides an excellent case study from New

tions.

population strives to accumu-

late wealth and gain power. Social change may be regularly stimulated by these strivings, although their assumed universality means that they do not alone explain why more powerful leaders successfully emerge in some places and not in others.

tion,

which

The

specific conditions

environment,

these

subsistence,

changes

occurred

(popula-

etc.) under

have

been

highly variable and will likely defy any at-

tempt at generalization. But on a broader level it can be argued that an explanation of

the origins of centralized power can be linked

to the alienation of labor rights from elders, big men, and other leaders of tight kin

bride-

wealth (which had come to include pearlshells), and thus they controlled young men’s

groups. I turn at this point to some

most

salient

archaeological

of the

indicators

of

ability to secure marriage partners. The gulf

these phenomena.

rank began to appear as successful leaders in-

To examine power and labor in the past, we

be married. Leaders invested in trade part-

reflect how leaders made changing use of the

between leaders and male followers widened and became more political, and a system of

ARCHAEOLOGICAL

IMPLICATIONS

can consider sets of behaviors with relatively pronounced material consequences that

creasingly demanded allegiance and contributions from other local men who needed to

labor of others. For instance, changing orga-

ners far afield, and they became essential links between their villages and the shell sources. These leaders thus came to coordinate group activities and community labor in a completely novel way. In short, an opportunity to control a prized, rare resource led eventually to a breach in baseline norms of marriage arrangement and a rupture of descent-group control over labor as leaders centralized power. In the end it is not the emergence of leaders and followers per se that we seek to identify archaeologically because persons filling these roles appear in at least nascent form in every society, and, indeed, struggles for elevated social position occur among all of the higher primates. Even though, as is well known,

nization of labor may be revealed through initial appearances of coordinated, large-

scale production of goods or through signs

that labor forces of greater sizes began to in-

teract cooperatively (as in major ventures involving multiple clans, villages, or polities). When the following kinds of transformations can be detected, it is reasonable to assert that

labor relations have changed to a significant degree (Arnold 1987; Clark and Parry 1990; Costin 1991; Hayden 199§a). Capital improvements, surpluses, cooperation. Considerable labor begins to be invested in capital improvements at resource

extraction loci (fishing platforms, terraces, irrigation systems, etc.), which then generate 28

POWER,

LABOR

RIGHTS,

staples in amounts that far exceed the an(e.g., the extended household or lineage). This development points to the initiation of cooperation

and/or

KINSHIP

prestige tools and display items (including elaborated tools, alcoholic beverages, complex boats, metals, fancy vessels, jewelry, and

nual needs of the largest local kin group interkin-group

AND

more)—many of them labor-intensive to produce—may have many material conse-

central

management to process and store staples for multiyear consumption, for hosting large feasts (see below), or for trade. Economic interdependencies, craft specialization, coordination. People begin to de-

gins to be supported by large subsistence

on a scale unlikely to be household- or lineage-based, a function of multiple, local-

and symbols or motifs associated with those vessels (Marcus and Flannery 1996) may pro-

Arnold and Munns 1994). These interdepen-

iors, as a number of recent and well-received studies have demonstrated. Settlement hierarchies. More complex economic and labor interrelationships may also be manifest in the loss of village autonomy and the rise of regional leadership centered in one or more communities. Leaders in central communities may draw on the products or participation of people in satellite/dependent communities. This may be evident in several ways, including unusual concentrations of elaborate architecture representing more expensive and labor-intensive construction materials, high proportions of exotic valuables, or concentrations of locally produced and prized status markers in small percentages of communities in a region. Certainly, this represents only a partial list; I encourage other archaeologists to consider these problems closely, with changing power and labor rights in mind, and to refine and expand this short roster. I would highlight major capital investments in permanent extraction or production facilities and sharply increased uses of intensifiable resource areas as common indicators of centralized (cooperative or coerced) labor organization. The first appearances of major technological innovations requiring considerable labor support, such as plank-boat making or bridge building, undeniably indicate significant overhauls to labor rights arrangements. Likewise, the emergence of large-scale competitive feasting activities and large-scale craft specializations also suggest new, higher-level political or economic organization. Moreover, and not

velop

massive

economic

quences visible in the archaeological record. Feasting. Major competitive feasting be-

efforts drawing on far more than a single lineage’s or clan’s labors. Architectural elements, food refuse, suites of serving vessels,

interdependencies,

ized intensification in specialized crafts (e.g.,

vide clues to the presence of this set of behav-

dencies, often based on distributions of con-

trollable unique resources, likely require administrative coordinators who guide the production processes and may themselves be specialists. Intensification of exchange, media of exchange. Exchange clearly intensifies—in tandem with capital improvements and economic interdependencies—and many subsistence and/or prestige goods circulate. Dramatic qualitative and quantitative differences from antecedent exchange systems arise, and specific media of exchange, such as tokens or beads, often develop to govern large-scale localized trade systems. Technological innovations,

external con-

tacts. Important technological innovations are stimulated and underwritten by aspiring leaders who have accumulated sufficient wealth, knowledge, or external contacts to bring new and productive technologies to their communities. Innovations in transportation may be especially important here, a point I explored a few years ago with regard to the plank canoe on the American Pacific Coast (Arnold 1995). Such innovations may ramify widely. As Kuhn and Sarther (this volume) imply, technological expansions can reflect, in material form, the creation of new webs of social interaction and power (see also DeMarrais et al. 1996). Hayden (1998) discusses prestige technologies in a similar light; indeed, he sees the development of prestige items as a rationale for gathering labor to manipulate in various ways. The emergence of 29

JEANNE

E. ARNOLD nonkin

mentioned above, the escalation of violence

labor in hierarchical

(intermediate)

societies (Arnold 1996a, 1996b). In this discussion I hope to have stimulated thinking along several lines: (1) to suggest that broad organizational rules related to power, rights to labor, and kinship are important in classifying societies (to the degree that types may remain heuristically useful); (2) to highlight the utility of a close examination of leadership and who holds the rights to draw on shares of labor in understanding the evolution of power; and (3) to underline the point that there is an important qualitative break in leadership and labor organization between small-scale kin-structured societies and formal hierarchical societies led by chiefs. This rupture can be archaeologically

and intensive factional conflict may signal important changes in leadership rules (Brumfiel and Fox 1994), although of course societies at all levels may exhibit structured violence. In my view the appearance of storage facilities or signs of increased sedentism alone do not herald deep-seated changes in power and labor organization (Arnold 1996a). As labor rights change, a significant proportion of the power held by traditional elders and other kin-based leaders in smallscale societies is usurped by emerging and powerful chief-like leaders (Figure 2.2). The

rising chiefs establish and then repeatedly exercise their rights to call on nonkin labor, consolidating their power. Earlier, I suggested

identified in a number of ways.

several ways this transition in power and

leadership might occur, including seizing rit-

Acknowledgments. I was pleased to accept Michael Schiffer’s invitation to participate in

ual or economic control. I expect such changes to be punctuated, disruptive transi-

tions, visible in archaeological sequences as

the 1997 Snowbird Conference. Mike pro-

change, preceded and followed by equilibria, and often eventually cycling back to simpler

comments as this chapter evolved. I have benefited as well from useful comments by

periods

of rapid

economic

vided a number of incisive and constructive

and _ political

organization (see Feil 1987; Scarry 1996). Recent detailed reviews of archaeological

Joyce Marcus on a 1998 draft of the paper. Susan Kus, reviewer Kathleen Morrison, and

Northwest Coast, and California provide ev-

pertinent

and

ethnographic

cases

from

Florida,

an anonymous

the

reviewer

additional

suggested

references.

several

Anthony

Graesch prepared Figure 2.1. To each, I am grateful, and I am responsible for any omis-

idence to complement the evidence presented here. Together, these data support the model of a basic distinction between rights to labor in small-scale societies and leaders’ rights to

sions and errors remaining herein.

30

3

Corporate/Network New Perspectives on Models of Political Action and the Puebloan Southwest GARY

M.

FEINMAN

From its inception one of the major strengths and key themes of American archaeology has been its endeavor to understand sociocultural similarities and differences through comparative analyses (e.g., Spencer 1997). Yet during the last half century such comparative

argue against the oft-held view that political complexity can be equated generally with (or measured by) marked centralization or rampant self-aggrandizement or individualism. A key theme here is that if centralization is defined as the greater consolidation of wealth and political power in the hands of a single individual or ruler, then larger, more hierarchical polities are not always (or by definition) more centralized. In contrast to a more monolithic perspective on hierarchies, this essay discusses and develops the distinction between two modes of political organization, which my col-

efforts have tended to focus principally on a

single cross-cultural dimension: hierarchical complexity. In the full scope of societal variation the dimension of complexity is unquestionably an important axis of comparison. At the same time, however, the general acceptance of certain monolithic notions regarding societal complexity (and its assumed direct

leagues and I (Blanton 1998; Blanton et al. 1996; Feinman 1995, 19974, 1997b) have re-

correspondence with economic stratification

and the centralization of power in the hands of a few people) has tended at times to obscure other important axes of organizational variation and to cloud our understanding of specific pathways of historical change (Blanton et al. 1993:10-23; Chapman 1996;

ferred to as the network or exclusionary mode, on the one hand, and the corporate mode, on the other. These modes refer to distinct strategies for achieving and maintaining power. They are viewed as strategies that have structural outcomes, and both strategies are found in all kinds of societies, from states to bands, and at different levels of social integration (from household on up). In this sense my use of strategy specifically is not intended to convey goal-oriented political-economic behaviors that necessarily achieve prescribed historical outcomes or teleological ends. In the network or exclusionary mode power is focused on individuals and their personal networks (e.g., descent genealogies, exchange partners). In contrast, with the corporate mode access to power is less individu-

Drennan 1996).

This chapter questions the traditional monolithic models that presume that hierarchical complexity, greater wealth stratification, and the increasing centralization of power always directly co-occur. I challenge the theoretical expectation that political hierarchy formation necessarily entails the stark

consolidation of power and wealth by a few

individuals or specific families, a concept central to theoretical perspectives as diverse as certain strains of neoevolutionism, Marxism,

and selectionism. To put it more succinctly,I 31

GARY

M.

FEINMAN

HIERARCHICAL STATES

CORPORATE

NETWORK

EGALITARIAN Figure 3.1. Two orthogonal dimensions for the comparison of societal organization.

By augmenting the theoretical tools that archaeologists can use to compare societal variation and change systematically, I endeavor to provide a broadened framework that can help us understand social organizational diversity in a wide range of societal contexts. At the same time, the approach offered here expands traditional processual perspectives (which tend to consider only hierarchical complexity in comparative fashion) through a greater consideration of diverse strategies of political-economic behavior. Following a general and comparative review of the corporate/network dimension, these conceptual tools are applied in a reconsideration of a specific geographical and his-

alized, more apt to be shared, and often a product of group membership. In most settings social actors typically have the ability to put people and resources together in limited and constrained ways toward goals and ends that also are circumscribed. In given cultural contexts power may be built in certain ways but not others (although this can change over time). For that reason corporate and network strategies frequently tend to be opposed. In proposing these modes our aim is not

to devise or apply a new societal typology. Rather, I present the corporate and network modes as the polar ends of a more continuous dimension that can be conceptualized as orthogonal to (or crosscutting) the familiar axis of hierarchical complexity (Figure 3.1). In other words, the corporate/network or exclusionary dimension is not meant to replace a concern with hierarchy but to expand our comparative theoretical concepts so that they may better account for more general patterns of societal variation and change. Both the corporate and network strategies can be envisioned at any degree of societal complexity because they represent different pathways of sociopolitical behavior or action.

torical context, the Puebloan Southwest, the

interpretation of which has generated more than a century of heated debate. Although some scholars view these historic and precontact societies as egalitarian, with purely consensual modes of decision making (e.g.,

Ciolek-Torrello et al. 1996; Graves et al. 1982:202), others, including myself, have in-

ferred that at least certain Puebloan populations were more hierarchically organized. Yet 32

MODELS

OF

POLICITAL

ACTION

until recently all parties have tended to approach the issue from a perspective grounded largely in a rather monolithic notion of lead-

This relationship stems from the basically le-

tion. Perhaps for this reason neither theoretical “camp” has been able to convince the

the land cannot exist without a supreme law-making authority whose decisions are

record. Here the consideration of the Pueblos be-

ethnographic and documentary materials. I suggest that the organization of many recent

This conceptualization is not at all surprising given that the word state is said to have been fixed as a generic term for the body politic by the political philosopher Niccolo Machiavelli early in the sixteenth century, at

egalitarian model, a traditional conception of

Machiavellianism today refers to the seizure,

galistic assumption that all political societies

are, or ought to be, united under a deterministic rule of law, and such a definitive law of

ership, stratification, power, and centraliza-

other or explain key elements of the empirical

final.

gins with the historic period and a review of

which time it was used in the form stato.

Pueblos does not closely correspond to the

maintenance,

a stratified/hierarchical society, nor a notion of utopian communalism (e.g., Saitta and Keene 1990). Rather, recent interpretations

These

NOTIONS

absolute

1968).

These individualizing roots of the state concept in Western thought and Renaissance practice clearly were influential to Morton Fried (1967:237), who wrote that “the state must establish and maintain sovereignty, which may be considered the identification and monopoly of paramount control over a population and an area.” Elman Service (1971) adopted a similar focus to leadership in his conception of chiefdoms. Almost by definition chiefdoms have been viewed as the rule over a social group by one individual. For example, in a classic overview article Robert Carneiro (1981:45) writes that “a chiefdom is an autonomous political unit comprising a number of villages or communities under the permanent control of a

considerations

are intended to inform our understanding of Puebloan societies, as well as our more general notions of societal variability and long-term change. MONOLITHIC

of

state makers, France’s Louis XIV, is said to have remarked “LVétat, c’est moi” (Fried

(e.g., Brandt 1994) seem to correspond to more corporate hierarchical formations. In subsequent sections of this paper I turn to the deeper, precontact past, where the corporate/network dimension sheds light on certain later Puebloan societies, as well as the oft-interpreted pithouse-to-pueblo transition 800-1100).

extension

Machiavelli, one of Europe’s early modern

of modern Pueblo sociopolitical organization

(ca. A.D.

and

power through the use of guile, fraud, force, and terror. Just more than a century after

OF HIERARCHY

During the last decades traditional models of political process and behavior in archaeology have tended to equate centralization (the concentration of wealth and decisionmaking power in the hands of a limited few) with increases in societal scale and complexity. This rather pervasive association of individual aggrandizement and power with political hierarchy comes rather easily to Western social science. In the International Encyclopedia of Social Sciences, the concept of state is defined as “the geographically delimited segment of human society united by common obedience to a single sovereign” (Watkins 1968:150). From its origins this normative concept of state has been closely linked to the concept of sovereignty, which was developed in the field of jurisprudence.

paramount

unit.”

Of

course,

in these

ev-

olutionary schemes chiefs lack the regular and institutionalized administrative structure typical of more hierarchical states. The equation of increasing hierarchy with greater centralization was solidified in archaeological thought through Kent V. Flannery’s (1972) seminal paper, “The Cultural Evolution of Civilizations.” Taking a systems approach, Flannery proposed centralization as one of the two core processes of cultural evolution. Although Flannery’s own thinking

33

GARY

M.

FEINMAN

on this topic has developed in some new di-

rections

(e.g., Marcus

and Flannery

organizational modes would leave rather distinct archaeological correlates. Building factions, amassing wealth, adorning and displaying oneself in an ostentatious way clearly are key facets of one pathway to power that I refer to as the network mode. But a key point here is that hierarchical organization also can take a somewhat more “faceless” and less centralized or individualizing form (which we refer to as the corporate mode). Furthermore, if we are to build convincing and broadly relevant comparative models, they cannot disregard important axes of variabil-

1996),

his important paper has had great influence across the discipline. The association of hierarchy with the centralization of power and the consolidation of basic resources by a few individuals has been argued to underpin the emergence of the state in the conflict theoretical approach of Jonathan Haas (1982: 212-213). At the same time it is a fundamental premise of Paul Roscoe’s (1993) practice approach. Recent postprocessual and selectionist emphases on individual strategies have further promoted the presumption that highly centralized and personalized leadership, control, and aggrandizement underlie most, if not all, hierarchical political organizations (although

ity. Likewise, if we are to construct convinc-

ing diachronic/historical models, we model and examine change along more one comparative dimension. Toward aims the next section of this paper more

see Kus this volume for a notable exception). This notion is most pervasive in recent discus-

must than those fully

defines the corporate and network modes.

sions of chiefly societies, where we see fre-

THE

quent references to accumulators (Hayden

CORPORATE

AND

NETWORK

MODES

In recent years many archaeologists have critiqued the assumption of directional

1990), aggrandizers (Clark and Blake 1994),

progress that lies behind some comparative

strivers (Maschner 1995), and entrepreneurial elites (Hayden 1995 a). Basically, these

approaches to long-term change (see also Gould 1988). Yet at the same time, most com-

models link the processes behind chiefly emergence to individual self-interest and the strategies of accumulation and personal networking that are undertaken to foster that self-interest. Although the specific factors

parativists would agree that the dimension of hierarchical complexity (the number of decision-making levels above basal domestic units) still has significant cross-cultural impli-

and the proposed sequences of change are not

cations

uniformly agreed on, each of these models views individual connections, status display,

for understanding

other

important

aspects of societal variation (as opposed to the progressivist notion or presumption of increasing complexity through time). For example, it would be hard to refute the statement that hierarchically organized societies generally have larger settlement sizes than socleties organized in less hierarchical fashion (e.g., Feinman 1995:259-261). This observed association should not be phrased as a general law or promote an expectation for simplistic thresholds, but it does have rather broad empirical support and raises interesting avenues for new research (e.g., Johnson

and unrestrained amassing of wealth as basic

and generalized foundations to inequality and its institutionalization. My argument here does not contest the viability of these ruler-centric models for specific regions, polities, and cultural settings. Likewise, I certainly do not take issue with the theoretical consideration and importance of differential self-interest per se. Yet I do question whether self-interest is always manifested in the same way, and whether it is necessarily as blatantly enacted in all cases as some of these networking models propose. More to the point, a central premise of this chapter is that there is more than one strategy of political action or pathway to power (see also McGuire 1983) and that these different

1982; Kosse 1990).

Although hierarchical complexity is recognized here as an important dimension for

societal comparison, recent comparative ap-

proaches (following the general foundational

tenets of Fried and Service) have tended to 34

MODELS

OF

POLICITAL

ACTION

Table 3.1. Political/Economic Typologies that Parallel the Corporate/Network Distinction

Network/Exclusionary

Corporate

References

Individualizing chiefdoms

Group-oriented chiefdoms

Renfrew (1974), Drennan (1991)

Gumsa

Gumlao

Leach (1954), Friedman (1975)

Prestige-goods systems

Big-man competitive feasting

Friedman (1982)

Wealth-based

Knowledge-based

Lindstrom (1984)

Material-based

Magical-based

Harrison (1987b)

Finance-based big man

Production-based big man

Strathern (1969)

Noncorporate organization

Corporate organization

Schneider, Schneider, and

Hansen (1972)

Wealth finance

Staple finance

D’Altroy and Earle (1985)

Wealth distribution

Staple finance

A. Gilman (1987)

Twem exchange

Sem exchange

Lederman (1986)

focus almost exclusively on this axis, often conflating complexity with centralization, expanding scale, and stratification in a rather monolithic fashion. When recognized, aspects of diversity that cannot be subsumed by

general political-economic modes or paths of action (corporate based and network based) are envisioned as dual strategies to reach and/or to maintain hierarchical organization. As stated above these modes are not seen as reified or immutable societal types. Likewise, the corporate-based and network-based strategies are not proposed as mutually exclusive categories. In fact, to a degree, these strategies coexist in the political dynamics of all social arrangements. However, because these modes also are structurally antagonistic (see also Strathern 1969:42-47), these alternative strategies of achieving and retaining power have been found to have different degrees of relative importance cross-culturally. Their relative significance also may vary ina single region over time (Blanton et al. 1996).

this overarching monolithic dimension often

then are treated as largely inexplicable in comparative terms because of culturally unique or historically specific factors. The inclination toward rather monolithic conceptions of complexity has been so dominant that it has obscured repetitive observations of cross-cultural variability that may reflect different modes and strategies of organization. The notion of heterarchy (Crumley 1979; Ehrenreich et al. 1995), a more multidimensional approach to variation, correctly underlines problems with monolithic perspectives on hierarchy, yet so far it has not specifically defined general axes of variation or cross-cultural archaeological correlates that will allow us to expand comparativist research. The corporate/network continuum is proposed as one such analytical axis or dimension. The

corporate/network

continuum

Therefore, one or the other mode often domi-

nates in any particular spatiotemporal setting (Blanton et al. 1996; Strathern 1969:42). I explicitly do not wish to conflate corporate or network modes with specific cultural or ethnic “personalities.” One cannot say that the Maya always followed network strategies, just as one would not characterize Maya society as always having been organized as states (e.g., Marcus 1995). The corporate and network strategies are neither static in a spatiotemporal sense, nor are they culturally bound in a manner that implies that specific groups of people follow specific

out-

lined here parallels similar distinctions in leadership strategies, organizational diversity, and political economy that have been

noted in earlier comparative analyses (e.g., Drennan 1991; Lehman 1969; Renfrew

1974; Strathern 1978) (Table 3.1). These two

35

GARY

M.

FEINMAN

strategies immutably. Likewise, once network- or corporate-based strategies are in place, one should not presume that the

ral contexts define and place specific cases along the corporate/network continuum. In an important argument that fore-

specific political-economic strategy will necessarily remain predominant or unchanging

shadowed our corporate/network distinction, Renfrew (1974) contrasted individual-

throughout a specific culture-historical or re-

izing and group-oriented forms of chiefdoms.

gional sequence. For example, the ancient

The cases he outlined were recognized as relatively similar in overall sociopolitical complexity yet organized in markedly differ-

Romans had a republican government that eventually underwent key transitions. At a later date, during the Roman Imperial period, power was more firmly concentrated in the

ent ways

hands of specific emperors. Political strate-

gies shifted greatly over these centuries, so Roman political behaviors as a whole cannot be categorically placed into either the corporate or network mode. Rather, it is the com-

1974:74).

For

Renfrew

link the social and geographic segments that constitute these populations. Collective labor

parison across Roman history that is most in-

and monumental public architecture were evidenced, whereas great stores and individual displays of personal or portable wealth were rarely apparent. Neolithic Malta, the henge-building peoples of the European third millennium B.c., and ethnographic Polynesia were interpreted as examples of these grouporiented arrangements (see also Feinman

triguing, with two key indicators of a more

network strategy—opulent ruler palaces and colossal portrait statues of the powerful emperors—first having appeared during the early Imperial period, when a series of authoritative individuals ruled the ancient city (Reid 1997). From a cross-cultural perspective the corporate and network modes may have their basis in the external and internal components of the “dialectics of control”

(Renfrew

(1974:74-79, 83), group-oriented polities deemphasized differentials in access to personal wealth and placed great importance on communal activities and group rituals that

1995).

The emphasis was entirely different in Renfrew’s individualizing forms, like those of

Minoan-Mycenaean

(see Giddens

1984:374; Spencer 1993). That is, those in leadership positions generally must balance to a degree what have been called “selfserving” and “system-serving” goals and interests (e.g., Flannery 1972:411). Although these dual concerns have been proposed often as opposing functionalist versus selectionist (or political) models (e.g., Hayden and Gargett 1990; Paynter 1989) or as representing different stages in the stepladder of cultural evolution, both self-serving and systemserving considerations generally are at work in most hierarchical arrangements. Aspiring leaders who address and resolve the wants and needs of followers (larger societal concerns) simply are more apt to be retained and supported by factions despite other potentially selfish aims and strategies of those individuals (see also Feinman 1995:262-263). Nevertheless, the ways in which these different considerations and strategies are emphasized and balanced in different spatiotempo-

Greece (Renfrew 1974:

74-79, 84). The egalitarian ethos was overcome as specific individuals were differentiated and privileged. Marked disparity in personal possessions existed, and the residences and tombs of rulers were far grander than those of the bulk of the population. Exotic trade wealth appears to have had a significant role in prestige accumulation. Some of the goods that were exchanged were elaborately crafted by specialists. In contrast to the

group-oriented formations, communal ritual

and public construction appear to have had lesser roles. Significantly, in a more recent comparison of Prehispanic chiefdoms in the Americas, Drennan (1991:283) found a similar distinc-

tion. In two cases (central Panama and Alta

Magdalena)

from

the

Intermediate

Area

(lower Central America and northern South

America), almost all public building was focused on individuals. Status differentiation was conspicuous. In the best documented of 36

MODELS

OF

POLICITAL

these cases (central Panama) tremendous volumes of portable wealth, reflecting great craftsmanship, were found in select burials. The items recovered also signal the importance of extraregional exchange links. Many of the burial goods evidently were meant for

ACTION

icas, both Renfrew (1974:74, 82-85) and Drennan (1991:284-285) have argued that long-term social change in chiefdoms has followed distinct paths that have involved some-

what different political-economic strategies and processes. Yet these different modes or

elaborate personal adornment. “A large por-

strategies also are evidenced in other societal

tion of the resources mobilized in the local

settings as well. In Strathern’s (1969, 1978) comparative analyses of nonstratified soci-

economies in central Panama and Alta Mag-

dalena went into competition over status dominance, focused heavily on the person of

eties in the New Guinea Highlands, two orga-

nizational strategies (home finance and home production) also are outlined. Home finance

the chief” (Drennan 1991:283). These cases illustrate clear parallels with Renfrew’s indi-

defines a mode in which the relative impor-

vidualizing chiefdoms and with what we call

tance of big men reflects their position in a

In contrast, a markedly different organizational mode was evident in two Early-Middle Formative-period Mesoamerican populations (Valley of Oaxaca, Basin of Mexico). In these regions collective works primarily were

1969:65). Through the external exchange of portable wealth items (e.g., shell valuables,

the network strategy.

network of financial arrangements (Strathern

live pigs), obligations and alliances are cre-

ated (Strathern 1978:75). Big men are able to solidify their factions through the local distribution of long-distance gifts derived from trade partners. As described for the Melpa of Mount Hagen and the Enga, this mode bears striking resemblance to the strategies employed by the Panamanian chiefs (see also Helms 1979). The acquisition of wealth through individual, entrepreneurial linkages and the use of that wealth by these charismatic figures to attract factions defines our network-based mode. With home production a big man’s prestige depends on the labor force of his own kin group or settlement to raise the goods needed for feasting and exchange (Strathern 1969: 42). For household or settlement heads, access to labor and land is critical, and it is derived largely through corporate descent. As described for the Maring and Siane, the eminence of a big man is achieved by obtaining

carried out to construct public spaces (plazas

and mound groups) for communal ritual. Although inequalities existed (as they did in Renfrew’s group-oriented chiefdoms), they were not expressed elaborately in house construction or burials. For example, at this time in the Valley of Oaxaca, all houses were made of comparable

materials

(earth

and

cane),

and the most elaborate burials were simply lined by stones and contained no more than a few exotic offerings. Certainly, craft specialization, long-distance exchange, and status

competition were not entirely insignificant in

these highland Mesoamerican regions during this period (e.g., Feinman 1991; Marcus and Flannery 1996). Nevertheless, I concur completely with Drennan’s (1991:283) perception that these activities consumed a relatively smaller proportion of the “chiefly domestic product” as compared to the aforementioned cases from the Intermediate Area.

Thus,

these

Formative-era

more land and labor than others (Strathern

1969:65). Siane big men do not accumulate great personal wealth (Strathern 1969:50). Intergroup ties, portable wealth, and financial maneuverings are far less important with home production than they are with home finance (Strathern 1978:98-99). In Siane

Mesoamerican

cases illustrate elements of both the network

and corporate strategies (see Drennan 1991);

however, considerable emphasis is on the latter mode, particularly when comparisons are made with the chiefdoms of the Intermediate Area. Focused primarily on these archaeological examples from Europe and the ancient Amer-

communities corporate groups are integrated

through initiation rituals that crosscut component social segments. Such ceremonies are described as much less significant among the 37

GARY

M.

FEINMAN

Enga and the Melpa (Strathern 1969:44). Home production fits comfortably within the corporate-based mode (Blanton et al. 1996), with its emphases on food production, land, labor, corporate descent, societal segments, and the integration of those segments through crosscutting integratory social and ritual mechanisms. Many of the contrasts noted by Strathern (1969,

1978)

for New

Guinea

the ability of Akwe-Shavante leaders to build factions beyond their lineages. These include age sets, which comprise community members of a certain age and gender, and a council of senior men that regularly discusses key village matters with the be’a. Akwe-Shavante age sets play a key role in initiation rituals and other ceremonial activities (MayburyLewis 1974:147). He’a lead community hunts and direct horticultural clearing, planting, and harvesting. They often serve a central role in various ceremonial activities (including distributions

populations

appear to parallel distinctions drawn across the world between two lowland South American peoples, Mekranoti-Kayapo and the Akwe-Shavante (Spencer 1993:42-45; Werner 1981). Mekranoti-Kayapo leaders (benjadjwyr) are described as “culture brokers” who cultivate external ties with individuals in other communities. In part these

of food), where they act on behalf of the com-

munity (Maybury-Lewis 1974:197-201). A he’a derives prestige from his athleticism, oratory skills, hunting abilities, sense of humor, and ceremonial knowledge; however, the size

links are made so that the leaders can obtain exotic wealth items, which are then dispersed

of his corporate faction is the basis of whatever power he may have (Maybury-Lewis

1974:197).

as gifts to secure and solidify the allegiance of followers. The prestige and factional support of a Mekranoti-Kayapo benjadjwyr hinge in

The

strength

of

an

Akwe-

Shavante leader is rather fleeting, and a single

community can have more than one he’a, each with his own corporate faction. Akwe-

large measure on that person’s network of external connections. This privileged net-

Shavante

leaders have no insignia of office

work of linkages to the external world often appears to be successfully transmitted from a

and few prerogatives, yet they are relatively

Given the importance of personal contacts

ally

successful at the manipulation of age-set organization, so that their own sons eventu-

benjadjwyr to his offspring (Werner 1981).

to prestige,

individuals

attract

attention

are

positioned

to

exercise

whatever

leadership capabilities they possess (Maybury-Lewis 1974:190). Whereas the Akwe-

through various means, including elaborate body ornamentation and the production of

Shavante parallel Strathern’s home produc-

handicrafts (Werner 1981:365-366).

tion and what we have called the corporate

Among the Akwe-Shavante, leadership also depends on factional support, but longdistance exchange contacts are not stressed, and exotic items do not have a central role in attracting followers (Maybury-Lewis 1974: 172, 205-2133; Spencer 1993:44). Akwe-Shavante leaders (he’a) derive support from a corporate group, which primarily includes the members of their patrilineage (MayburyLewis 1974:169). A powerful he’a may attract other coresidents of his village, as well as clan and lineage members from outside his community. The personal eminence of a he’a is therefore based on the support of his lineage and of any other men who are prepared to join his faction (Maybury-Lewis 1974: 148). Several crosscutting associations foster

mode

in

certain

key

ways,

Mekranoti-

Kayapo political behavior corresponds more closely to Strathern’s home finance or the net-

work mode. To synthesize this far-ranging discussion, the corporate mode emphasizes communal ritual, public construction, food production, large cooperative labor tasks, social segments that are woven together through broad integrative ritual and ideological means, and suppressed economic differentiation. Despite the presence of large architectural spaces, Renfrew (1974:79) aptly described individuals in such polities as relatively “faceless” and “anonymous.”

In

contrast,

the

network

mode places greatest significance on personal prestige, wealth, power accumulation, elite 38

MODELS

OF

POLICITAL

ACTION

Table 3.2. Tendencies of Corporate/Network Modes. Network

Corporate

Concentrated wealth

More even wealth distribution

Individual power

Shared power arrangements

Ostentatious consumption

More

Prestige goods

Control of knowledge, cognitive codes

Patron/client factions

Corporate labor systems

Attached specialization

Emphasis on food production

balanced accumulation

Wealth finance

Staple finance

Princely burials

Monumental ritual spaces

Lineal kinship systems

Segmental organization

Power inherited through personal glorification

Power embedded in group association/affiliation

Ostentatious elite adornment

Symbols of office

Personal glorification

Broad concerns with fertility, rain

agerandizement, highly individualized leadership, lineal patterns of inheritance and descent, personal networks,

social formations (or “egalitarian societies” in the sense of Fried). This discussion challenges the unilinear proposition (Spencer 1993) that cultural evolution necessarily moves either simply and directly from more consensual to simultaneous (individually focused) forms of organization or toward greater centralization. Rather, corporate and network strategies are presented as organizational modes that, to a great degree, crosscut societal variation in relative hierarchical complexity and stratification. One might view the corporate-to-network continuum as a dimension for the comparison of political economies that runs perpendicular or orthogonally to the long-recognized axis of hierarchical complexity. That is, it is argued that one can have societies organized either corporately or in more network fashion at each different degree or level of hierarchical complexity.

long-distance ex-

change, exotic wealth, princely burials, and the specialized manufacture of status-related craft goods (Table 3.2). The corporate/network continuum bears some similarity to the distinction previously drawn between societies having simultane-

ous (network) and porate) hierarchies

sequential-ritual (cor(Johnson 1982). Of

course, Gregory Johnson’s (1982) discussion differs in its narrower concern with decision making, avoiding direct tie-ins with economic considerations. Simultaneous hierar-

chies (which conform to the more traditional notion of hierarchies) are defined as social

arrangements in which a few central individuals exercise integration and control over a larger population (Johnson 1982:403), whereas sequential-ritual hierarchies are described as relatively more consensual and egalitarian formations in which ritual and elaborate ceremonies are argued often to have key roles in the integration of modular social segments 378-381).

(Johnson

1982:405,

Corporate

hierarchies may

be character-

ized by more power sharing, greater depersonalization of rule, and less flaunting of wealth than is traditionally conceptualized for hierarchical societies. Yet corporate hierarchies should not be equated with political or economic equality or utopian communalism. In corporate hierarchies certain individuals are more apt from birth to rise to ruling positions than others (based on kin affilia-

1989:

However, in contrast to Johnson’s discussion regarding sequential and simultaneous hierarchies, I have argued that corporate strategies are not restricted to nonstratified 39

GARY

M.

FEINMAN

tions or other factors). And once in such po-

Elsewhere, my colleagues and I (Blanton et al. 1996; Feinman 1997b) have discussed ur-

marked disparities in wealth, depictions of individualized rulers) that are used to define and distinguish different forms of sociopolitical behavior in the archaeological record (cf. O’Shea and Barker 1996; Wason 1994). Such correlates may be completely appropri-

(A.D. 200-600) as having been organized as

marily through network strategies; however,

sitions, these individuals can make key decisions that fundamentally affect the basic social

and

economic

well-being

of others.

ate for polities in which power is derived pri-

ban Teotihuacan during the Classic period

a corporate hierarchy. Although leadership at that Central Mexican metropolis was ap-

they may be less apt where more corporately organized strategies predominate. In the re-

contemporaneous Maya), it certainly could not have functioned as a fully communal so-

discuss the ancient Southwest.

parently

faceless

maining sections I build on this foundation to

(in stark contrast to the

ciety with an urban population of roughly

100,000 people. Yet at the same time there are no personalized or named rulers, no clear

palaces,

relatively

minor

disparities

THE

HISTORIC

and

AND

recent Puebloan

ETHNOGRAPHIC

PUEBLOS

Scholars have long debated the nature of pre-

in

wealth, and no wealth-laden graves (for the

contact and modern Puebloan political orga-

much that we still do not understand about the government and economy of Teotihua-

tributed to these controversies, part of the problem lies in the inference that Puebloan

period

A.D.

200-600).

Although

there

nization. Although many factors have con-

is

societies, past and present, do not neatly con-

can, strong corporate codes, redistributive

rituals associated with fertility, and a segmental aspect that included multihousehold

form to the monolithic model of societal complexity, nor can they be easily or unequiv-

plans were key features of its organizational structure (e.g., Cowgill 1992a, 1992b,

taxonomies. To a degree the debate over pre-

apartment 1992c,

compounds

1997;

and

Manzanilla

ocally shoehorned into the familiar societal

quadripartite

1997;

1988, 1992; Pasztory 1992, 1997).

historic southwestern political organization has its foundation in disagreements over more recent Puebloan ethnographic and historic accounts and the relevance of the more contemporary formations for understanding the past. Many interpretations of ancient Puebloan egalitarianism stem from the liberal use of the direct historical approach or the application of more recent analogs to interpret the past. Consequently, it is instructive to begin with a reconsideration of the nature of historic Pueblo organization (with a focus on

Millon

To this point I have endeavored to illustrate that hierarchical formations (at different degrees of complexity) may exist without high degrees of centralization, blatant expressions of economic stratification, or highly personalized leadership. When these familiar aspects of more centralized network organization are absent, more corporate or-

ganizational modes seem to be found (as at Classic-period Teotihuacan, in Renfrew’s group-oriented chiefdoms, and among the Akwe-Shavante). In each of the latter, political offices rather than specific personages seem to be emphasized, and corporate codes that may help achieve order and integration through ritual are found (see also Blanton et al. 1996; Feinman 1995). Furthermore, if we recognize the existence of corporate forms of organization, then we must reevaluate and broaden the correlates of hierarchical societies

(like

palaces,

elaborate

the western Pueblo). This discussion, as well

as my subsequent comments on southwestern archaeology, are prefaced with the recognition that although Puebloan settlements share a generically similar architectural plan, specific aspects of these settlement layouts, as well as Puebloan organization in general, encompass considerable diversity. During the last decades the egalitarian concept has undergone considerable reevaluation (e.g., Feinman 1995; Flanagan 1989), with most scholars recognizing that few, if

burials,

40

MODELS

OF

POLICITAL

leaders of ceremonial societies and then worked for the leaders by society members (Forde 1931; Parsons 1925:87). By implication, access to the means of production (e.g., land, irrigation technology) was affected by group (sodality) affiliations and not evenly

any, human societies are entirely without so-

cioeconomic differences based on age and sex. Yet if the expanded notion of an “egalitarlan society” is to retain some comparative meaning in accord with Fried’s (1967) original usage, then such social systems ought minimally to have (1) basically consensual (as opposed to coercive) decision-making structures along with a high degree of household autonomy, (2) no inherited differences in access to power or the means of production,

distributed at birth (see also Upham 94-100).

and (3) the absence of institutionalized posi-

article

Elizabeth

Brandt

(1994:

12-16) argues that the historic Pueblos had leaders and status hierarchies. Positions of leadership could be relatively permanent and were associated with prestige. Although there generally was not a single leader, councils of ceremonial authorities had varying degrees of coercive power. These religious leaders controlled esoteric knowledge and ceremonial property, including sacred objects, songs, and chants, as well as, in some cases, the calendar, specific information about resource distributions, and aspects of irrigation technology (key to agrarian production). Leaders, through their role as clan authorities, could oversee storehouses of food or other goods that were redistributed in feasts. According to Edward

Dozier

(1970:154),

(see Ladd 1979:482; for Hopi see Waters 1963:130-131). Certain kin segments were

more highly ranked, and this was reflected in their larger quotient of leaders. In addition the social role or office of a council member existed apart from a specific individual. Positions did not depend on the charisma of a specific individual and their ability to build a faction. Rather, leadership roles were to a degree institutionalized and associated in large part with specific clan affiliation (Watts 1997). Fewer leadership positions existed than there were capable people to fill them. These aspects of leadership are classically nonegalitarian (Service 1971), and this is not a picture of a truly consensual society (see Cordell 1989). Yet in other ways historic Puebloan organization did not parallel Service’s traditional, and narrow, definition of chiefdoms. Power was not centralized in the hands of a single ruler (Brandt 1994:9-10), nor was trade or production monopolized by a single individual. In life and death material wealth generally was not hoarded or monopolized, although sacred knowledge was so controlled

the

discipline of religious-political authorities could even be “despotic,” with individuals losing property, houses, and/or land and potentially facing eviction from the community as a consequence. Many means of corporal punishment could be employed (Upham 1989:96). Clans also controlled land, which was inequitably distributed across these sodalities (e.g., Levy 1994:237; Whiteley 1985: 369-370). Among the historic Hopi the more important the ceremonial responsibilities of a clan, the better the quality of land held by that group

1989:

In a specific review of historic Zuni, Todd Howell (1994:64) adds that the social role of a council member was an institutionalized office that existed apart from individual personalities. Specific leadership positions were determined by clan membership, and the representation of clan segments in positions of authority was heavily skewed toward those clans of highest rank. Thus structurally, modern Puebloan leadership differs from the traditional notion of the “egalitarian society.” Leadership positions were ascribed, although not necessarily in a lineal (parent-to-child) manner. Nevertheless, members of certain clan or kin segments were much more apt to be leaders than were members of other clans

tions of leadership. In egalitarian societies those intrapopulational inequalities that do exist should strictly reflect age, sex, personality, and experience. The modern Pueblos do not conform to these egalitarian criteria. In an important overview

ACTION

(Levy 1992; Titiev 1944:66). In

the past Hopi land was ceded to important

41

GARY

M.

FEINMAN

(Howell 1994:89). Modern Pueblo society is

Pueblo social structure. To enable a commu-

dizing leaders, and differential access to basic

segments must cooperate for the good of the whole. This cooperation is fostered through

characterized by leadership without aggran-

nity to function most effectively, all social

resources, technology, and information without ostentatious wealth accumulation. Based

the mediation of clan leaders, with each hav-

on the rich ethnographic literature, it is important to consider the ways in which these

ing a prescribed role. For a ceremony to be conducted properly, and for the ritual prayers to work, each corporate segment must fulfill its ritual role (Tyler 1964:156, 169). For the Hopi the entire course of creation and history

societies were organized and integrated, fo-

cusing on the key roles of ceremony, ritual, architecture, and corporate codes. Pueblo officials were always ceremonial

is unfolded annually in the ritual cycle (Waters 1963:125). At Zuni, familial relations

leaders with mythological sanctions for their positions (Titiev 1944:59). Societal order was

serve as a metaphor for interclan relations at

partly maintained by calendric katsina ritu-

the level of the community, with the highest

als, which were organized by certain ceremo-

ranking religious office referred to literally as

Outside of specific performances, masked katsina impersonators could implement discipline and punishment or organize work parties for subsistence-related tasks (Adams

los correspond in key ways to the corporately organized hierarchical societies discussed earlier (Table 3.2). We have noted (1) hierarchical offices that were not tied to great dis-

nial authorities (Titiev 1944:66) whose positions were hereditary (Waters 1963:167).

the “House Chief” (Watts 1997:23-24). In many senses the historic/modern Pueb-

1991:7). Katsina dances also were associated with food redistribution, ceremonial feast-

parities of wealth,

(2) shared power

in the

hands of ritual/political councils (e.g., Ladd

ing, rain, and fertility (Adams 1991:7). Important rituals were conducted at the heart of the pueblo in centrally located kivas and plazas. Historically, the kivas (elaborate,

1979:488),

(3) large public ceremonies,

(4)

the maintenance of order in part through rituals of fertility and the control of sacred knowledge, (5) strong corporate codes and

semisubterranean ritual structures) with the

ideological order (see also Ortiz

1962),

(6)

sipapus (small holes in the floor) were viewed as passageways to the Underworld (e.g., Mindeleff 1989; Shafer 1995; Waters 1963), bridges across the multitiered universe to the supernatural. With the beneficent influence of the supernatural, masked katsina dancers often issued from the kiva at the core of the pueblo, where their performances were watched by the rest of the community (Adams 1991:12; Mindeleff 1989:117). Traditionally, the performers were sprinkled

sive food production rather than wealth distribution or a marked emphasis on prestigegoods exchange. Access to power and the means of production were to a large degree

ritual performances were central to community survival and well-being, reinforcing societal values and positioning (as well as maintaining) the community amid the larger universe (Plog and Solometo 1997:177).

sequently, it is perhaps not surprising that Fred Eggan, whose early work (Eggan 1950) is often cited by archaeologists (e.g., Longacre et al. 1982; Vivian 1970) as evidence for the nonstratified nature of the ethnographic

to a lesser degree Acoma society is paralleled in the world of the katsinas (Adams 1991:13; see also Swentzell 1988). It is the fabric of

death that “no modern ethnologist or social anthropologist believes that the modern Pueblo are egalitarian” (Eggan 1991:107).

ritual events that both recognized and inte-

grated

societal segments,

(7) faceless

(even

masked) rule and discipline (with the mask reinforcing the corporate code and esoteric knowledge to wear the mask [Adams 1991: 12]), (8) communal organization of labor, and (9) an emphasis on staple finance/inten-

with cornmeal, which is the basis of life. The

ascriptive (based on sodality affiliation). Con-

The political structure of Hopi, Zuni, and

pueblos, wrote only a few years before his

42

MODELS

OF

POLICITAL

ACTION

Table 3.3. Features of Large Late Precontact Puebloan Societies Egalitarian Formations

Conforming Features

Apparent Anomalies

Rarity of princely burials Similar residential settings

Amassed labor efforts

Monumental public construction

Damped household access differentials

Lack of identifiable “rulers” Intrasettlement social segmentation

Multimodal settlement pattern Marked disparities in storage Restricted access to ceremonial features

Specialized production and labor divisions

Select differences in access/burial Large aggregated populations Chiefly Polities

Monumental public construction Amassed labor efforts Multimodal settlement pattern Marked disparities in storage Restricted

access

to ceremonial

features

Specialized production and labor divisions

Rarity of princely burials Similar residential settings Damped household access differentials Lack of identifiable “rulers” Intrasettlement social segmentation

Select differences in access/burial Large aggregated populations

THE LATE PRECONTACT

remarked: “Traditionally, the prehistoric Southwest has been considered an area that supported only egalitarian societies.” Elite burials seem to be rare, hoards of prestige wealth are even rarer, and ostentatious personal accumulation (as in the historic Pueblos) is for the most part not evident. There are few “elite residences” that we all can

PUEBLOS

The above review of more recent Puebloan organization offers potential insights for reconsidering our models of the late precontact American Southwest. Although I do not suggest that the corporate-hierarchy perspective on the historic Pueblos can be extrapolated simply or directly to the past, it does provide some new avenues in which to rethink old dialogues (e.g., Cordell 1989; Lightfoot and Upham 1989; Nelson 1994; Reid and Whittlesey 1990). The debate over whether ancient Puebloan social organizations were egalitarian has been clouded by the assump-

agree on. Although wealth items and labor-

intensive goods clearly are not evenly distributed in the Southwest, they do not show indications of having been restricted narrowly to a privileged few. Concerning highly decorated pottery, “exchange may not have been as widespread and distributions might not have been as uneven” (Hegmon and Plog 1996:32) as previously supposed. For the prehistoric Puebloan Southwest, the control of exchange networks and prestige goods generally does not seem to have been at the nexus of power. Nor have we found much indication for elite managers personally controlling or administering Puebloan economic production. But does this imply that prehistoric Puebloan organization was always nonhierarchical and egalitarian? To paraphrase Lynne Sebastian (1991:115), there is better evidence for leadership than specific leaders

tion that hierarchical, nonegalitarian social

systems would necessarily be correlated with agerandizing individuals, a strict centralization of power, and the personalized control of wealth. This presumed isomorphism has affected advocates of both sides of this debate, influencing their theoretical expectations, as

well as their interpretation of the archaeological record. After decades of rather contentious debate over precontact Puebloan social organization, we see that some of the proposed criteria for hierarchical organization have withstood the challenge better than others (Table 3.3). As Cordell and Gumerman (1989:5) have

43

GARY

M.

FEINMAN

Comparative distributions of storage space are significant. As illustrated more than a decade ago for the Little Colorado region, small sites (whether pithouse or pueblo) have a minimal amount of storage space relative to habitation space (Lightfoot 1984:96). Interestingly, for small sites this ratio remained remarkably constant through time. Over the sequence, larger pueblos always have almost an order of magnitude more storage space relative to habitation space than smaller pueblos. A comparison of larger sites revealed that this ratio increased through time with larger, more aggregated settlements. In other words, larger sites did not simply have more storage capacity than smaller settlements; they had greatly expanded storage space per capita.

(a position in line with that suggested for the historic Pueblos). From my vantage many ancient Pueblos conform to the correlates of a corporate hierarchy. Egalitarian societies are rarely associated with multilevel settlement hierarchies or communities larger than 2,500 people

(Feinman

1995:259-261;

Kosse

1996). Yet communities of this size were present in various areas of the Plateau Southwest after A.D. 1100 (Adler 1996a:106). These larger communities generally formed the upper tier of multilevel settlement patterns in which smaller sites often clustered around larger ones with uninhabited open areas in between (Adler, ed. 1996). Furthermore, as George Gumerman (1972) recognized more than 25 years ago, the large sites were functionally and architec-

Subsequent analyses (Hantman 1989; McGuire and Saitta 1996; Saitta 1991) have supported and amplified Lightfoot’s original

turally distinct from outlying villages (see also Adams 1996; Dean 1996; Kane 1989;

findings. Here it is worth remembering that there is no compulsion for surplus output in

Kintigh 1996; Roney 1996). These focal settlements included ceremonial spaces, some of which were restricted in access (Crown and Kohler 1994), as well as food and water storage facilities that often were absent or represented in proportionally lower frequencies at smaller settlements. The construction of different kinds of sacred spaces and ritual facilities at larger settlements suggests that social roles and activities at these big sites were distinctive. Ritual activities either played a greater or more diversified role at the larger sites, and/or these sites were more apt to service and integrate surrounding populations. Consequently, it seems insupportable to imply (e.g., Reid and Whittlesey 1990; Whittlesey 1986) that large nucleated pueblos were simply several small pueblo units writ large. Proponents of nonhierarchical organization in the Puebloan Southwest have never adequately explained the often skewed distribution of ceremonial and storage facilities at larger sites. Larger pueblos tend to have more

the domestic mode of production, but rather

the political life provides the stimulus to produce (Sahlins 1972:91, 123; see also Arnold this volume). Furthermore, once specific pueblos were occupied for a number of years, there is no apparent subsistence rationale why per capita storage needs would be markedly different at large versus small sites. Likewise, I see no purely ecological basis why people at larger pueblos should have been able to produce or store much more food per capita. Based on a preliminary but stimulating quantitative analysis by Carla Van West (1996), it now appears that Plateau environments potentially were more productive and pueblo inhabitants generally were further from population pressure (even in dry years) than we previously have been led to believe (e.g., Johnson 1989). They also could have produced far more surplus than previous interpretations have postulated. Likewise, the late prehistoric trend toward aggre-

elaborate, and frequently more restricted, rit-

gation cannot be uniformly explained by simple demographic increase or demographic pressure (Crown et al. 1996). Although more such detailed studies from additional regions are needed, these findings are important because, through the refutation of alternative

ual facilities than smaller sites. Restricted access is generally more apparent at larger pueblos (Adler, ed. 1996; Crown and Kohler 1994), and specific ceremonial features (e.g., great kivas, great formal plazas) are absent in smaller communities.

44

MODELS

OF

POLICITAL

ACTION

hypotheses, they tend to strengthen the link between political-ceremonial relations, sur-

Fowler and Stein 1992; Fritz 1978; Stein and Lekson 1992) who have proposed that sig-

dicative of intensive food production have

the ritual landscape as a whole. At minimum,

nificant cosmological meaning is conveyed in the structure of the large ancient pueblos and

plus production, and agricultural intensification (McGuire 1984). Agrarian features in-

large multistory pueblos replicated the quadripartite and multitiered structure of the

long been correlated with larger Puebloan

sites (Plog 1974). How storage is distributed and accessed also is interesting. In pithouse communities, early pueblos, and smaller late precontact pueblos, storage space generally is directly associated with (or accessed from) particular habitations. In contrast, in larger pueblos storage space is often associated with kivas, other ceremonial facilities, or in a more general manner with room blocks (Dean 1996; Plog 1995; Saitta 1991). In the latter cases it is difficult to tie specific storage areas with certain families or individuals. Although this may seem to contradict the expectations for individualized

(network)

hierarchies

universe and linked individual households into community “big houses.” Shared architectural designs and ceremonial features undoubtedly communicated commonly held values possibly manifest in related rituals. From this perspective corporate Pueblo power was rooted in the creation of (and restricted access to) ceremonial space and so ul-

timately in ritual performance and the control of sacred knowledge, as well as food production and consequently land. Because these basic factors previously were argued to be those at the heart of hierarchical, nonconsensual relations in the historic Pueblos, their apparent importance for the ancient Pueblos should not be surprising.

where

storage is centralized in the hands of a single ruler, it fits the expectations for a more corporate-based organization in which leadership is tied to specific offices and/or corporate group membership (Table 3.2). In such situations marked differences in access and power are more explicitly manifest between different sodalities or large coresidential segments rather than between a small subset of elite households and the remainder of the population. This general pattern also would

COMPARISON

TO OTHER

MODELS

It is perhaps useful to compare this corporate hierarchical model of the later precontact Pueblos with other current and related views that also have moved beyond the monolithic egalitarian/nonegalitarian debate. Gregory Johnson’s distinction between simultaneous and sequential hierarchies represents an important expansion of Fried’s and Service’s original conceptualizations. Whereas the former (simultaneous hierarchies) conformed to the traditional idea of hierarchies, the latter (sequential hierarchies) were defined as an “egalitarian alternative.” More specifically, Johnson argued that sequential hierarchies provided a means to organize larger groups of people (often only for seasonal or temporary durations) without any marked intrapopulation differences in wealth or power. Positions of leadership were seen as shared or only event specific. Such social formations were observed to often achieve a degree of coordination and integration through a heavy emphasis on ritual activities and the ceremonial linkage of modular segments into larger wholes (Johnson 1989).

conform with a model of restricted, but not

universally open, sharing across a commu-

nity (Hegmon 1989).

The formal construction of large planned ceremonial/residential spaces, as well as extensive road systems, indicates the likely presence of mobilized labor and_ hierarchical forms of leadership. If we look at larger pueblo structures holistically, they required a greater labor outlay than generally recognized for egalitarian societies (cf. Adler 1989). As in the modern Pueblos, some researchers (Johnson 1989; Lowell 1996; Saitta

1991) have noted component, possibly corporate, social segments, which through the built environment were integrated into larger social wholes. I also concur with a growing number of scholars (e.g., Cameron 1996;

45

GARY

M.

FEINMAN

In a later paper Johnson (1989) directly applied the notion of sequential hierarchies to the Puebloan Southwest. His ideas found considerable and immediate resonance among researchers (e.g., Reid 1989; Spielmann 1994) who generally view the past and

Yet Saitta’s definition of communal formations also includes the notion that all societal members participate in production, that they have access to the necessary factors of production, and that the entire population determines the division between essential and

present Pueblos as egalitarian. In contrast to earlier theoretical views, which tended to

surplus labor. I believe that kind of consen-

sual decision-making structure is a bit too utopian. The model may be more appropriate for smaller and more mobile settings. J am less sanguine whether any sedentary population with hundreds or even thousands of

equate egalitarianism with relatively small

populations, Johnson’s sequential-hierarchy model provided an attractive construct in which larger groups and settlements theoretically could be integrated without marked disparities in political position or wealth. Although I concur with Johnson’s (1989) sequential-hierarchy model in its emphasis on ritual and the ideological integration of

members can make and implement decisions

to divide land and surplus, organize defensive strategies, determine access to ritual knowledge, manage issues of succession, and deal with other matters in such a consensual or completely open manner (see Johnson 1978, 1982). Such a communal structure would be

component Puebloan segments, I take issue with his fundamental description of the an-

cient Pueblos as egalitarian. Power sharing, the absence of linear descent, and even elements of an egalitarian ideology do not necessarily signify an egalitarian society as out-

especially problematic when military compe-

tition and threats from external populations

were often a serious concern, as some scholars (e.g., Haas and Creamer 1993; LeBlanc 1997; Wilcox and Haas 1994) have alleged.

lined above. Like the modern Pueblos, some

ancient Pueblos had, I suspect, positions of leadership. Those people who filled the positions were more apt to come from some groups than from others, and they probably participated in making decisions that affected other people’s lives in fundamental ways. I also see much in concert between the corporate model and the interpretive perspective adopted by Dean Saitta and his colleagues (McGuire and Saitta 1996; Saitta 1991, 1994, 1997; Saitta and Keene 1990), which views the Pueblos as an example of a communal social formation. Like corporate hierarchies the conception of communal social formations (Saitta and Keene 1990:205) avoids the presumed isomorphism among hierarchy,

Certainly, power and decision making were not participatory or open to that degree in the more recent historic Pueblos, which had more formal, clan-affiliated, and institutionalized mechanisms for leadership. THE PITHOUSE-TO-PUEBLO

TRANSITION

It is fair to ask what is gained through the proposition that many later precontact pueblos were organized as corporate hierarchies. At the minimum this interpretation helps direct us from the acrimony of the dichotomous egalitarian/nonegalitarian debate by providing a framework that can account for multitiered settlement systems without great disparities in the distribution of wealth, marked architectural and storage differences between large and small sites with few “princely” graves, and large collaborative labor projects

centralization, and stratification that so often

pervades our general theoretical models. The communal model also recognizes that power may be shared. According to Saitta’s model, communal labor mobilizations provided a kin-based mechanism through which labor was amassed and surplus production was produced well beyond immediate household subsistence needs.

without clear elite residences. Yet, more im-

portant, this model both raises potentially new questions and inspires possible new directions for hypothetical answers. For example, it provides a theoretical basis

to decouple (in some situations) institutional-

46

MODELS

OF

POLICITAL

ACTION

documented to have been much more gradual (e.g., Diehl 1996, 1997a; Rocek 1995). The

ized positions of leadership from centraliza-

tion (see discussions in Nelson 1995 and Rautman 1998). Alternatively, we might ponder when and how a relatively corporate organization became established in the northern Southwest. That question led me to a reconsideration of the oft-discussed pithouse-to-pueblo transition that occurred between A.D. 700 and 1100. During these centuries most settlements in the northern Southwest shifted from semisubterranean, usually rounded freestanding dwellings into aboveground, rectilinear pueblo architecture. Yet this transition was not limited to a shift in house form and construction; it generally also included a change in the location of storage features from subterranean cists to

next major increment in maize use apparently did not occur until well after the pithouse-topueblo shift (Minnis 1985). Thomas Rocek (1995) reports that in one region of southcentral New Mexico the ubiquity of maize remains basically stable in a comparison of a late pithouse and an early pueblo settlement. Although detailed botanical data are patchy across the Southwest, the above studies, along with coprolite (Minnis 1989) and isotopic analyses

(e.g., Matson

and Chisholm

1991), have revealed that there was no systematic increase in corn usage at early pueblo settlements, and therefore the correspondence between maize dependence and the pithouse-to-pueblo transition was far from uniform. If dramatic increments in settlement duration and maize dependence were not directly timed with the pithouse-to-pueblo shift, then what factors may have been responsible for it? Some scholars have focused on the techno-

surface rooms, a formalization of ceremonial subterranean architectural features (kivas),

changes in mortuary practices with a greater proportion of on-site (and subfloor) burials, and the adoption of contiguous room blocks (e.g., Kane 1989; McGuire and Schiffer 1983; Rocek 1995; Shafer 1995; Wilshusen 1989). Of course, the timing and specific details of these shifts varied somewhat from region to region. Diverse interpretations for the pithouseto-pueblo transition abound. During the last three decades most researchers have seen subsistence and settlement shifts as the impetus. Basically, the transition has been understood as a response to an increasing reliance on agriculture and greater sedentization (e.g., P. Gilman 1987; Plog 1974). Yet recent finegrained botanical analyses have cast serious doubts on the subsistence-settlement model. The presence of corn in the Southwest long predates the pithouse-to-pueblo transition (Schiffer 1972a; Wills 1988); more important, however, a significant increment in corn usage appears to have occurred with the advent of pottery and decreasing mobility around A.D. 200 (Wills and Huckell 1994). If agriculture and more sedentary lifeways were the key “kickers,” why did the pithouse-topueblo transition not occur then? Subsequent to A.D. 200, changes in maize usage and settlement continuity have been

logical advantages

(less pest infestation)

of

aboveground storage (Shafer 1995) or design efficiencies of surface architecture (McGuire and Schiffer 1983). Yet the suggested labor minimization techniques and temperature resistance of surface versus subterranean vernacular architecture are not uniformly agreed on (P. Gilman 1987). More important, architecture is part of society’s means of nonverbal communication (e.g., Blanton 1994; Deetz 1977:93; Rapoport 1990), and it seems un-

likely that fundamental changes in house construction, dwelling shape, the contiguity of domestic units, public architecture, burial practice, and storage patterns that were all part of the pithouse-to-pueblo transition would be altered simply to achieve functional or cost-benefit advantages in temperature control or dwelling upkeep. Following earlier discussions (Hegmon 1996; McGuire and Schiffer 1983; Shafer 1995; Whalen 1981), I argue that a more fundamental sociocultural transition lies behind the pithouse-to-pueblo shift. The specifics of this change were not entirely uni-

47

GARY

M.

FEINMAN

form over space and time, partly because the organizational “baseline” (the late pithouse villages) was in itself highly variable (e.g.,

that found for the preceding pithouse era (Kane 1989). The shift to aboveground storage often (al-

where they would have been somewhat open to the community; in others, such as the SU

dences that were positioned in front of the connected storerooms (Hegmon 1996). Al-

Wills r991). In some, like Shabik’eshchee, most storage facilities were outside houses,

though not always) entailed a direct connection of these storage facilities to specific resi-

site, subfloor pits were inside domiciles. Of

though the transition from extramural sub-

course, like late pithouse organization, the

terranean pits to such posterior storage com-

where the transition occurred first (McGuire

privatization of storage (e.g., P. Gilman 1987; Plog 1990), it is difficult to see how food

arching socioeconomic understanding of the pithouse-to-pueblo shift remains elusive, few

those households that were part of the same coresidential compound. In times of need

partments has been interpreted as a kind of

early pueblos also were variable (Hegmon 1995:50), especially in the Anasazi area,

could

and Schiffer 1983:296). Although an over-

have

been

privately hoarded

from

each individual in a compound would know

would see these changes as simply reflecting

where food was stored and who had it. A restricted sharing strategy (Hegmon 1995, 1996) may have been in place in which food was shared freely and generally with coresidential and closely associated households and less frequently during feasts with other residential segments. Through food sharing and kiva ritual, nuclear family units were tightly knit to the other dwelling units that shared their room block, and these residential

a transition from egalitarian to nonegalitarlan society. Nevertheless, certain general aspects of the pithouse-to-pueblo transition have been noted. The transition from freestanding pithouses to blocks of contiguous rooms that housed more inclusive (modular) residential

units would seem to be a rather fundamental social change with major implications for the organization of labor (Whalen 1981). That the early pueblos often incorporated ceremonial architectural features directly into the residential complex itself also indicates a high degree of integration within these dwelling units. Ceremonial architecture entailed greater investment in the pueblos than the earlier pithouse villages (e.g., McGuire and Schiffer 1983:295). In certain areas of the Southwest a hierarchy of ritual features appears to have existed (Wilshusen 1989). A subset of the nonresidential subterranean structures (and

segments

were

in turn

interconnected

and

nested with others (Lightfoot 1994). I propose that the initial pithouse-topueblo shift was a transition to a more corporate form of political action with little initial or associated increment in political complexity. In line with corporate formations (Table 3.2) we see the growing importance of multihousehold corporate labor systems, an increasing importance of communal ritual and ceremonialism, the formalization of ritual spaces, and the integration of modular social units, whereas household storage and wealth distributions were relatively even (perhaps even less pronounced than they had been in some earlier pithouse communities [e.g., Lightfoot and Feinman 1982]). Here it is interesting to recall that many traditional Puebloan mythological accounts describe an ancient shift from localized lineages to more diverse (less genealogically linear) clan-based communities that integrated distinct social segments (Chang 1958).

the associated households) was more elaborate and oversized and appears to have served intercommunity functions (e.g., Blinman

1989; Kane 1989; Potter 1997a; Wilshusen 1989). Communal rituals and feasting seem to have fostered both social integration and a degree of social distinction (Kane 1989; Potter 1997a); however, the traditional markers

of socioeconomic differentiation (e.g., mortuary distinctions, wealth concentrations) were muted and comparable in degree with 48

MODELS

the

OF

POLICITAL

the Zuni designation of the key ceremonial authority as the “House Chief” (see discus-

Harry Shafer’s (1995) symbolic analysis of

pithouse-to-pueblo

transition

in

ACTION

the

sion above). In this regard the manifestation

Mimbres area also is relevant. Shafer (1995: 42) argues that the rectangular pueblo house

of this cosmological

representation

in the

Southwest at a time when other aspects of a

(with rooftop entrance by ladder and a rec-

tangular hearth directly below the roof hatch) serves as a representation of the multilayered universe (Otherworld, Middle World, Underworld). The ladder to the rooftop hatchway is suggested to serve as a

more corporate structure were set in place would seem significant.

many native North and Middle Americans to link the universe’s tiers, thereby providing a connection to the supernatural (Shafer 1995). Ina comparable manner Dolores-area

pueblo transition. But after A.D. 1000, with additional aggregation and greater intercommunity competition (e.g., Adler, ed. 1996; Saitta 1997), more hierarchical (yet still largely corporate) formations emerged in certain regions of the Southwest. Interestingly (as noted above), storage facilities generally were even somewhat less directly tied to residential architecture in these later pueblos. Consequently, one might postulate that increments in complexity/hierarchy in the late pueblos were actually correlated with a less concentrated and perhaps more collectivized deployment of food storage. If this interpretation of changing storage practices can be more firmly supported in the future, then we would have to reject the notion that hierarchical development is uniformly or necessarily characterized by the blanket emergence of self-serving from system-serving institutions (cf. Spencer 1993). Obviously, this broad-ranging discussion has raised as many issues as it has answered. Why were corporate formations so enduring in the Puebloan Southwest? What prompted their emergence and spread during the pithouse-to-pueblo transition? Nevertheless, by considering the prehistoric Southwest along two orthogonal comparative dimensions rather than solely through a monolithic lens of hierarchy, one can begin to understand and integrate important aspects of this complex historic tradition and to place the recognized changes in a broader comparative context that does not simply characterize or dichotomize all southwestern societies under the rubric egalitarian (or hierarchical). At the same time, if we agree that many later

metaphor

(Colorado)

for the axis mundi,

believed

For the most part later pueblos preserved

elements of the basic corporate organizational mode and the architectural cosmogram

by

kivas (associated with the early

pueblos in this more northerly region) are argued to have included sipapus, portals to the Underworld (Wilshusen 1989). Most dwelling units, as well as larger communities, were associated with a kiva. In the early pueblos the basic multilayered plan of the universe therefore appears to have been replicated in the vernacular and ceremonial architecture of the community as a whole. This interpretation would seem to represent a classic example of what Blanton (1994:8-13) has defined as canonical nonverbal

communication,

where

basic

cultural

principles held in common by a community are conveyed through the built environment. This more collective representation of the axis mundi provides a stark contrast to the hierarchical and ruler-centric (networkoriented) Classic Maya (A.D. 250-900), who

portrayed their ruler himself as the axis mundi (Gillespie 1993). The point here is not to raise any kind of diffusion as an issue; I suspect that the notion of the multitiered universe is widespread and temporally deep among native North and Middle Americans. Rather, it is to suggest that part of the Puebloan corporate code was one in which the community as a whole (or a representation of its totality) served as a collective representation of the universe. This has clear parallels to modern Pueblo practice, where each corporate segment is required to complete the ceremonial round in appropriate fashion and

49

that emerged

during the pithouse-to-

GARY

M.

FEINMAN

Puebloan populations were corporately organized, that will help us understand why and how they responded to the later challenges and perturbations in some of the ways they did. CONCLUDING

where (see, for example, the variability doc-

umented in Feinman and Neitzel 1984, as well as in many works cited above). Likewise, the nature, periodicity, and magnitude of the benefits that accrue from leadership (and the relations between leaders and other members of a society) may be very different in societies organized in corporate as opposed to network hierarchies. To construct a better and more convincing theory of long-term social behavior and change, we must recognize and build on this structured consideration of variability rather than compress it, ignore it, or treat it as if it all were beyond productive

THOUGHTS

The corporate/network (exclusionary) dimension introduced here is offered as a means to enhance and expand broadly evolutionary or processual comparative approaches rather than as a complete replacement for that school of thought. For comparative analyses hierarchical complexity remains a key analytical axis. Yet the cor-

comparative investigation.

porate/network dimension provides a means to account more generally for specific varia-

If we can agree that this variability is meaningful and relevant, then the archaeo-

ity. As we have

must

tion within levels of sociopolitical complexseen from

logical correlates

the discussion

be

of hierarchical

broadened

and

societies

reconsidered.

In

above, corporate formations have tended to create interpretive enigmas for scholars re-

viewing the archaeological record our expectations for hierarchical organizations must

ity alone. The consideration of the orthogo-

eties. That is, those archaeological indicators

liant for explanation on the axis of complex-

mirror the recognized diversity of such soci-

nal corporate/network dimension serves to broaden and strengthen our analytic arsenal

(e.g., princely burials, palaces, massive disparities in wealth, depictions of individual

for our study of societal variation and longterm change. A more multifaceted and multidimensional framework for the examination of long-term change eventually should allow us to incorporate aspects of cultural variabil-

grandizing strategies should not be considered the sole or ever-present correlates of decision-making complexity. Instead, we also must devise and search for the archaeological correlates of more corporate hierarchical formations (e.g., massive and coordinated construction projects, administrative offices, insignias of rank, monumental ritual spaces, central and uneven food storage, shared corporate codes). Although this recognition may well make the analytical duties of the archaeologist more difficult, it also may enable us to understand a series of anomalous cases and guide us away from certain overly simplistic and mechanistic readings of the archaeological record. More specifically, the proposed corporate/network model as applied to the Southwest provides a productive pathway away from the acrimonious debate that has permeated recent discussions of ancient Puebloan sociopolitical organization. The model accounts for empirical information about set-

and to build a better theoretical foundation

ity into

our

comparative

perspectives

rulers) associated with more network or ag-

in a

way often unconsidered by monolithic and unilinear frameworks, without having to completely jettison comparison by treating each and every historical transition and sequence as entirely unique. Consideration of the corporate/network continuum also forces archaeologists to focus on power and the different ways that it is achieved and maintained. The perspective provides a framework for the comparative examination of political action from the vantage of both leaders and followers. By doing so it can guide us from current perspectives that blanket our view of past societies with the familiar entrepreneurial and individualistic excesses of the late twentieth century. Certainly, such aggrandizer models work in some

cases, but they do not appear

to fit every-

tlement 5O

patterns,

architectural

differences,

MODELS

OF

POLICITAL

ACTION

ritual structures, and storage features that have never been fully explained by the egalltarian viewpoint, and it also reconciles why

work or individualizing mode to more corpo-

trappings

tionary shift in organizational complexity (e.g., Service 1971) or a dramatic change in subsistence or sedentism. At the same time, this diachronic analysis may help us under-

rate formations. By doing so I have provided a basis for understanding a major and wide-

princely tombs, highly elaborate elite residences, and hoarded prestige wealth—the

of aggrandizing

rare among

spread cultural transition that does not appear to be a consequence of either an evolu-

chiefdoms—are

the prehistoric Pueblos.

The

framework presented here also links the an-

cient Southwest to recent perspectives on the modern Pueblos. Control of ritual, sacred

stand some of the factors and conditions that

knowledge, and staple finance seem to underpin power and hierarchy in each. This inter-

prompted the emergence of corporate modes of sociopolitical behavior more generally. If

west to a fertile body of theory that has enabled us to understand other past societies where hierarchy has not entailed the centralization of power, individual aggrandizement, or blatant stratification.

forces and cross-cultural factors that may underlie organizational variation (along dimen-

we can gain an understanding of the selective

pretation also connects the ancient South-

sions in addition to that of complexity), then

we will have made an important contribution to the studies of long-term change and social theory.

Finally, I have interpreted the pithouse-to-

pueblo transition as a shift from a more net-

51

4

Abandonment Conceptualization, Representation, and Social Change MARGARET

C. NELSON

Abandonments offer a context for examining social changes at varied scales. Within intermediate societies the movements entailed by abandonment are especially interesting because they may be as much a strategy for using landscapes as the failure of a particular structure or adaptation. I discuss the unique aspects of abandonment research that

unstated assumption that people move in lock-step within households, villages, larger

communities, and regional groups. Change is conceptualized as uniform within households, communities, and sometimes within

regions. Second, we must become much more

aware of the impact of our statements on

nonarchaeologists. We have communicated

deepen our understanding of social processes in middle-range societies. I examine the place

of abandonment

research

to the public and to professional nonarchaeologists contrasting images of the “science of

in the study of

archaeology” finding “the answers” or of the “relativist archaeologist” embracing all views, suggesting that any perspective is as reasonable as any other. In addition, we have presented the archaeologist as a detective in a mystery novel. The past is a mystery rather than a history. These images are detrimental to our contributions in contemporary society and social research. I discuss abandonment research as a way of looking at these conditions and moving toward a stronger basis for social theories and a clearer presentation to nonarchaeologists of the sense of archaeology and the past.

these societies, changes in the way abandonment research has been conceptualized and

the impact of this conceptualization on research, and the future directions of research

that will contribute to clarifying the variability in human social behavior within interme-

diate societies and the continuity between

past and present. My perspective is strongly

influenced by a focus on the archaeology of

the North American Southwest.

Throughout this essay I emphasize two changes crucial to all archaeological research and that can be addressed easily in abandonment studies. They have an impact on the quality of our social theories, the contribution of anthropology within contemporary societies, and the future of archaeological research. First, we must take a less normative approach to understanding behavior. Archaeologists for many decades have called for attention to variability in human behavior, yet the majority of our statements about the past and about human behavior rest on the

THE

CONCEPT OF ABANDONMENT THE STUDY OF MIDDLE-RANGE SOCIETIES

IN

Within foraging societies when groups move away from a locale, we say they are employing a mobility strategy. When centers in state-level societies are depopulated, we study collapse (although Sinopoli [1994] explores the mobility of capitals within the his52

ABANDONMENT

of hunter-gatherer with this end of the scale

tory of the Mughal empire). Mobility is a primary focus of research on small-scale social groups (e.g., Binford 1980; Kelly 1983). Collapse is equally pervasive as a topic of study

and subsistence farming with the middle

range. But communities within middle-range societies generally move less frequently, dissolve less easily, and aggregate for longer than do those considered “egalitarian.” On the

by those interested in cities and states (Yoffee and Cowgill 1988). But among people who

other end of the size and complexity scale,

live in social groups more complex than the former and less than the latter, we have varied ways for labeling and therefore understanding movements. We use the label abandonment if we are interested in “the leaving,” such as the depopulation of regions or discontinued occupation of a site or structure, and apply the label migration research if our focus is on “the resettlement.” Why we conceptualize the processes of movement differently in societies of different scales is an interesting question that would provide a more self-conscious understanding of our explanatory stances in archaeology, but it is not the focus of this paper. Rather, I take for granted the use of the term abandonment and its complement, migration, in the study of movement in middle-range societies and examine them as processes, strategies, outcomes, and causes of social change. Understanding the processes of change contributes to better understandings of social life, perhaps applicable to various scales of organization.

middle-range societies are less complex than

states and nations, although they may participate in the actions of or form structures and take actions to resist states and nations (Ben-

der 1990). States and nations have a kind of institutionalized power that is broad, deep, and long-lived. Power relations in middlerange societies may be more often negotiated and more easily undermined than in more complex societies (e.g., Bender 1990; Earle 1977; Potter 1997b; Whiteley 1988). Dissolving social relations and moving among social categories is difficult in state-level societies. Between states and bands, middlerange societies may be characterized as having relatively firm residential focus on places and as subsisting on concentrated resources. They comprise primarily farmers who hunt and gather and live predominantly in aggregated settlements with limited institutionalized leadership (short-term and/or of limited authority). All of these conditions are highly variable, however, and I list them not as a way to classify communities in the past but as a way to delimit the context I am addressing. My working definition subsumes the categories that in the past have been referred to as tribal and chiefdom organiza-

The category “middle-range” or “interme-

diate-level” society is difficult to delimit but not for lack of trying (Feinman and Neitzel 1984; Fried 1967; Service 1962). The problem resides not only in the observation that the category is extremely variable, as docu-

tions. Feinman and Neitzel (1984) use the phrase “pre-state sedentary societies.”

mented by Feinman and Neitzel (1984), but

in the usual problem of creating nominal categories in a continuously varying universe. The term intermediate implies a scale more complex than small, highly mobile groups

I take time to describe my understanding of the category “intermediate societies” because several aspects of this category are particularly important to abandonment research, differentiating it from similar research on a smaller scale and on more complex social contexts. These aspects are mobility, aggregation, place-focused residence, and the nature of leadership. As Rocek (1996) has pointed out, mobility and sedentism are not opposite ends of a continuum; they can be incorporated within the land-use strategies of any group. Most of our ideas about mobility,

that were once referred to as bands (Service

1962) or egalitarian societies (Fried 1967). Delimiting this end of the scale is problematic because we know that small-scale societies have relations that involve different forms of leadership, the longevity and strength of which varies, creating complicated though not complex social forms (Bender 1990). Fur-

ther, we cannot equate subsistence focus with organizational scale, ruling out the equation

53

MARGARET

C.

NELSON

however, derive from studies of small-scale societies in which population densities are quite low and strategies for resource use are

ering ties or rejecting ideas important to the majority of people (Whiteley 1988). New ideologies may form or be adopted as people

1989; Kelly 1983). In the context of some population packing and where resource use is targeted to specific places (as with cultivation or salmon fishing), mobility is organized dif-

Understanding abandonment processes, then, contributes to our understanding of the complexities of social life within intermediate societies. And it is an important arena for ex-

ergy, resources), and access, all of which are

of behaviors among and within communities. The mobility theories that apply to foraging

extensive rather than intensive (Cashdan 1983, 1984, 1990; Halstead and O’Shea

move and form new communities 1991; Adler 1993; Herr et al. 1996).

ferently. Thus, to move may involve considerations of ownership, investment (time, en-

(Adams

panding our understanding of the variability

socially negotiated in ways that differ from

groups and the collapse theories that explain

those in foraging societies. Farming promotes

state-level abandonments are not appropriate to settled cultivator—hunter-gatherers.

a degree of land ownership (Bender 1990; Kohler 1992; Preucel 1988) resulting from

Middle-range

societies differ not only from

the limited distribution of arable land and investments in preparing land for cultivation. Aggregation, well documented among for-

forager and state-level societies; their structures and practices vary greatly both among and within communities. The strategies ap-

terns among many who depend on tended or concentrated resources (like cultigens or salmon). This is not to suggest that aggregation is a permanent or universal condition of life outside of foraging but that it is more common. Thus, to move away from an aggregated settlement is a complex social process. It involves considerations of access and support, whether the move is to another aggregated settlement or to a more dispersed pattern of regional settlement. I have noted, in my discussion of mobility, the influence of place-focused land use on movement. The kinds of investments and access questions that arise may be complicated to resolve in the context of limited institutionalized leadership. Finally, the nature of leadership influences movement. Many argue that social arrangements are relatively flexible (Adler 1990; Horne 1993; Johnson 1989; O’Shea 1989: Upham 1984) or that people in middlerange societies are relatively autonomous; they vote with their feet (Netting 1993:276), sO movement is a strategy for addressing so-

other. “Every economic system has a diverse set of roles available, and households adopt

agers (e.g., Steward 1933; Thomas 1973), is short-lived in comparison to aggregation pat-

propriate for one community, kin group, household, or individual may not be for an-

different

strategies

to take

advantage

of

them” (Wilk 1991:28). We should expect a varied and complex array of behaviors and social relations. For example, the founding households in communities have different options and roles from the latecomers (see, e.g., Haenn 1999); therefore their options for abandonment of that community will differ. (Herr et al. [1996] and Shafer [1996] have be-

gun to explore these differences in the history of communities.)

Abandonment captures the interest of archaeologists and the public because it is the obvious process influencing the remains we see—archaeological sites are not currently resided in by living people. Why did people leave? How did they leave? When did they leave? In this sense abandonment research is the arena of investigation that focuses on why, how, and when people left places. It is easily transformed for the nonarchaeologist into the pursuit of a mystery. But this perspective is restrictive and detrimental to archaeology. Abandonment is a process of transformation from one way of using the landscape to another, within sites, locales, and regions. In this essay I limit my consider-

cial issues (e.g., Dean et al. 1994; Halstead

and O’Shea 1989). But movement is also limited by commitments to the ideas that successfully hold groups together in middlerange societies. Movement may involve sev-

54

ABANDONMENT

ation to movements that result in the absence

established by each household and the nature

people change by moving residence; in fact, perceiving abandonment as part of an ongoing process of transformation allows us to link the past to the present. People are constantly transforming themselves, their land, and their communities. The study of abandonment behavior contributes to the development of social theory from the perspective of both action and change. Moving and transforming the use of land at any scale is probably a conscious, socially negotiated decision and is certainly more often than not a conscious and variable

Also, movement by any household impacts the abandoned community in relation to the ritual, social, and subsistence obligations carried by that household. It impacts the recipient community similarly in relation to the demands created by its presence. Four perspectives on abandonment behav-

of active residence. It is not mysterious that

of each

ior direct current research.

set of actions (Adler 1993). What conditions

cides? What is the nature of the negotiation?

(Wilk

1991).

First, abandon-

the former limits our understanding. Second,

How does the move impact those who stay and those who leave and those with whom the departing later join? The action of moving is a social one. Abandonment is an aspect of ongoing social change; in the context of change, we gain insights about the social dynamics of groups. In middle-range societies occur more

economy

ment is a process, not an event. It begins well before any actual move and continues to impact people and modify landscape use well after a move. Because the archaeological record is the product of cessation of residential use of places, it has been easy to think of this cessation as an event rather than a process, but

a move? How is it accomplished? Who de-

decisions to move

household

abandonment is studied at different scales. I refer to scale in two senses: the size of the area that becomes unoccupied and the analytic scale at which we examine the nature of movement (segments of settlements, whole settlements, local areas, and broad regions).

Third, residential abandonments occur in several different ways, including temporary,

often at the

scale of individuals and small groups (households?). Within the Southwest there are few examples of movements en masse by large

partial, anticipated, among many other aspects. These vary within and among settlements and regions. Fourth, abandonment transforms the use of places. It is not to be confused with discard or relinquishment of ownership. The latter is a modern, Western view of abandonment that is unfortunately applied to interpreting the past. I examine these aspects of abandonment research as they influence our understanding of middle-

community groups (Haury 1958; Herr et al.

1996; Schlegel 1992; Upham 1984). Also, movement of individuals and households may have been rather frequent, at least in marginal environments, because as Johnson (1989) argues, mobility is an important survival strategy (for the Southwest see Cordell and Plog 1979; Hard and Merrill 1992; Reid et al. 1996; Schlanger 1988; Varien 1999; Zedeno 1994). Thus, the social dynamics of

range societies.

Abandonment is a process (Cameron [1993]; Fish et al. [1994]; Lightfoot [1992: 67]; Schiffer [1972b, 1985]; see Anthony [1990] and Herr at al. [1996] for a similar view on migration). Thus, it should not be studied in the absence of attention to how places were settled and the dynamics of changing settlement organization. Abandonment research simply emphasizes the process of moving away. For example, Reid argues that migrations into the Cibecue Valley in eastern Arizona in the A.D. 1200s decreased mobility, created uncertainty and competi-

negotiating movements occurs constantly in

face-to-face relations among people with different kinds and amounts of information. The dynamics of gender relations within and outside households, of age relations, of wealth and knowledge differences all impact and are impacted by this constant change. For example, the decision by one household to leave a village and settle in another is not the same as that for another household. Options depend on the history of relationships

55

MARGARET

tion over access to resources, and led to the abandonment of some sites and areas and the occupation of others (Fish et al. 1994: 157-161; Reid et al. 1996). The social history

C.

NELSON

settlement. She describes a shift in residential focus at Oraibi toward the oldest ceremonial structures and communal areas at the village. Those who moved from Oraibi to establish a

of mobilizing human resources is as much an influence on movement strategies as are the

new settlement did not continue to construct the same ceremonial structure (Whiteley

1993).

quite different social processes than did the

conditions of the natural environment (Stone

Abandonment processes vary according to scale (Cameron 1991a, 1993; Cameron and Tomka 1993; Cordell 1984). The social processes of regions, sites, and areas within sites differ markedly and contribute to understanding varied social dynamics of change (Fish et al. 1994:136). The causes for depopulation of regions or reorganization of regional settlement involve social relationships not only within villages but within regional networks and are tied to the conditions of the regional landscape. These can more easily be related to broad scale climatic changes than to local degradation effects. The causes of intrasite abandonments may be much more idiosyncratic.

Cameron

has

noted

1988). The shifts within the village involved

resettlement. Abandonment

of ceremonial

place being left to the new home, what can be transported, and the condition in which the abandoned place should be left (Adler 1993; Schiffer 1985). The extent to which these can be planned depends on the causes of the move and the nature of the pull from other places

that

(Anthony 1990; Cameron 1995; Lipe 1995).

Many archaeologists have contributed to an in-depth understanding of the material implications of different degrees of planning, distances of moves, and permanence of move (e.g.,

1975;

Binford

1978;

Brooks

Schiffer 1985; Stevenson 1982; Wilshusen

1986 [these are selected sources from different regions; a full set of references would be lengthy}). This methodological focus on recognizing the behaviors encompassed by different abandonment strategies has contributed to recognition of the variability in behavior and

structures

archaeological record. As many

ceremonial structures is an act of rejection or

have noted

(Cameron 1991b; Joyce and Johannessen 1993), occupants of a single site behave differently and by implication employ different strategies in the process of moving. This realization has produced a more complicated past to explain and one in which models of abandonment and migration can no longer generalize about the residents of a site or re-

“ideological closure” of the ideas and practices that defined and integrated a community. Regional moves may be so disruptive that they entail such action (this is my idea, (1991b,

Baker

1993; Cameron 19914; Joyce and Johannessen 1993; Kent 1992, 1993; Lightfoot 1993;

because these could no longer be guarded against intruders. In contrast, local movements may not be associated with ritual closure at abandoned sites because the ceremonial structures continue to be watched and visited. It is also possible that ritual closure of

not Adler’s). Cameron

ways.

move, how the move will be made from the

gional abandonments require ritual closure burning)

in various

dential and nonresidential moves, rapid and gradual abandonments, and short- and longdistance movement represent some ways that abandonment takes place. A range of considerations is involved in any decision to move residence, including the time frame of the

“abandonment of structures or activity areas is a constant process in many settlements” (1993:5). Two examples illustrate the effects of scale of movement. Adler (1993) examines the differential treatment of ritual structures depending on whether abandonment involves shifting occupation within a site or mieration from a region. He argues that re(through

occurs

Temporary and permanent movements, resi-

1992) docu-

ments the historic reorganization of groups within the Hopi village of Oraibi with departure of some and shifting of others within the 56

ABANDONMENT

gion. Contextualizing the process of move-

perspective

is difficult but an appropriate level of analy-

homeland because it is part of their essence

ment to the level of household and individual

of identity

and

community

in-

tegrity, people maintain contact with their

sis, when combined with broader scale analysis, for understanding the social processes involved in moving. I am not suggesting that we know what people think about their options but that we consider the place of households and individuals in the structure of communities and how that place influences their options, their strategies, and ultimately their behavior. Abandonment transforms the use of places. Schiffer has stated that abandonment is “the process whereby a place—an activity area, structure, or entire settlement—is transformed to the archaeological context (1987: 89). Schiffer’s model of formation processes also recognizes that discontinuation of one kind of use need not imply discontinued use (Schiffer 1972b, 1985, 1987; see also Cameron 1991a). There are many examples in which residential use ceases but other uses of a site or parts of a site or a larger locality do not. Within the Southwest, modern Hopi and Zuni continue to visit places where they once resided in large villages and use them for collecting resources and performing rituals (Ferguson 1995). Adams and Adams (1993) and Reid (Fish et al. 1994:162) argue that residents of previously densely occupied regions in the North American Southwest continued to collect resources after they had abandoned them as loci of residence. The regional networks of those who continued to occupy these depopulated regions shifted toward other centers of growth. Among the Late Woodland residents of Western New York, locales abandoned as residences and primary field areas continued to be used as hunting and collecting areas. In fact, by the historic period, at least, people were extremely protective of these hunting areas. From a strictly functional and ecological perspective, “continued utilization of land would have extended the duration of claims on the best lands” (Adler 1996b:3 55). Adler argues that continuity of access and claim is related to previous investments. From the

and their history. E. Charles Adams (1997) documented the continued use of an abandoned farming village in the Homolovi area in northeastern Arizona for the ritual burial of birds. Many native people state that prehistoric sites are not abandoned; their buried ancestors still occupy those places. Returning to those “living” places to communicate with ancestors and bury relatives may be more common than we currently perceive (see, e.g., Blake 1998). Continued use for different purposes is more pervasive and extensive when moves are not across great distances (Adler 1993). But even at the level of regional abandonment people return to use their former homeland and retain rights to the land (Fish et al. 1994:145; Schwartz 1970). CHANGING THE FACE OF ABANDONMENT STUDIES: CONCEPTUALIZATION AND RESEARCH

QUESTIONS

Explaining abandonment is an obvious focus for archaeological research and is of particular interest to that segment of the public for whom the abundance of once-occupied places

provides

an

intriguing

mystery.

In-

deed, it has always been popular to study abandonment. But the way in which abandonment has been conceptualized has influenced the design of research and the kinds of inferences made about social change. The earliest studies of abandonment addressed the causes (warfare, raiding, environ-

mental degradation, insect infestation, climatic change), with abandonment as the effect (Cordell 1984:304-312). Although the causes of social changes are important, many of these studies interpret abandonment as a passive effect rather than as an active process. As a result, consideration of the outcome of an abandonment was not a central issue (but see Schiffer 1972b, 1985). In addition, abandonment was seen as the failure of people to thrive. To some archaeologists and many interested nonarchaeologists abandonment is 57

MARGARET

C.

NELSON

ing within a region that allows for longterm continuity of land use. Schlanger and

equivalent to quitting. General disinterest in the study of migration beginning in the 1960s has, until recently, limited our understanding of the role of movement in continuity (but see

Wilshusen (1993) examine the local shifts in

residential locations within the Four Corners area of the Southwest that occur within the framework of long-term occupational conti-

Schiffer 1972b and Schwartz 1970).

Studies of cause have become more com-

nuity at a broader regional scale. Within a

plex, considering not only the conditions that

similar region Varien (1999) documents the local shifting of small groups within a broad

contribute to leaving but those that draw

people out of one site, locale, or region to another. Fish et al. (1994:143) state that “aban-

regional community that maintains use of the

most productive places over many centuries. Reid and others (Fish et al. 1994:157-161; Graves et al. 1982; Reid et al. 1996; Zedenio 1994) analyze regional population shifts

donments during late prehistory (in the Southwest) can be viewed as the outcome of interplay between conditions in previous homelands and the availability and percep-

within the mountainous region around the Mogollon rim in Arizona during the 1300s

tion of alternatives” (additions and emphases are mine). For example, they discuss demo-

from small to large settlements as a reorgani-

graphic and climatic changes associated with

zation responsive to changing demographic,

abandonment of some sites and areas in the

social, and environmental conditions. Their

mountain Mogollon area of eastern Arizona during the early 1300s but emphasize the growing perception of threat and uncertainty

work assesses the social dynamics of incorpo-

rating ethnically diverse groups in the large reorganized villages. My own research with Michelle Hegmon in southwestern New Mexico documents the shifting of population from aggregated to dispersed settlement

that led to abandonment of smaller sites and ageregation into large, defensively located villages (Fish et al. 1994:157-161). I will return to the “perception” of options later. Fish

within a region as a strategy that allows for

et al. (1994) and Lipe (1995) discuss pull factor of extremely large, aggregated communi-

continuity in regional occupation (Nelson [1993, 1999]; Nelson and Hegmon [1996]; see also Nelson and Anyon [1996] for a similar study). Outside the Southwest Marian White (1963) documented shifting village location on a decadal cycle that suited the cultivation strategies of Late Woodland occupants of the western New York region. Horne (1993) has coined the terms /ocational stability and occupational instability to contrast continued commitment to the use of places with the actual continuity of occupation. Regional reorganization carries with it various social implications, brought into the spotlight in recent abandonment research. Adams (1991), Reid and his students (Reid et al. 1996), Mills and her students (Herr et al.

ties that drew people from one region to the next during the very late prehistoric periods throughout the Southwest. More recently, the role of movement in occupational continuity and reorganization of the use of regional landscapes is recognized by archaeologists and ethnographers (Fish et al. 1994; Johnson 1989; Nelson 1999; Varien 1999; Zedeno 1994). Abandonment of the residential use of places is seen as a strategy (Cameron 1993, 1995; Horne 1993). Native

people in the Southwest believe that movement has always been part of their way of life (Naranjo 1995), resulting in the residential abandonment of many places. Ferguson (1995) reminds us, however, that residential movement away from a place does not represent discontinuity in the use of the place. People remain attached to places through repeated visitation (Cameron 19914) and burial of their kin, who continue to reside there after physical death. Abandonment is now examined as a strategy for demographic shift-

1996), and Zedefio (1994) have focused on

the problems of integrating groups that have different social and possibly ideological practices. The context of Reid’s work is one in which small sites are abandoned as people shift to large village communities. Mills is examining the reorganization of local settle58

ABANDONMENT

ment

with

increased

immigration,

positing

tional continuity. Recent research documents

the view that new structures for community integration bring diverse groups together (Herr et al. 1996). The appearance of two

several major regional depopulations during the late prehistory of the Southwest (Cameron 1995; Fish et al. 1994). Long-distance

new forms of community architecture, large enclosed plazas and large subterranean com-

movement

munal structures, suggests a change in ideology and the contexts of community social relations. Fish and Fish (1993) find the same developments in communal architecture with reorganization within the Hohokam region in southern and central Arizona during the Classic period. In essence these researchers are concerned with the problem of resolving diverse perspectives that result from the coming together of varied social groups who have abandoned their previous residences. The appearance of large community structures and perhaps a new ideology of community may be examined from the functional perspective of the mechanisms of adaptation, as well as from the perspective of practice. Communal structures may “serve to integrate” diverse groups, but how do individuals come to participate? Regional

reorganization

may

involve

western Pueblo areas. Adams

New

Mexico,

Hegmon

of people in-

(1991) argues

for the emergence of an ideology of ancestor worship as a means of integrating the disparate groups that formed the large communities in the late prehistoric Southwest; this ideological change correlates with an increase in the size and a decrease in the number of communal ritual structures. Most recently, research on abandonment has brought to archaeology a clearer recognition of the variability within human societies. The abandonment of a village is an excellent context for examining social processes that are extremely variable. The options facing any individual or household in an intermediate society are based primarily in their kinship network and other relationships established over time. Thus, the options across all households are extraordinarily diverse. The village, which once appeared fairly integrated and homogeneous, becomes a diverse array of small groups or individuals acting according to their perceived options. This variation, of course, exists within the structures of communities, but such structures may be fairly

a

shift away from large sites toward small settlements. Within the Mimbres region of southwestern

by a large number

volves considerable change. In the Southwest dramatic changes in the organization of communities around large enclosed plazas occurred with movement out of the Four Corners and into the northern Rio Grande and

and I

have examined some of the social processes and changed contexts created by such a shift (Hegmonet al. 1998). Within large villages of the Classic Mimbres period, styles were relatively uniform across a number of media (ceramic and architectural). Once villages were abandoned and some segment of the population moved to farmsteads, styles became less uniform, and many new styles were integrated with existing forms and practices. Although this may be perceived as an adaptive shift that accommodates the formation of new groups or networks, it is also a dramatic change in the structure of the communities within which people interact on a daily basis. In this new context the nature of negotiation, the availability of options, and the limitations on actions would be different from the earlier village setting. Residential abandonment does not always result in reorganization and regional occupa-

flexible and fluid (Bender 1990).

Two examples illustrate this variability. In the Southwest many archaeologists have argued that agricultural land surrounding villages was and still is controlled by clan or kin-group leaders but that more distant fields, marked by field houses, indicate individual ownership or at least “practical control” (Kohler 1992). Kohler suggests, for the northern Southwest, that only a limited number of families may have had the right to claim land and establish field houses. This limited access is interesting in light of the abandonment and reorganization in the Mimbres region in the mid-1 1oos. In the east59

MARGARET

C.

NELSON

ern Mimbres area, during the twelfth century, villagers abandoned their large settlements, some moving to their field houses that were remodeled to accommodate more residentially permanent settlement by a few households (Hegmon et al. 1998; Nelson 1993, 1999; Nelson and Hegmon 1996). If Kohler’s notion is applicable to this context, only some of the Mimbres villagers had this option. What did the others do? Who stayed and who left? This process could entail considerable conflict and negotiation. From an-

[1994:17]). Wilk’s (1991) ethnographic research among the Kekchi Maya in Belize has led him to argue that broad social and economic networks and the general ecological context set limits on the options available to households, but within those limits he documents considerable variation and change. In the context of such diversity the changes brought on by abandonment and movement

farming land held communally to cultivating

processes of social change more clearly than have previous studies that rarely address di-

may

best be understood

as variable within

households and communities. Studies that focus at the household—or even within house-

hold (e.g., Wilk 1991)—and community scale simultaneously may be able to examine the

other perspective, the Mimbres case poses interesting social issues. If people shifted from privately held land, how did the relationships

among individual households change? What constituted the communities and what held them together? These issues are ignored when large site abandonment is viewed as a uni-

verse action.

Finally, abandonment research is an excellent arena in which to examine the nature of negotiation and power in intermediate societies. Whiteley’s (1988) study of the historic fissioning of the Hopi village of Oraibi made clear the role of individuals in positions of authority in orchestrating the abandonment of part of the village. Although we have argued loud and long about the nature of power and authority in intermediate societies (for the Southwest see Reid [1989] and Upham et al. [1989]), we have not addressed how relationships of power are related to movement. Power may derive from the control of resources (Brandt 1994; Hantman 1989), of

form event with implied regional abandon-

ment or simply as a structural failure. A second example is from San Mateo Ixtatan in the Maya Highlands of northwestern Guatemala (Hayden and Cannon 1984), a similarly diverse settlement. Among the approximately 500 households of contemporary subsistence farmers, none depend entirely on cultivating their plots in the village. A few families hunt extensively, some have distant secondary fields, others produce items for local trade or transport them long distances to large cities, and some leave the village on occasion to do wage labor. This varied economic mix supported the continuity of this community, making possible the aggregation of so many people in the agriculturally marginal highlands. I see this case as similar to aggregated villages in the North American Southwest and perhaps many other areas of the world. The diversity of subsistence strategies also implies different practices and sets of daily interactions, as well as the formation of different networks. These form a large part of the options available to people for

people (Saitta 1994; Sebastian 1991; Vivian 1970), or of ritual knowledge (Brandt 1994;

Potter 1997b; Spielmann 1998), each having different implications for the nature of negotiation and access (Bender 1990). In addition, in intermediate societies we see “considerable variability in the ways leadership positions are transferred from one generation to the next” (Feinman and Neitzel 1984:61), sug-

gesting that negotiation is a constant and crit-

ical dimension in community-level social relations. Access to resources and negotiation

of options are derived through social media-

tion (Adler 1996b:339; see also Haenn 1999

moving from one place to another. As Anth-

for an interesting contemporary study of rural farmers in Oaxaca). Similarly, within households access and power can change, impacting the options available to different in-

ony (1990) and Cashdan (1983) have argued, people are most likely to be able to move to familiar places, where they have established relationships with others (also see Zedefio 60

ABANDONMENT

dividuals. For example, heads of households or kin groups, if they are perceived as the holders of land, may have authority to slough

that are the outcome of selecting among options or priorities. Although we may not know fully what options were perceived, we may be able to understand those that were acted on, as well as a range of possible op-

members—i.e., cause them to have to aban-

don their home—under certain conditions (Wilk 1991). The nature of gender and age relationships impact this process in the sense that they are integral to owning and controlling resources and knowledge. Perceptions of abandonment as a process have expanded our ideas about the multiple causes, dimensions, and outcomes of movement. Among the many dimensions two are especially important for improving our ap-

tions. For example, I have discussed varia-

tion in subsistence strategies within communities that are related to access to land and resources and possibly to ownership and authority. Within a community these create different options and relationships of access. I have also noted that there are different kinds of power and that power may be fleeting. It may be worthwhile to consider what options are created for different individuals or groups and how a flexible structure impacts

proach to social change: variability in behav-

ior and negotiation of options. The variable strategies and possibilities for individuals,

and is impacted by negotiations. Similarly, access to information Is variable within com-

households, and communities create a myr-

iad of outcomes rather than the smoother trajectory so commonly presented to describe change. Recognition of the role of negotiation expands our approach to social dynamics from categories of inequality or difference to individual actions with respect to those categories. Our expanded view of outcomes leads to greater appreciation of the role of movement in strategies of land use and therefore attention to the aspects of social relations that promote or facilitate movement. FUTURE

munities, providing different structures of knowledge within which people operate. We may be able to delimit these structures and evaluate the social contexts of abandonment decisions and actions. Behavioral approaches to archaeology have made the greatest contribution to understanding variability in abandonment behavior. As Schiffer states, “The ultimate goal of a behavioral archaeology is to furnish... explanations for variability and change in human behavior” (1995b:251). Although attention has been paid to the technical choices involved in forming and using artifacts (e.g., Schiffer and Skibo 1987) and depositing them (Schiffer 1972b, 1987), the social context of the choices could be more fully explored. Various kinds of abandonment

DIRECTIONS

I have presented some information about the study of prehistoric abandonment in intermediate societies as a way to address efforts at building social theory. A range of approaches can be brought to bear on the new developments in abandonment research. Agency-based,

behavioral,

and

(rapid, gradual, long-term, short-term, distant, near, etc.) may be perceived as processes

processual

approaches all have contributions to make in the future development of social theory addressing movement. Agency-based research is concerned with negotiation and perception of options, among many other aspects of social life. The focus is on structure and practice—structure in the sense of defining the contexts that influence the actions of people, practice in the sense of the actions of people. Shennan (1993) has argued convincingly that archaeological deposits are evidence of practices

resulting from social choices within a given set of conditions and as behaviors that create new conditions, in turn defining a new range of choices. Linking the behaviors that precede abandonment to those that follow and create new conditions would contribute substantially to our understanding of the variable nature of social change. For example, various households within a community may abandon their residences in different ways. Which segments were last to move? Which show evidence of anticipated return? What 61

MARGARET

C.

of archaeological research to nonarchaeologists. Abandonment studies, in particular, are an excellent arena for communicating our perspectives on human society. People are fascinated with abandonment. We can either

conditions may have contributed to the existence of different options for these different households? How might the different abandonment strategies be linked to different patterns of movement (long-distance, perma-

play up the mystery, as we have so often, or

nent, incorporation into existing settlements, creation of new settlements)? The answers to

we can use this fascination as a foothold for having a voice. To play up the mystery is to mislead. It is not a mystery that people left places; actually, it isa mystery that they managed to stay in some places in the Southwest. The topic of abandonment, which draws

these questions can help us better understand

the transformation of human social relation-

ships. Processual archaeologists constitute the majority of those addressing questions about human

movement

in

“intermediate

people’s attention, is an invitation to commu-

soci-

nicate about the complexity of human societies. People have always had to negotiate, strategize, coax, and respond. Conditions have changed, but social life has always been complicated; “patterns of inequality, power

eties.” Recent research has moved the study of the processes of movement toward a more complex understanding of cause that includes push and pull (Anthony 1990; Cameron

NELSON

1995; Lipe 1995). In addition, the

complexity of processes is recognized in the

differentials, and situations of domination and resistance arise in all societies” (Shennan

variable nature of abandonment within sites,

1993:53).1t was not any simpler to be a gatherer than it is to be a suburbanite. The options and dangers are different; the networks and structures are different. But many of the basic concerns are the same. When we act, we influence our conditions in ways that are positive and negative. These positives and negatives develop as we continue to act and live with other actors. How people move through

localities, and regions. The challenge for processualists is to look at the variable nature of human communities for conditions that contribute to that variability. Understanding those conditions may contribute to our abil-

ity to conceptualize the array of social processes involved in the formation, change, and continuity of communities. Our attention in this conference was directed toward the building of social theory. We are also aware that our approaches to presenting our understanding of human behav-

their lives, making choices and mistakes, is at least as interesting as the collapse of any

structure if presented in an accessible way. Archaeologists have considerable knowledge about human society that often is not accessible to nonarchaeologists because we mask it in antiquarian images of the past.

ior, of the past, and of the practice of archae-

ology to nonarchaeologists significantly influence the future of archaeology. I raise this issue here because I believe we can do a much better job of communicating the value

62

5

Elements of a Behavioral Ecological Paradigm

for the Study of Prehistoric Hunter-Gatherers ROBERT

L. KELLY

Archaeologists have used various forms of evolutionary thinking in recent years to account for patterns in the archaeological remains of foraging societies. One of these forms is borrowed from the field of evolu-

at any one time there will be limits. One of the major limitations for archaeologists, and especially archaeologists of hunter-gatherer societies, is that archaeology records patterns in behavior as they are manifested over long

I will discuss several elements of this paradigm that need further thought before archaeology fully implements it, and I will consider the links between behavioral ecology and some of the other approaches discussed in this volume. There are three assumptions to this chapter. First, I take the position that rather than being dogma, the source of off-the-rack answers, a researcher’s paradigm ought to be a way to ensure continuous learning about the past. The paradigm followed here is that of behavioral ecology, and its derived models serve “not as lawful statements about reality but as structured forms of inquiry, more interesting to stalk than to live by” (Winterhalder 1987:313). Second, I take the position that our explanations of the past must revolve around what we can realistically know about the past. We cannot know everything about the past but can only hope to understand patterns in behavior that we are able to infer from archaeological data. This means that the models we use must be compatible with the data at our disposal. New perspectives and techniques will open our eyes to new sources of data, but

chunks of time measured in hundreds or thousands of years. Nonetheless, large-scale changes in human behavior are still the result of change in the way individuals made behavioral decisions in the past. The theories developed to explain the past must take into account the fact that we deal with behavioral decisions that were made by individuals but are tested against assemblages that are the conglomerate results of actions and decisions of many individuals. Third, the approach] take is colored by the fact that I expect archaeology in the twentyfirst century to be quite a different animal from its twentieth-century predecessor. In North America prehistoric archaeology will change as a function of NAGPRA and the ensuing dialogue that it has encouraged (or forced) between archaeologists and Native Americans. Although I think this dialogue is beneficial (Kelly 1998), others lament it because they see it as requiring the destruction of archaeology as a science. This is not a necessary result. Science is a strategy for maintaining honesty in the intellectual enterprise because it makes explicit the conditions under which the researcher will accept that an

spans of time—decades at best and often in

tionary or behavioral ecology. In this essay

63

ROBERT

L.

lationship between a behavior and its reproductive consequences (e.g., see Hawkes 1996; Hill and Kaplan 1988a, 1988b), but behavioral ecology does not argue and need not argue that people must be aware of a behavior’s reproductive consequences or that behavior is genetically linked (including, certainly, those behaviors that are of most traditional interest to anthropology). The majority of the human phenotype consists of culturally rather than genetically transmitted behavior. We receive our genes from only our biological parents, but we can acquire our culture from

idea is wrong. Indeed, for archaeology to achieve any goal (be it drawn from a materialist or some other paradigm) archaeology must become a better science, developing better, less ambiguous ways of interpreting material remains and better, more verifiable ways of explaining human social change. Strongly situated with a scientific approach, behavioral ecology, for me, offers a very good learning strategy for understanding why variation exists among human societies, and it is compatible with other ways of reconstructing, interpreting, and explaining the past

many people (see Boyd and Richerson [1985] and Cavalli-Sforza and Feldman [1981] for

(Kelly 1995). WHAT

IS BEHAVIORAL

KELLY

discussions of different transmission types).

ECOLOGY?

concerned with the differential persistence of

By being an influential role model, a single individual may have no offspring and yet

especially prevalent in, but not limited to, the

ulation.

Behavioral ecology is an evolutionary science

alter phenotypic frequencies in a human pop-

variability in behavior over time. It has been

Unlike another evolutionary approach, evolutionary archaeology or selectionism (Dunnell 1980), behavioral ecology tends not to rely explicitly on the concept of selection

study of foraging and horticultural societies. The field has devoted attention largely to in-

teractions between people and their natural environments,

examining

foraging

group

except that selection is assumed to have been the process that created the human mind.

size, group movements, and especially food selection. Initial answers were found in some

Space prevents me from taking up the debate between these two paradigms (see Boone and Smith 1998; O’Brien and Holland 199§A, this volume; Schiffer 1996; Spencer 1997;

simple economic relationships between humans and their natural environments, but the

search has now expanded to include how foraging choices affect social interactions (and

vice versa), examining sharing, marriage, and

Teltser, ed. 1995), but it is important to ask

life, mating

come more common because its bearers become more common through time. This would entail either (a) that the trait is genetically controlled or (b) that through increased survivorship of offspring and subsequent enculturation by the biological parents, a trait becomes more common because the individuals being taught that trait become more common. The former scenario seems unlikely for most things that interest cultural anthropologists and archaeologists; for example, corner notching or side notching of projectile points

why behavioral ecology does not discuss selection more explicitly. Briefly, practitioners of evolutionary archaeology use selection in two ways. First, there is the claim that if a trait increases the reproductive fitness of

parental investment (for a review of application in hunter-gatherer ethnology, see Kelly 1995). Sharing, for example, is affected by the size of food resources and/or the variability in their encounter rates, but it is also a function of who has power in other areas of choices,

and

attitudes

bearers of that trait, then that trait will be-

toward

children (e.g., Hawkes 1996). How decisions are made and how behaviors are passed on, modified, and/or discarded is a central concern in applying behavioral ecology to the ethnographic and archaeological record. Behavioral ecology does rely on notions of “fitness” because it assumes that humans possess an innate drive to reproduce and to see their children live to reproduce (Cronk 1991; Smith and Winterhalder 1992:27). Some behavioral ecologists focus on the re64

A

BEHAVIORAL

ECOLOGICAL

PARADIGM

and Kelly 1987), is simply uninteresting. Stated bluntly, selectionist archaeology has yet to “solve” any long-standing problems of prehistory. There are two reasons for this. First, the insistence among selectionists that the source variation present in a population is not a subject worthy of study (Boone and Smith

is not a genetic trait. The latter is possible, but I suspect that it would be uncommon because humans are rarely enculturated solely by their biological parents; thus successful traits

can be copied by nonrelated individuals.

Second, and alternatively, proponents of selectionism argue that it is the replicative success of a trait that matters, not the repro-

1998; Schiffer 1996). For Darwinian

ductive success of its bearers (Leonard and

selec-

Jones 1987). Whether one behavior or artifact meets a goal better than another is important, but what determines which behavior or artifact will become more common through time is not necessarily only a function of how well that behavior or artifact meets its direct goal. If there is a change in hunting techniques over time, for example, it

tion to operate, variation is assumed to be undirected, not related to the selective forces at work. But clearly humans can act with intention relative to some goals, invent new technologies to meet perceived needs, and alter their behavior over their life spans (see Boone and Smith 1998). Second, evolution-

with different techniques plays a significant

trait. The use of behavioral ecology’s models,

is highly likely that the return rate associated

ary archaeology has avoided using models that could predict the replicative success of a

role; but other factors, such as the amount of

especially

those

that

model

“opportunity”

costs and examine the trade-offs that someone makes in adopting a new trait, could be helpful here. And as I will elaborate on below, these models could benefit from a greater knowledge of artifacts’ “perfor-

prestige associated with different hunting methods, may also condition which behavior

becomes more common over time (see discus-

sion of the Numic expansion below). This latter approach is the more useful, but so far discussions and examples of this process as selection have been obfuscatory or uninteresting. Schiffer (1996) points out, in

mance

characteristics”

(Schiffer

and

Skibo

1987, 1997). Thus, elements of behavioral ecology and behavioral archaeology could

assist a selectionist archaeology in construct-

critiquing Dunnell’s (1989a) only application

of an evolutionary approach—explaining the Middle to Late Woodland transition—that the approach contains blatant contradictions. Neiman (1999) has also shown that one of the premier examples of selectionist archaeology—the shift from thick- to thinwalled pottery in the Late Woodland of the

ing tests that could determine the extent to which selection shapes the past behavior recorded by archaeology. Behavioral ecology has been successful largely because it has proceeded by constructing and testing models that have a substantial

rather than gradual change expected from selectionist theory. Schiffer provides a better

ing, for example, the amount of calories in a

central

eastern

United

States—is

a

empirical content

(and thus are falsifiable).

These models require baseline data on forag-

rapid

kilogram of ricegrass or the time it takes to find a caribou or to harvest biscuitroot (something akin to behavioral archaeology’s artifact performance characteristics). Ethnographic observations and experiments provide these data (see Kelly 1995:Table 3.3). The models then make predictions—for example about how diverse diet should be (the diet breadth model), which places to search

discussion of artifact selection in explaining

how class and gender, as well as artifact function and cost, account for why the gaspowered automobile was “selected” over the electric version in the early twentieth century (see Schiffer 1995a; Schiffer et al. 1994). Relabeling “adaptation” as selection, as Abbott et al. (1998) seem to do in their explanation of the differences in the stone tool technologies of mobile vs. sedentary societies (Parry

for food

(the patch choice model),

or how

long to forage for food in a given patch (the 65

ROBERT

L.

KELLY

marginal value theorem)—based on the goal of foraging return-rate maximization. These models implicitly entail some very general bi-

things in the most energetically efficient way. First, there may be a time lag between environmental change and behavioral change. The degree to which this is a factor will de-

understanding cultural directives (e.g., bring home as much meat as possible, spend time

vironment is. The more variable it is over time, the less well adapted a particular behav-

duce few children, acquire prestige, acquire

ioral responses to environmental change might be quick or slow, depending on a number of factors, for example the availability of the material or knowledge for a new technology or a new technology’s impact on existing prevalent behaviors (see below).

ological directives (to reproduce and see that their children survive) but are also directed at

pend on how variable and fluctuating an en-

ior might be. In addition, some of the behav-

with offspring, produce many children, prowealth, be good to your relatives). Behavioral

ecological analyses require that goals be defined, but analysis is directed at using the models as ways to ascertain and explore foragers’ goals and why they might differ among individuals and/or among populations. Behavioral ecological studies of huntergatherers have focused on foraging behavior because it is logical that foraging is closely related to survival (Smith 1991:24). The field began with the simple, and simplistic, assumption that the goal of a forager should be to forage optimally, that is, to maximize the net rate of food intake per unit foraging time. Mean foraging efficiency is assumed to be closely linked to fitness although this assumption has not seen any explicit tests (see Hill and Kaplan 1988a, 1988b; Smith 1991). Although the focus on foraging efficiency may seem an undue imposition of Western capitalist values on a noncapitalist social form, Smith (1983:626) points out that under most conceivable conditions we could still expect foragers to maximize their foraging efficiency, and several ethnographic studies have successfully tested this assumption (e.g., Alvard 1993,

1995; Hawkes

Second,

because

cultural

evolution

is a

process that transmits existing biological and cultural information differently from genera-

tion to generation, we do not expect that any

particular behavior will be the most efficient one but that people will select a strategy that achieves goals better than other existing ones. Thus, we expect behavior to only tend toward optimization. Third, the existing frequencies of behaviors in a population may affect a given individual’s choice of which strategy to follow. Early in its application, behavioral ecology assumed that the consequence of a behavior was not affected by its frequency in a population (e.g., the choice of which seeds to gather

or which animal to hunt) but only on characteristics of the resource (e.g., caloric content, time to locate, time to harvest). However, most behaviors have some social impact. Their meaning and value within a population are partially a function of their frequency. Foraging behavior, for example, has some social significance beyond its efficiency, and the

et al. 1982; Hill

and Hawkes 1983; O’Connell and Hawkes 1981, 1984; Smith 1991). Efficient foraging can, however, be undertaken to minimize the time spent foraging—because other activi-

decision to forage for particular foods or by using a particular technique is affected by the

frequency of other strategies in a population. One individual may, for example, choose to

ties, for example trade or child care, are more

critical than getting more than the minimal amount of food needed—or, conversely, to

maximize prestige because many others are

already hunting efficiently. These situations

maximize the amount of food gathered, this

can be analyzed through game theory (see Smith and Winterhalder 1992), which can predict a stable mix of behavioral variants— an evolutionarily stable strategy—that differ in their energetic efficiencies, none of which

in cases where more food contributes to survival more than nonforaging activities (see discussion in Kelly 1995:5 5-56). Nonetheless, there are four reasons to ex-

pect that individuals will fall short of doing

66

A

BEHAVIORAL

ECOLOGICAL

PARADIGM

are as optimal as would be predicted by a

there are still issues that need to be worked

1990, 1992, 1993, 1996).

logical problems.

simple optimal foraging model (see Hawkes

out to apply ethnographic lessons to archaeo-

Last, when foragers select foods and foraging tactics, they must balance a goal of

Modifications for Human

return-rate maximization with a desire to reduce the risk inherent in foraging and with

Foragers

The initial generation of models treated hu-

the fact that while foraging they are forgoing

man foragers as if they were feeding “on the fly,” but human foragers rarely consume food

Figuring out how foragers juggle these goals is the focus of new generations of models

monly central-place foragers (see, e.g., Bird and Bliege-Bird 1997). These models also

Behavioral ecology may have begun with

back, but humans often modify those foods

the

opportunity

to

do

something

in such a fashion. Instead, they are more com-

else.

(e.g., Winterhalder 1983; Winterhalder and Goland 1996; Winterhalder et al. 1999).

have had to be revisited in that animals that are central-place foragers may bring food before doing so—drying meat, for example, or partially processing seeds. How much foragers will process food before returning with it is a complex decision-making process that must take into account factors such as how much weight and/or volume a forager can carry, how much energy must be expended in returning to camp (which is largely but not

some simple models and assumptions, but rather than use them as off-the-rack explana-

tions it has used them as ways to learn and

explore why people behave in the ways they do. These models have proven useful in understanding ethnographic cases. Several applications of behavioral ecological models to archaeological data have verified some of the models’ predictions, notably ones that use the diet breadth model to test the prediction that a reduction in high-ranked resources (through climatic change or population growth and subsequent resource depletion) results in greater dietary diversity (see Broughton 1994a, 1994b, 1997; Janetski 1997), as well as diet breadth and risk reduction models to examine the origins and spread of plant domestication (e.g., Diehl 1997b; Gremillion 1996; Winterhalder and Goland 1996). PROBLEMS

IN APPLYING

BEHAVIORAL

MODELS

entirely a function of distance), and so on,

balanced against the overall return rate from the food collected (see Barlow et al. 1993; Barlow and Metcalfe 1996; Bettinger et al. 1997; Bird and Bliege-Bird 1997; Jones and Madsen 1989; Madsen and Kirkman 1988; Madsen and Schmitt 1998; Metcalfe and Barlow 1992; Rhode 1990). Second, models have had to be modified to take into account the division of labor. It is inappropriate to model human behavior with all foods as potential choices when, in fact,

some of those foods may be effectively out of bounds. If women are not hunting, for example, then there is no use in modeling their foraging choices with large game included (Jochim 1988; Simms 1987; Zeanah 1996). This is significant because women generally

FROM

ECOLOGY

The initial models employed in behavioral ecology were borrowed from zoological studies of animals and insects (zoologists had previously borrowed these from economics). Humans

take seeds, tubers, and small game, resources that tend to have low return rates, whereas

do not behave like other animals,

men tend to focus on large game, resources that have high return rates and that can therefore be taken more profitably at longer distances from camp. This means that efforts to model residential camp locations must sort out women’s from men’s target resources. Both of these factors have been taken into

and it rapidly became apparent that modifications had to be made to accommodate human patterns of foraging. Additionally, most work with behavioral ecology that is cited by archaeologists entails work done with living foragers (see Kelly 1995). Although there are archaeological applications, 67

ROBERT

L.

great (as a function of declines in wetland productivity or an increase in upland productivity), we expect to see greater residential mobility, with the marsh being but one stop on a seasonal round. Linking the model’s predictions to the paleoclimatic record, I generated some expectations about when the wetland should be the focus of settlement. The archaeological record appears to hold up those predictions (Kelly 2000).

account with recent research in the Carson

Desert of western Nevada, an area where an

extensive wetland is bordered by mountains containing pinyon and bighorn sheep, among other resources. Zeanah (1996; Zeanah et al.

1995)

employed

a sophisticated

GIS

KELLY

ap-

proach to determining resource distributions and then predicted where the best returns from women’s vs. men’s foraging were to be found. From this Zeanah was able to predict

probable locations of residential camps (as

small game such as rodents and that men were hunting large game such as bighorn.

Documenting Variation within Populations Because evolutionary approaches presume that change over time is a function of the differential transmission of behavioral variants

there is good reason to expect such a division

variation within as well as between popula-

opposed to daily task sites). In employing such a model Zeanah assumes that women

were primarily gathering plants and taking

over time these approaches must foreground

This is based on ethnographic analogy, but

tions. Archaeologists have long documented variation but still tend to interpret the archae-

of labor in prehistoric cases as well (Kelly

1995).

I have

also employed

optimal

ological record in the emic terms of culture history (Dunnell 1986), focusing on modal

foraging

models in the Carson Sink (Kelly 2000; Larsen and Kelly 1995). Instead of modeling

behaviors, the most common

things that

men’s and women’s diets, however, I devel-

people did in the past. Part of the reason has

ing experimentally derived return rates (e.g., Lindstr6m 1996; Simms 1987) and allowing

gists brought to the study of prehistory, but it

oped a series of expectations of the return rates from various food-getting activities us-

to do with the theoretical premises (notably, a normative concept of culture) that archaeolo-

is also a function of the fact that behavioral variation within a population is difficult to

for how much of a resource could be carried back

in an

average-sized

burden

document archaeologically. Part of the issue is that the relevant unit within which variation should be measured is difficult to define (Pueblo Bonito, Chaco Canyon, the area

basket.

These models end up requiring some precise information—burden-basket size, how much

of a resource can fit into that basket, how

defined by Chacoan outliers, the Southwest,

much does meat reduce in weight by drying,

North America?). This is made more difficult when sites and assemblages can only be assigned to a long time period, which is often the case for hunter-gatherer archaeology. And before one can take variation in artifact form to be significant in terms of the behavior that is of interest, other factors must first be controlled for, such as sampling error, raw

how far can one walk ina day carrying a load, etc. Armed with such data, I generated four models: (1) foraging from a residential camp within the wetland, (2) foraging from a residential camp in the mountains, (3) foraging in the mountains from a camp in the wetland, and (4) foraging in the wetland from a camp in the mountains. To sum up, in years when the wetland in the Carson Desert was productive, the “best” choice (the one that provides the highest net return) is if some people (we presume women) forage in the wetland while others (we presume men) hunt bighorn sheep in the mountains. However, when the difference between the productivity of the wetland and that of the surrounding upland is not

material constraints,

artisan errors/idiosyn-

cracies or artifact life history (Schiffer and Skibo 1997). Dunnell (e.g., 1986) has long appreciated the need to reconcile archaeological theory development with systematics. A focus on documenting variation in behavior requires an approach to artifacts that focuses more on 68

A BEHAVIORAL

ECOLOGICAL

PARADIGM

attribute behavior over time and space than on the construction of types. But it also requires models that recognize that the variation from which some behavior emerges as the dominant form may be forever out of our reach. This makes our endeavor difficult, for it is hard to understand why one choice is made of those available if we do not know what those other choices were. One alternative may be to increase the regional and temporal scales of analysis. We

chaeological and experimental studies, performance characteristics provide critical baseline data in the same way that ethnoarchaeological and experimental studies provide foraging return-rate data for different food resources. We know many of the performance char-

regions, larger even than those associated with settlement studies of the 1970s and

ity,” sharpness, and durability; the fracture mechanics of stone—the relationships be-

havioral archaeologists call artifact “performance characteristics” (e.g., Longacre 1991;

Longacre and Skibo 1994; Schiffer and Skibo 1987;

may need to study adaptations across large

Skibo

1992). Derived from ethnoar-

acteristics of stone, raw material “knapabil-

1980s, ones that permit us to see large spatial

tween striking angle, force, raw material, and

and temporal patterns in behavior. Looking at data drawn from many regions means that

resulting flake shape (Amick and Mauldin 1989; Andrefsky 1998; Knecht 1997; Magne

the model employed has to predict differ-

1985;

Odell

1996; Whittaker

1994). We

ences across space and/or time—and there-

know something about projectile technology,

example in her examination of the temporal

1991), manufacturing breakage rates (e.g., Titmus and Woods 1986), and the relation between stone points and the organic component of projectile technology (Ellis 1997).

fore it would need to be theoretical, not simply descriptive. Beck (1995) provides one

including penetration and breakage rates (e.g., Hutchings and Briichert 1997; Knecht

and spatial distribution of early side- and

corner-notched projectile points in the Great Basin. Relying on the notion of performance

characteristics and experimental data, she ar-

There is also a volume of literature examining

gues that corner-notched points have a lower failure rate than side-notched ones, and

the “organization” of stone-tool technology, “the study of the selection and integration of strategies for making, using, transporting, and discarding tools and the materials needed for their manufacture and maintenance” (Nelson 1991:57; see also Bleed 1986; Carr 1994; Johnson and Morrow 1987; Kelly 1988) and behavioral chains, or chaines operatoires (e.g., Sellet 1993). (I am leaving aside here the equally large literature on use-wear analysis.) This literature has vastly improved our understanding of the basic parameters within which a stone-tool technology operates and has developed some important concepts for its study. To be integrated into a behavioral ecological approach, however, these data and concepts need to be expressed in the same terms as the theory that is part of behavioral ecology, namely, that people weigh their options and opt for those that provide the highest benefit:cost ratio given opportunity costs (see O’Connell [1995] on the importance of this specifically in ethnoarchaeological research).

thus the former “outcompeted” side-notched

points. She notes, however, that there are other factors at work, including delivery technology, function, raw material (e.g., availability, type, nodule size), and manufac-

turing and maintenance costs. And this leads to a second issue in the application of behavioral ecology. A Model of Technology Grounded in Principles of Behavioral Ecology Where behavioral ecology has been implemented in archaeology, it has been with data that more directly reflect subsistence choices, floral and especially faunal data. But the majority of the archaeological record of huntergatherers consists of stone tools and the products of their manufacture and reuse. For behavioral ecology to be implemented we need an approach to stone tools that is grounded in the principles of behavioral ecology. Of special importance here are what be69

ROBERT

L.

KELLY

time it takes to manufacture the tool and on the expected “return” from using the tool. Bipolar technologies entail less investment, but, I suspect, if they produce only small flakes, that each flake will be useful for only a

A simple example: a hunter has just killed a deer, and it is a long way back to camp so he must field butcher the animal. He does not have a tool with him, but there are small nodules in the nearby stream from which he could remove some small flakes bipolarly, or he could walk 5 km to a known quarry site, produce a large biface or some large flake tools, and return. The former choice entails less “up-front” time, but the latter may produce tools that do the job more quickly. On what basis would a decision be made? The approach of behavioral ecology is evident in research on the nature of mobile tool kits (e.g., Goodyear 1989, 1993; Kelly 1988; Shott 1986). Shott (1997) explains a diminution in arrow point size during the Woodland period in the eastern United States as an effort to reduce the cost of hunting increasingly

short

time.

A

hafted

biface,

on

the

other

hand, will entail a higher up-front cost but may be usable for a longer period of time and provide higher returns and return rates (because the haft could enhance leverage and the tool’s cutting ability; see Tomka 1999). The exact cost and benefit of a technology will depend on the particular cost of quarrying and manufacture, which is produced by local raw material distribution and quality and the functional demands of the particular technology. The question that arises is, What is the relationship between stone tools and resource

return rates? Models of stone tool procurement and use focus on how tools maximize some aspect of the tool’s utility but (understandably) do not ask how the manufacture

more scarce game. By decreasing point size

the up-front cost of hunting is reduced and the overall return rate from hunting raised. Simms et al. (1997) also make implicit use of foraging theory in predicting the differences in the labor invested in ceramics between nomadic foragers and sedentary horticultural-

of a tool in one way vs. another affects the return from the activity in which the tool will be

Using simulation, Kuhn (1994) made explicit

employed. Bifaces may be worth the initial investment in their manufacture because they maximize utility relative to transport cost

rather than larger multifunctional tools. The

ring in nodules of adequate size and quality is very far away) that the tool must be usable for

probably one of the most significant “costs”

for the immediate task (e.g., for some period

ists in the Fremont area of the Great Basin.

(but see Kuhn 1994). But what if the initial investment is so high (e.g., raw material occur-

use of optimization in determining that mobile tool kits should, contrary to expectations, consist of many small, finished tools

lengths of time far greater than that necessary

quarrying and transport of raw material is

of time longer than the resource to be processed is available)? In this case investing

entailed in a stone-tool technology, and there

are some estimates of quarrying (Carambelas and Elston 1992; Elston Raven 1992). In developing a foraging model that resource transport costs into account, calfe and Barlow

rates and

in bifaces would lower the utility of the activity in which the tool is to be used. The relationships between foraging return rate and technology are not simple because the return

takes Met-

rate is partly a function of the technology involved, as well as characteristics of the re-

(1992) noted that such a

model should be useful in making initial predictions about what kind of stone tool a person should employ under different circumstances if the goal is to maximize the return rate associated with that tool’s use. When an individual decides to manufacture a stone tool, he or she has a choice among various kinds of tools and ways to make tools. Which is chosen may depend largely on how much

source itself. Figure 5.1 is a graphic representation of what is at present only a hunch. The three curves suggest three different relationships between resource returns and the cost of technology. We assume that a resource provides 6000 kcal and that 30 minutes are needed to process it with bipolar flakes, 20 minutes 7O

A

BEHAVIORAL

ECOLOGICAL

PARADIGM

oO Oo oO

RETURN RATE (kcal/min)

600

Oo

©

|

— © ©

NO= @® Oo Oo Oo ©

_ Oo Oo

Biface

|

|

10

i

|

.

\T

20

30

40

90

MINUTES TO FIND RAW MATERIAL

Figure 5.1. Hypothetical relationships between stone-tool manufacturing technologies, the return rate that results from their use, and the time necessary to find raw material of sufficient size/quality to manufacture a tool. A biface provides a higher return rate but requires a larger investment; bipolar technology provides a lower return rate but takes less time to implement. This diagram suggests that in this hypothetical case, if it takes more than 25 minutes to locate raw material suitable for biface production that bipolar or casual flake tools would be used.

with larger flake tools, and 10 minutes biface (for postencounter return rates 300, and 600 kcal/hr, respectively). rates assume that no time is devoted

with a of 200, These to tool

here. The specific relationships will depend on the distribution of raw material and the nature of the resource. But I suspect that the general relationships indicated by the graph, if not the specific values, are generally correct. It suggests, for example, that if it is going to take 15 minutes or so to locate material to manufacture a biface (material of sufficient size and quality) but small nodules sufficient for bipolar knapping are near to hand, then the forager should use bipolar flakes and not waste time seeking stone for a biface. If it will take 25 minutes to locate any stone, and that

manufacture. If a tool is not at hand, however, then one must be made. The cost of

manufacturing a tool should be subtracted from the return of the activity it is used in to compute net gain. In Figure 5.1 the x-axis reflects the time it takes to find raw material. We assume the time to be 5 minutes for a bipolar flake tool manufacturing episode, ro minutes for large flake tools, and 25 minutes for a biface. We assume that resource processing requires five bipolar-flaking episodes, or two large flaketool-producing episodes, or a single biface. Figure 5.1 is not a predictive model but, again, only a graphical representative of the thinking behind the approach advocated

stone is adequate for making a biface, then

the forager could manufacture simple flake

tools only ora biface. A biface, in fact, may be chosen, even if it takes more time, as it could

be used for future resources (thus, perhaps we should “average” the cost of this particular biface over future food-extraction opportu71

ROBERT

L. KELLY

information, encoded in symbols rather than in DNA, is “inherited” by one generation from another. One might assume, then, that it operates according to a set of principles of inheritance as biological evolution does (Bettinger 1991; Boyd and Richerson 1985), but this does not mean that those principles are the same as or even analogous to those gov-

nities). This particular example may seem unrealistic because it is always assumed (but probably not always true) that a forager carries tools adequate to the anticipated task or knows where raw material can be found quickly. Still, we could work the example from the perspective of a forager sitting at camp thinking about whether to invest time in foraging with the tools at hand or to spend time on tool manufacture. At some point

erning biological change.

To understand cultural change we must

know how changes in adult activities affect

people had to make these sorts of decisions, and all choices would have entailed costs and

enculturation, the process whereby children learn their culture. For example, Hewlett

benefits. Experimental data are needed to construct predictive models using this approach. We do not know, for example, if there is a difference in the return rate of harvesting grass with a simple flake vs. a hafted biface. We do not

(1991) notes that poorer male hunters among

the Aka spend more time with their children, whereas the children of better hunters spend less time with their fathers. How does this affect the development of these two sets of chil-

dren? How does a population’s valuation of different activities shift as a function of

have adequate data on the time it takes to quarry and manufacture different kinds of

changes in the way children are raised? What role does intrasocietal variation in encultura-

tools. If a resource can be harvested with a simple flake tool just as easily as with a biface,

tion play in cultural change? Research into

then flake tool production will probably be

this area is really just beginning (e.g., Boyd and Richerson 1985; Cavalli-Sforza and

chosen over biface production, especially if raw material for bifaces is not readily avail-

able but nodules suitable for flake tool production are. A number of archaeologists, myself included (Kelly 1988), have argued or assumed that bifaces are a more efficient way

Feldman

1981;

Hewlett

and

Cavalli-Sforza

1986), but it is necessary for the development of archaeological models that can anticipate changes in the rapidity and direction of cultural change as a function of changes in the way children are enculturated. Although the connection between enculturation studies and behavioral ecology has not been explored much, we might expect that ecological factors affecting adult activities that in turn affect child rearing could play a large role in cultural change (Kelly 1995: 153-156). There appears to be, for example, a large difference in the modal personalities of children who are largely socialized by their parents vs. children who are largely socialized by their peers. The former see resources as limited and seek to acquire them through individual manipulation, especially of technology; the latter see resources as not limited, but controlled, and the key is not to increase production individually but to find social ways of acquiring resources from those who already have them. This is of course an over-

to carry flake tools than a casual core, that is,

that more usable flake edge can be removed froma biface than any other core form. But to my knowledge no one has tested this hypothesis systematically. In sum, much basic information needs to be collected for behavioral ecology to be implemented in archaeology.

Models of Cultural Transmission of Information Humans pass on information between generations about kinship, subsistence, religion, morals, aesthetics, tools, and so on. Because cultures change through time, some information is obviously not passed on, and some new information (brought in through diffusion, independent invention, or errors in the enculturative process) is added. In other words, information is selectively passed on, making culture analogous to genetics, in that 72

A

BEHAVIORAL

ECOLOGICAL

PARADIGM

pattern and one that is expected from the theory of behavioral ecology given the con-

simplification, but it points to one factor— the systematic effects of enculturation—that

straints of child care on women (Kelly 1995: 262-270).

could bias a member of a society toward accepting one behavior over another, for ex-

Even

ample choosing to hunt individually rather

with well-warranted

assumptions,

however, we still must grapple with the fact that archaeological assemblages are nowhere

than communally, hunting with a bow and arrow rather than with an atlatl, or living

near as fine-grained as ethnographic foraging

close to rather than far from relatives.

data. Is this a problem? In his ethnographic

analysis of Inuit hunting, Eric Smith (1991) found that the diet breadth model, probably

Archaeological Data Record a

Composite of Activities

There is also a problem of scale. Applications

the most commonly employed model in ar-

pends not only on the food resources avail-

over short rather than Jong time scales. That

chaeological applications of foraging theory, was a more accurate predictor of food choice

of behavioral ecological models to ethnographic data have shown that food choice deable and the technology used in their acquisition but also on whether the forager is,

does not bode well for the application of behavioral ecology in archaeology. But Smith’s

unmarried or married, a parent or not because these different statuses create different

to archaeology; would longer time scales mask some variability but still be capable of showing temporal trends in dietary diversity? This is an essential question to an archaeology based on behavioral ecology. Many questions about prehistoric dietary change come down to questions about the diversity of resources taken and/or the intensity of resource use. But if the data we turn to are long-term conglomerates of food choices, then the effects of different seasons and different years may be inseparable. We see the maximum ranges of food choice of a long time period (Madsen 1993; see also Grayson and Cannon 1999; Simms and Zeanah

long time scale was still but a blink of the eye

for example, male or female, young or old, opportunity

costs

of

different

behavioral

choices. If these factors are important to understanding how subsistence changes over time, then how do we bring behavioral ecological models to bear on archaeological data that are conglomerates of activities (Grayson

and Delpech 1998)? The botanical and faunal remains in a site can be the results of years, if not decades or centuries, of the activities of men and women, young and old foragers, and so on. There may be an isolated case here and there, but I do not know of any unambiguous way to systematically separate something even as fundamental as men’s and women’s activities for the archaeology of hunter-gatherers, although there are some notable efforts being made (e.g., McGuire and Hildebrant 1994). But for the most part we must rely on the strength of ethnographic analysis (rather than analogy) to provide archaeologists with strong reasons to warrant particular assumptions that are to be used in models about, for example, male/female choices of food resources. Although some argue that the assumption that men hunt large game and women gather plant food and opportunistically take small game is unwarranted and an imposition of western culture on prehistory, it is a very strong ethnographic

1999). Better ways to measure dietary diver-

sity meaningfully given the exigencies of the archaeological record of ancient foraging societies are needed (see Grayson and Delpech 1998 for an example). Computer simulations that construct archaeological assemblages are needed to ascertain what kinds of foraging questions can realistically be posed to the archaeological record of hunter-gatherers. Such work has existed for some time in archaeology (e.g., Thomas 1974) but has not been conducted within the framework advocated here. Those who would take up the task must bear in mind the coarse resolution of hunter-gatherer archaeology. Extremely complex, sophisticated models will be of no 73

ROBERT

L.

use if they make unrealistic demands of forager archaeology.

KELLY

1994; Bettinger and Baumhoff 1982) argues

that the pre-Numa were nomadic and focused on resources such as large game (that provide high return rates); the Numa, on the other hand, are alleged to have been less no-

The Need for Fine-grained Models But even while recognizing that the specific assumptions of fine-grained models may not be verifiable archaeologically (e.g., did chil-

madic and had a wider diet breadth, includ-

ing small seeds and pinyon nuts, utilizing some different technologies (e.g., seed beaters) and tactics (e.g., green cone pinyon procurement). Bettinger argues that the Numa out-competed the non-Numa, who migrated,

dren forage for themselves?), it is also clear

that extremely simple, unsophisticated models will be of no use. The reason was pointed out above: something always competes for one’s time. If a man goes hunting, he may

went extinct, or intermarried and lost their language.

forgo opportunities to participate in prestigious ceremonies. Sometimes people act in

terms of instantaneous returns, but at other times they act in terms of some return that will not be manifested for some time to come

Numic and pre-Numic lifeways, as envisioned by Bettinger, may have entailed different trade-offs for men and women. Among many foragers the sharing of meat from large

In the past, foraging models focused solely on trade-offs between food choices: do I go

can be “show-offs” and spend time hunting large game to acquire prestige (and possibly

game is a primary way that men acquire prestige. Hawkes (1996) argues that some men

(e.g., building up favors or raising children who will support parents in old age).

more mating opportunities). Others, she sug-

for deer or pickleweed? But the choice to go

foraging (never mind what the foragers are going to forage for) competes with the choice

gests, are “providers” who seek nonshared and hence less prestigious resources but pro-

trading or religious activities. Foraging models do assume that a forager tries to maximize the rate of return, but this is always seen in re-

(Recent reanalysis of the Ache data on which Hawkes based her conclusions suggests that

vide more food directly to the family larder.

to spend time with children, or spouses, in

show-offs are young, unmarried men and

older men with no dependent offspring; L.e.,

lation to other options, an activity’s opportunity costs. In addition, foragers may be aiming at a short-term goal (maximize the

amount

of food

brought

home)

one strategy is more likely at different stages in a man’s life [Wood and Hill 1999].) Bet-

tinger argues that pre-Numic peoples relied

or some

more heavily on large game than the Numa; if true, this would suggest that they had the

long-term goal (minimize the time spent foraging in order to spend time with trading

potential to have a higher frequency of showoffs than in Numic populations. Pre-Numic men might have been reluctant to give up hunting even if in the long term that choice made them less competitive with the Numa because they might have been unable to overcome the immediate perceived loss in status and benefits that would result from forgoing an opportunity to hunt in order to collect small game or to care for children (see also Whitley [1994] for another example of the possible relationships between pre-Numic and Numic men and women). Women also make trade-offs, and theirs may be linked primarily to children and their

partners to arrange a good marriage for a

young daughter). Modeling these sorts of be-

havioral trade-offs is really just beginning in ethnographic studies (see Hawkes 1996). They are by no means simple or straightforward, and their archaeological study requires sophisticated modeling. The issue of the Numic migration into the Great Basin provides an example. Some argue that about 1000 years ago speakers of Numic languages migrated into the Great Basin of western North America from a homeland in southeastern California and replaced (or subsumed) the existing “preNumic” population. Bettinger (e.g., Bettinger

74

A

BEHAVIORAL

ECOLOGICAL

care. Research with the Hadza and Bushmen suggests that if children do much of their own

offs is essential for the testing of ideas drawn from behavioral ecology.

foraging, higher population-growth rates can

result. This research also suggests that the amount of foraging children do is linked primarily to the environment—how easy it is to get food and how much danger children are in while foraging (Blurton-Jones et al. 1989, 1994; Hawkes et al. 1989, 1995, 1997). These environmental factors would have been the same for the Numa and the

BUILDING

cultural, and that means

then, is behavior maintained, even if ineffi-

cient in the face of changing circumstances, because it is a flag of affiliation? And is this question relevant to the scale of behavioral change that archaeological data record (especially that of foragers)? We do not know. But asking this question means that archaeologists will have to consider the interaction of ethnic groups, and this means that we must traverse some unfamiliar terrain: the ways in which people categorize themselves and are categorized by others and how groups of

ent notion of child care, one that resulted in their leaving children at camp and that al-

people

lowed them to forage more efficiently than pre-Numic women (and presumably permitted them to raise more offspring to adulthood)? Enculturation plays a role in determining subsistence choices and behavioral an

environment

can

that ethnic differ-

ences, however they are perceived, can become boundaries. These in turn help establish the costs and benefits of behavioral choices and the ensuing value of trade-offs by, for example, shutting off resource areas. When,

pre-Numa. An additional factor, however, is

and

BRIDGES

But I think there is even more to it. People are

whether women feel compelled to forage with their children. In some foraging societies women feel a need to do so; in others they do not. This is related to the environment in that in extremely arid places the heat stress that might result from carrying a child while foraging may make it impossible for a woman to gather sufficient food. But it is also related to the availability of caretakers (Kelly 1995). What if Numic women entered with a differ-

trade-offs,

PARADIGM

who

view

themselves

as

different

interact under different circumstances. In addition to environment, technology, and popu-

lation density,

ethnicity—how

individuals

relate to others they perceive as different— conditions behavior. Does this matter? We have already noted that people can perceive the cost and benefits of behavioral trade-offs differently. If different perceptions of the trade-offs are associated with different “ethnic” groups, then one

have

more than one adaptive “peak,” as Bettinger (1991) has noted. Differences in child care and large-game hunting could have constructed two adaptive peaks between the Numic and pre-Numic. This is an important point to which archaeological applications of behavioral ecology have not paid sufficient attention. But did Numic women leave their children at home more than non-Numic mothers? We will certainly never know. But it is clear that behavioral differences that are summarily described as “ethnic” could predispose a group toward adopting or not adopting a technology or behavior. It should be possible to construct models that predict the results of different behavioral trade-offs and to translate these into predictions that are testable with archaeological data. Recognizing and modeling these behavioral trade-

could see how one group might replace another, but ethnic differences in this case

would need to be translated into behavioral

differences. Here we could think of ethnic

groups as pseudospecies, as the units undergoing adaptive change (Bettinger 1994).

But ethnic groups are not entirely like species because they are cultural not biological constructs. And here we can see some connections between behavioral ecology and other approaches discussed in this volume. The perception of difference is not a given but is part of the adaptive process. What do I mean by this? Consider these two facts. First, foraging models can predict how groups 75

ROBERT

L.

Mean Per Capita Return Rate

might interact under different circumstances, whether they will share freely, or trade, or defend their resources depending on the overall return rate of their environment (given their technological means of exploiting that environment) and the extent to which environmental returns are correlated (see Kelly 1995; Winterhalder 1997). Two groups whose return rates are not correlated—because they take different resources or because the climates of their respective local environments are different (a poor harvest in one group’s region occurs during a good harvest in the other’s)—may share, whereas two groups

KELLY

Phase A

Phase B Per capita return for new members if current members ‘skim’

N n Increasing Group Size Figure 5.2. A modification of Smith's (1985, 1991) group size model. The mean per capita re-

whose resources are correlated, such that a

turn rate rises until a maximum

bad year for one is always a bad year for the other, may fight over resources during bad years. Second, for any given activity there is an “optimal” group size, one that maximizes the individual rate of return. Borrowed from Smith (1985, 1991), Figure 5.2 shows how group size can be related to the per capita return rate in some cases. In this figure the per capita return increases for the foraging activity until a group size of N is reached. Up until a group size of N, additional foragers add to the per capita return rate; adding members after N foragers does not enhance the activity, but because the food gathered must now be shared with more people, the per capita return declines. One could imagine, for example, that communal net hunting (as among the Efe or Aka pygmies) is made more efficient with more foragers because more beaters or net monitors means fewer animals escape—until a saturation point is reached at which no more beaters or net monitors are needed. Smith pointed out that where N is somewhat more than 1, a dilemma is produced: current members will not want additional people to join because the per capita return rate will decline, but because the per capita return rate of a group with a size greater than N will still produce a higher return rate than foraging alone, lone foragers will want to join

is reached at a

group size of N; after this, additional group members do not increase task efficiency and lower the mean per capita return rate. At a group size of n the mean per capita return is the same as foraging alone. Thus, two basic phases should characterize any group formation process: a phase of inclusion (A), in which potential joiners must be courted, and a phase of exclusion (B), in which potential joiners will be repelled or added at a cost to the new members, perhaps in the form of a smaller share of the returns than current members receive.

refuse to “share.” One possible outcome is that members allow others to join, but at a

decreased per capita share—current mem-

bers skim off the top before sharing (see discussion in Kelly 1995:314). This fact establishes two phases of group formation, noted as A and B in Figure 5.2. And if we can shift scales here—from the short time scale involved in the formation of a net-hunting group to the long time scale involved in, for example, the formation of competitive villages—we can see two important

phases in group formation that might be reflected in a coarse-grained archaeological record. In Phase A current group members should want others to join, and groups would compete for members. One can imagine that this could be accomplished by a rhetoric and manipulation of symbols of inclusion— “discov-

ering” kin relations, converting the economic

the group. Tension can arise therefore be-

unit from a household to a lineage, for ex-

tween members and joiners, and members may face retaliation from joiners if they

ample. Because those being solicited would

76

A

BEHAVIORAL

ECOLOGICAL

have options, there would be little potential for exploitation—if not satisfied with one group, a person could “vote with his or her feet,” shifting allegiance to another. Such a situation may arise where population density is relatively low, where individual mobility is

expect different sorts of hierarchical societies to arise, depending on the spatial extent of resource fluctuations (Kelly 1995). In that for-

mulation I focused on population density as the driving force and discussed the effect that different environmental circumstances had on the relations between men and women. Discussion here suggests another dimension of variability: the extent to which people have choices about which group to join. Optimal group-size models are relevant to understanding how people decide to let others be “accepted” or not—in this case, to be part of their “ethnic” group or not. The world today unfortunately provides ample evidence that people discover or create ethnic differences in situations of competition, as

still possible, and where one group is not con-

sistently in a favored position (e.g., located on a secure food source such as a prime location of a river with a large, reliable salmon run or on prime agricultural land that has a consistent water source). If I understand Feinman (this volume) correctly, it is under these circumstances that something like his “corporate” political organization could form. If population continues to grow (or if envi-

one group attempts to shut off membership. Ethnicity, and ethnic affiliation, is part of

ronmental conditions degrade), group for-

mation may reach Phase B. Here groups have

reached their maximum size, and now potential members will be shut out or admitted at a cost. It is under these circumstances that we would see a rhetoric and manipulation of symbols of exclusion. To maintain as high a return rate as possible current members will have to decrease the share of those who join after the group

has reached

gies are born

that

adaptation, not just a unit of adaptation. We

need models that combine sharing and group-size models to predict when ethnic differences will be permeable and impermeable boundaries. The

its maximum.

This results in hierarchical organization, and because it requires some justification, ideolo-

“explain”

why

PARADIGM

CONCLUSION

study of evolutionary process requires

that we study variability in prehistoric lifeways and that we examine the nature of cul-

tural transmission of information. A linking

some

benefit from the activities of others. It is under these conditions that some individuals

of the models of behavioral ecology, models that focus on the trade-offs associated with competing activities, with models of cultural transmission may be of great value in testing hypotheses about evolutionary change. These hypotheses must be compatible with the nature of archaeological data, which for hunter-gatherers means units of analysis that

can come to command the labor of nonkin, as Arnold (this volume) notes; this could include

slavery. Hayden (1998) argues that there are agerandizers in every society and that they appear where it is possible for them to increase production. I suggest that they arise in Phase B and in those cases where it is not only possible for them to participate in self-aggrandizing activities like competitive feasts (such as the potlatch) but where the ordinary folks who support aggrandizers (for prestige is not taken, but given; see Kelly [1995:306]) acquire some benefit from the cost of supporting an aggrandizer. (This does not mean that they are doing well under a “petty captain” but that they are doing better than whatever the alternative is.) Elsewhere I have pointed out that we could

will be conglomerates

of activities.

Such

a

fact will require that the study of evolutionary processes be amenable to analysis of data drawn from large areas and long time periods. This may limit us to the study of processes that are manifested over such large scales of analysis and force us to leave fine-grained processes, although built into archaeological models, to ethnographers of living societies. Additionally, the interpretation of archaeological remains must be consistent with the theoretical paradigm used to explain

77

ROBERT

L.

KELLY

tinue to denigrate it within our own field? We cannot hope for anthropologists of different backgrounds and inclinations to “all get along” if the postmodernists think the “scientists” are evil and if the scientists think post-

the changing frequencies of those remains. For hunter-gatherer archaeology this means that stone-tool production and use should be examined with the framework of optimal foraging models. Finally, the issue of ethnicity and the role that group construction plays in evolutionary change can be examined within a framework of behavioral ecology, especially if it incorporates models of enculturation and cultural transmission. A few concluding comments. In this conference we were asked to try to build bridges among the various theoretical paradigms described in the volume. I have made a small movement in that direction. But by focusing on behavioral ecology, I may create the impression that I think it the only worthwhile paradigm for the twenty-first century. I do not. Let me make this point clear. Most everyone agrees that diversity is good, that a “thousand archaeologies” (Schiffer 1988) are desirable. Yet in conversation, and often in print and in reviews, we do not make good on this claim. The vitriol that anthropologists have been spewing out at one another in recent years has not gone unnoticed by the popular media, and many in and out of the field predict anthropology’s demise. They might be right. Only a few an-

modernists are morons.

How do we physicians heal ourselves? First, the discipline structurally and culturally should reward efforts at unification. At present the culture of anthropology is such that we reward polemic, bombast, and show-

manship rather than the serious testing of ideas. Showing respect is not just nice; it’s good scholarship. Having said that, I still assert that a scientific approach is an essential element of any effort to learn about the past because of its emphasis on the testing of ideas against empirical data. But having said that I must also confess that postmodernism’s emphasis on reflexivity is important in showing us where our thoughts and background place

blinders on us. And this brings us back to par-

adigms. Because paradigms are ways of looking at the world and because there is no one right way to look at the world (although there are wrong ways), continuing to promote one

way of looking at the world, as we so often do, is not especially useful or interesting. “Our ways of thinking are like snakeskin: without them we perish, yet we must con-

thropology departments have actually split (Stanford, most recently), but the fact that

tinue to shed them in order to grow” (Kelly 1992:263). It is for this reason that I emphatically state that behavioral ecology is not the one right way of looking at the world and repeat that I use it because I find it a useful tool for learning about human behavior and prehistory. Someday, I suppose, I will shed it when it is no longer useful.

any should do so is alarming. After all, are not anthropologists the ones who preach the value of diversity? The Snowbird conference was not so much about theory as it was about paradigms. Par-

adigms are ways of looking at the world; they can be thought of as different cultures. How can we teach the value of diversity if we con-

78

6

Food, Lies, and Paleoanthropology Social Theory and the Evolution of Sharing in Humans STEVEN

L. KUHN

AND

CATHERINE

Like our closest living primate relatives, the African hominoids, human beings are intensely social animals. It follows that our hominid ancestors exhibited a similarly high degree of sociality. Yet just as it is certain that humans and human ancestors have always lived in groups, it is equally certain that the sizes and compositions of those groups, as well as the ways in which social relations are structured, maintained, and negotiated, have changed radically over the evolutionary history of our taxon. Understanding how the structure of human societies may have changed since the hominid line first diverged some five million years ago has been a constant theme in paleoanthropological research. This essay addresses current and

future directions

for the development

SARTHER

erature of evolutionary ecology. We argue that extending existing forms of social theory to the evolution of cooperation will ultimately require explicit consideration of the cognitive structures underlying both human and nonhuman social interactions. In the final section of the paper we propose a highly provisional scenario for the evolution of cooperative economies based on food sharing and outline a series of potential archaeological indicators for changes in group size, social organization, and degrees of cooperation and sharing. Any productive discussion of theory must begin with an explicit statement of research agenda. There is no denying that archaeologists today work from a wide range of epistemological and ontological positions. At the same time it is difficult to escape the impression that many apparently profound differences in “ways of knowing” the past ultimately stem from wanting to know different things about the past. The facts and phenomena of greatest interest to paleoanthropologists are not necessarily central concerns for other archaeological researchers. Most important, scholars working on later time ranges can take for granted many of the very same elements of behavior and cognition that paleoanthropologists seek to investigate. Thus, what are seen to be crying needs for the

of

social theory relevant to human evolutionary studies. It focuses on the evolution of sharing and economic cooperation within social groups. These are behavioral tendencies thou ght to be typical of modern-day humans but relatively uncommon among nonhuman social organisms, including the higher primates. A brief history of archaeological approaches to food sharing among PlioPleistocene hominids, beginning with Isaac’s pioneering work in the early 1970s, illustrates how these issues have been addressed archaeologically and sets the stage for a more general consideration of the evolution of

study of human

ancestors may

seem either

moot or peripheral to other practitioners of anthropological archaeology.

cooperation in social animals. Our treatment

of sharing behavior draws heavily on the lit79

S. L. KUHN

AND

C. SARTHER

On first consideration it might seem that social theory has little to do with paleoanthropology, a field that unites archaeology

ancestors stretches back over two and a half million years, but the density of data points (“sites,” artifact and fossil-yielding deposits)

Pleistocene humans

pared to later time periods: there are many more documented sites in northern Arizona

is extremely low in both time and space com-

and biological anthropology in the study of

and human

ancestors.

This area of inquiry is commonly viewed as a largely empirical research domain, domi-

dating to a single 200-year-long phase of late

Anasazi prehistory than there are archaeological localities representing the first million

nated by arduous fieldwork and exhaustive

laboratory studies of scarce traces of ancient

human behavior and little informed by high-level theory on the whole. In reality,

years of the hominid record in sub-Saharan

Africa. Moreover, because the effects of post-

depositional processes of disturbance are ex-

evaluation of the sparse and ambiguous Pleis-

acerbated by age, Paleolithic archaeological sites tend as a rule to exhibit low or at least

tocene archaeological record should require an especially comprehensive theoretical foundation. Social theory is often considered

suspect levels of spatial integrity and chrono-

the purview of classic social-scientific and “mentalist” approaches to the human condition, whereas paleoanthropologists tend to favor materialistic approaches derived from

logical resolution (e.g., Petitt 1997; Stern 1993, 1994). Apparently instantaneous,

tions about social interaction are central to many evolutionary models: both biological and cultural reproduction are channeled through the same relations among individuals that constitute social systems. In fact, some of the most hotly debated and extensively researched questions about Pleistocene hominids revolve around the structure and composition of their social groupings. Notions about the evolution of economies based on sharing and Machiavellian intelligence have put the social realm at center stage. Although paleoanthropologists are as interested in social phenomena as any of their anthropological colleagues, human evolution research places different demands on theory than do many other forms of inquiry. First and foremost, an evolutionary perspective 1s vital. Archaeologists debate the relevance of different schools of evolutionary thought to the explanation of changes in human culture and behavior over the past few thousand years (e.g., Barton and Clark 1997; Boone and Smith 1998; Dunnell 1980, 1989a;

metric dating, it is impossible to reliably link multiple occurrences in time (should they be discovered). Thus, it is unlikely that we will ever have a very clear and comprehensive picture of hominid life at any particular moment in the remote past. On the other hand, the Paleolithic record is absolutely unique in that it presents evidence for trends in ecology, technology, sociality, and other aspects of behavior that played out over tens or hundreds of thousands of years. Paleoanthropology provides us with an opportunity to study changes in human behavior over evolutionary time scales, so it is only appropriate to apply theory grounded in evolutionary principles. More so than their colleagues in other branches of archaeology, paleoanthropologists must also be concerned with the interaction between biological and _ behavioral evolution in the human organism. Some ad-

high-resolution “snapshots” of past behavior are extremely

O’Brien

1996a,

this

volume;

O’Brien,

few

and

far between,

and,

given current limits in the accuracy of radio-

evolutionary and ecological theory. Yet no-

vocates

of “neo-Darwinian”

or “selection-

ist” approaches to archaeology argue that it is pointless to separate the biological and behavioral (Dunnell 1980; O’Brien and Holland 1995a). This would seem to be a rhetorical position, adopted so as to validate the application of Darwinian models of evolution by natural selection to the archaeological record of human material culture over the past few thousand years. Although it would

ed.

1996; Schiffer 1996; Spencer 1997). However, approaches founded in evolutionary theory are ideally formulated for the purposes of paleoanthropological research. The archaeological record of humans and human

80

Foop,

LIES, AND

PALEOANTHROPOLOGY

be inaccurate to suggest that there has been no significant biological evolution in the hu-

quently

associated

with

analyses

of eco-

relatively recent changes such as dental reduction (Calcagno 1989, Smith et al. 1986)

with subjects as wide ranging as exchange and sharing (Hawkes 1992; Winterhalder

man species since the end of the Pleistocene,

nomic or ecological phenomena (foraging theory), but formal theorists have also dealt

are

1996a, 1997), the sizes and compositions of

comparatively

minor

(but

cf.

Foley

1991). In contrast, it is certain that the fundamental physical and cognitive capacities of

social groupings (e.g., Dwyer and Minegal 1985; Lee 1994; Smith 1981, 1985; Winter-

ceding two and a half million years. Any ideas about social and other behaviors that

ing 1991), the development of social inequality (Boone 1992), and even the construction

halder 1986), information networks (Lans-

hominids changed profoundly over the pre-

of monuments (Neiman 1997). Formal the-

we bring to paleoanthropology must be ca-

ory tends to rely on highly simplified, mecha-

pable of coping with fundamental changes in

nistic models. The dynamic behavior and wider implications of formal social-theoretic

organically based capacities and propensities

of the social actors themselves; furthermore, these ideas must accommodate the behavior

models

VARIETIES

OF

SOCIAL

are explored through

mathematical

solutions or computer simulations. Although

of both modern humans and intelligent social creatures that were, nonetheless, not exactly human.

this approach is assiduously reductionist, it is a reductionism of the methodological rather than the ontological variety (Brandon

1996:180-181). The structural aspects of formal models are either abstracted logically

THEORY

A theory is a general statement about the way the world works, based on a series of unques-

from “first principles” or derived from com-

parative observations of a range of social or-

tioned assumptions or premises. Social the-

ganisms

ory (as we understand the term) refers to any general proposition about the dynamics or structure of social relations, including the mutual interactions between social life, environment, and material culture. The term is not, or should not be, confined to certain subsets of theory or certain aspects of social behavior. For the purposes of the present discus-

and

(e.g., Foley and Lee 1989; Tooby

Devore

1987).

Most

important,

these

theories make minimal assumptions about the tendencies and capacities of the organisms involved, such that they apply equally to humans, social animals, and the simple automata of computer simulations. Much of the social theory traditionally used in the social sciences is of a very different sort. The range of ideas falling under this rubric is truly vast, and it is patently impossible to do justice to it here. Nonetheless, much of the social theory applied in the “human sciences,” whether of neoevolutionary, radical, structuralist, postmodern, feminist, or some other ideological stripe, does share some basic characteristics. First, it tends to rely on narrative or experiential models of human social behavior rather than rigorously structured formal models. Fur-

sion, theoretical notions about society in use

by anthropologists and other social scientists today have been divided into two broad groups: formal theory and the more traditional theory of the social sciences. This simple dichotomy subsumes a remarkably broad set of approaches, many of which are mutually contradictory, and it would almost certainly disturb the authors of many of these ideas. The following classification of theories is entirely structured around the problem of identifying ideas that are most applicable to human evolution research. What we call formal social theory is derived mainly from the biological sciences, particularly evolutionary ecology, although it has found wider application (such as in economics). Formal models are most fre-

thermore, traditional social-science theories

of society are typically holistic rather than reductionist: although they may emphasize one or two dimensions of human existence, they frequently seek to encompass many more. Finally, the traditional models used in the

81

S. L. KUHN

AND

C. SARTHER

at ideas about social behavior that both do justice to the diversity and complexity of social relations among modern humans and are capable of integrating uniquely human characteristics into a common framework with other nonhuman social creatures (including our own forebears). It is difficult to imagine that the broad range of human social behavior could be reconstructed from first principles. Whereas formal theory may provide the most useful working models for application in an evolutionary time scale, ideas

social sciences are based on fundamental (although often unexamined) assumptions about universal aspects of human nature. They generally assume, among other things, that individuals communicate through the use of symbols, plan around expectations about the future, and recognize and respond to the intentions of others; in short, they pre-

suppose that actors in social interaction are

physiologically, cognitively, and psychologically human. The structure of the models is derived exclusively from human experience and history. Whether informed by broad

about how social relations are actually real-

ized within and among human groups are the only source of verisimilitude for these simple

comparative analyses or detailed case studies,

the focus remains exclusively on the actual

human condition. In reality, these two great families of the-

models. Formal models can be used to ana-

tionists such as Darwin and Wallace were strongly influenced by Malthus and other so-

but theories of society from the human sciences provide vital perspectives on the unique

ory may have common

lyze the general evolutionary motor behind long-term changes in human social behavior,

roots. Early evolu-

social sciences has grown much more profound than it was in the late nineteenth cen-

historical directions taken by the evolutionary process. The principal obstacle to combining elements of formal and social-scientific theory is

apply evolutionary concepts to problems in psychology, sociology, economics, and other

each other in mutually agreeable and intelligible terms. Formal models tend to be struc-

cial thinkers of the time. However, the con-

ceptual separation between biological and tury. Interestingly, more

recent attempts

getting different social theorists to speak to

to

tured in terms of material or economic para-

fields may represent a renewed convergence

meters such as time, energy, or reproductive

of thought. It should be clear that what we have re-

with the limitations and goals of human evo-

success. Some traditional social-scientific theories are also constructed around these currencies, but most deal with more abstract

type more exclusively associated with the social sciences. Most important, formal models

cepts necessary for linking formal and classic

ferred to as formal theory offers a better fit

(and more uniquely human) phenomena such as status and power. We argue that the con-

lution research than does social theory of the

social scientific social theory will come from an understanding of the nature and evolution

make fewer assumptions about the nature of social actors. Just as we know that our ances-

of human cognition. On one side evolutionary theorists will need to integrate increasingly realistic models of cultural transmission into their theoretical formulations: various “dual-inheritance” models (e.g., Boyd and Richerson 1985; Cavalli-Sforza and Feldman 1981; Laland 1992; Richerson and Boyd 1978, 1992) would seem to be a major step in this direction. The other side of the endeavor will involve a critical examination of the fundamental cognitive underpinnings of distinctive human social arrangements today. Most of us would agree that traffic in symbols has

tors were not exactly like us, we must also ac-

knowledge that capacities on which tradi-

tional theories of social behavior rest may not have existed at all times in the ancient past. In addition, most formal models are based on

evolutionary theory, or at least are structured in terms of the same currencies and variables, whereas many classic social-scientific theories are not. Nonetheless, the “traditional” theory of the social sciences has an important role in human evolutionary research. The ultimate goal in identifying or developing theory useful for paleoanthropology is to arrive 82

Foop,

LIES, AND

PALEOANTHROPOLOGY

not removed humans from an evolutionary arena, but it has changed the rules of the game somewhat. With that premise the obvi-

sion of labor along gender lines (see Kelly 1995:261-292 and references therein), sharing seems to be a mainstay of hunter-gatherer

ous next step is to determine just which rules have changed and why.

existence, particularly among mobile, egalitarian foragers living in small groups.

This discussion now turns to the evolution

In this regard the contrasts with our closest

of sharing and cooperative economies, a set of issues closely bound up with notions of

biological relatives, group-living primates,

are especially striking. Examples of apparent sharing have been reported by primatologists

how human and animal societies work and

how hominid social arrangements may have changed over the course of the Plio-

Pleistocene.

The

questions

of when

(e.g., Anderson and McGrew 1984; Boesch and Boesch 1994; Ferrari 1987; Perry and

econ-

Rose 1994), but the very fact that observers

omies based on food sharing first appeared

take

ductive

food on a regular basis, but this accounts for

and how they might have evolved have excited considerable debate and much proresearch

among

Paleolithic

archae-

to

describe

individual

a small portion of the total diet, and the sharing often takes place ina restricted social context (e.g., alliance formation among males) (see, e.g., Boesch and Boesch 1989; de Waal

ambiguous evidence for food sharing and base camps in the earliest Plio-Pleistocene record. Consideration then shifts to the more in human

trouble

episodes attests to their comparative rarity. A few nonhuman primates do appear to share

ologists and human ecologists. We begin with a review of debates regarding decidedly

general issue of cooperation

the

1989; King 1994:65-70; 106-113; Stanford 1998).

and

McGrew 1992: A good deal of

nonhuman societies. Finally, we conclude by suggesting directions for future develop-

what has been described as sharing among primates appears to be better described as

making

tential benefit of having exclusive access to a resource does not outweigh the cost of de-

ments in both general theory and criteria for

“tolerated theft,” situations in which the po-

these social and cognitive models

more amenable to empirical testing. “PASS ME THE WILDEBEEST”: EVIDENCE

FOR SHARING

fending

1987;

King

any nonhuman primate as it is to recent, mobile hunter-gatherers (Hawkes 1992). Al-

though humans do not always cooperate and they are not the only organism that shares, the development and elaboration of economic cooperation among humans is unequaled among higher organisms. In an influential series of papers published

in the 1970s and early 1980s, Glynn Isaac provided the first comprehensive model for the emergence of generalized sharing, division of labor along gender lines, and other supposedly unique aspects of human be-

1990) can be counted among nature’s altru-

ists. However, although these behaviors are neither unique to humans nor typical of all economic

Blurton-Jones

tral to the social interaction and foraging of

PLIO-PLEISTOCENE

interaction,

(see

of individual cases, food sharing is not as cen-

IN THE

Although concepts of ownership and private property vary widely among ethnographically documented hunter-gatherers, sharing is “the sine qua non of hunter-gatherer culture in the minds of contemporary anthropologists” (Kelly 1995:161). Humans are not the only animals among which individuals allow others to consume resources that they have obtained. Social carnivores (e.g., Kruuk 1972; Mech 1970), nesting birds, and eusocial insects are also known for this kind of behavior. Even female vampire bats (Wilkinson

human

it

1994:65-67). Regardless of what one makes

DISPUTED

havior

(Isaac

1971,

1972a,

1978a,

1978b,

1983).' The components of this systemic food-sharing model, known alternatively as the “home base hypothesis,” included the following: 1) increasing consumption of meat; 2) food-sharing, specifically in the form of

cooperation

and food sharing seem to be more highly developed in humans than in any other organism. Along with remarkably consistent divi83

S. L. KUHN

male

provisioning

of mates

AND

and offspring

C. SARTHER

vial action, Binford (1981b)

argued further

that early hominids were probably marginal scavengers rather than intrepid hunters. The implication that scavenging precluded food sharing because of the ostensibly small quantities of meat involved led some researchers to view scavenging and a home base pattern of land use as mutually exclusive options (Rose and Marshall 1996:311). Binford’s application of middle-range research to these issues (e.g., Binford 1981rb, 1985) and Isaac’s own willingness to further

(e.g. Lovejoy 1981), sexual division of labor,

and socialization organized around home bases; 3) pair bonding and increasing male investment in child rearing, with a concomitant lengthening of infant dependency; and ultimately 4) increasingly sophisticated communication skills that developed as hominids socialized at their home bases. Rather than stressing bipedalism, hunting, or technology as a prime mover, Isaac envisioned a feedback system of adaptations centered on the flow of resources through social networks. The cornerstone of Isaac’s formulation was really the provisioning of mates and offspring at semifixed points on the landscape, which was seen as the beginnings of something essentially human. The primacy of this aspect was sup-

explore the problem (e.g., Isaac 1983) have

led many members of at least two generations of paleoanthropologists to conduct relevant taphonomic, actualistic, and ethnoarchaeological research (e.g., Blumenschine 1987; Bunn 1986; Bunn etal. 1980; O’Connell et al. 1988; Schick 1991).* Many of these projects have extended research programs that had

ported by analogy to living hunter-foragers, among whom food sharing, sexual division

of labor, and social activity focused at home bases or campsites were easily observed.

preceded or appeared at about the same time as Binford’s critique (e.g., Behrensmeyer

tological and archaeological evidence indicated an emergence of this fundamentally

tion and consumption, raw material and animal part transport, site formation processes,

Pleistocene: 1) bipedalism, which freed the

mality of food resources in paleoenvironments continue to demonstrate the complex-

1975; Brain 1981). Studies of meat acquisi-

Isaac argued that three aspects of the paleonhuman

behavioral

complex

by

the

and the availability, predictability, and opti-

Plio-

hands for carrying and provisioning; 2) the association of stone tools and bones from di-

ity of these issues and inform attempts at archaeological interpretation. These studies

verse taxa in considerable density on Leakey’s “living floors” East African

thought

transport,

have contributed to a fundamental reassessment of the archaeological evidence on which

or Isaac’s “Type C” sites at Plio-Pleistocene localities,

to indicate

and

food

off-site hunting,

Isaac’s inferences were based. Successive au-

food

consumption

thors have reinterpreted the accumulations of

(Isaac

bones and stone tools known as “Type C” sites as “central places,” to which food was

1978a, 1978b; Leakey 1971); and 3) increases in brain size, which were linked to enhanced

communication

abilities,

among

transported

so that

larly used

“depots”

it could

be consumed

close to a safe refuge (e.g., Blumenschine et al. 1996; Isaac 1983; Schick 1987), or as regu-

early hominid forms. The most closely scrutinized component of

Isaac’s model was the archaeological one.

on habitual foraging

One of the first critiques of the home base hypothesis came from Binford (1981b), who

routes (Potts 1984, 1988). In these more recent scenarios the transport of food is not

related interpretive concepts in archaeology were based on a series of unwarranted

Rose and Marshall (1996) bring together multiple lines of evidence to reassess hy-

linked to sharing.

forcefully argued that Isaac’s model and other

assumptions

and could be challenged

potheses about the evolution of food sharing.

on

They suggest that under certain ecological circumstances predation pressure and territorial defense actually promoted both the formation of larger social groups and the use

taphonomic, ethnographic, and ethnological grounds.

Citing

alternative

processes

that

could have led to the accumulation of bones and stone tools, including carnivores or flu84

Foop,

LIES, AND

PALEOANTHROPOLOGY

man

of home base—like places (Rose and Marshall

Fruth

and McGrew

1996)

near water

and

other relatively predictable resources. Transport of food to group-defensible places would have been a response to increased predation risk and to competition with both nonhominid carnivores and other hominid groups. However, they also argue that over time food sharing and social activities were increasingly associated with these locations. In support of this line of reasoning the authors cite the aforementioned primatological evidence of food sharing (particularly meat) and the habitual use of favored locations

shall 1996:317). Monogamy, sexual division

of labor, and male provisioning are not included as components of the resourcedefense model because the authors agree that the dearth of solid archaeological evidence, contradictory primatological and ethnoarchaeological evidence (Hawkes 1990; Stanford et al. 1994), and insufficient paleontological data (e.g., McHenry 1992) cannot support the presence of any of these institutions in the early Pleistocene. Regardless of what was actually going on Plio-Pleistocene,

it seems

that

is

ultimately

ary context has been reexamined. Despite the productive research it has generated, the home base hypothesis itself, and the focus on the interpretation of a few archaeological localities, also obscures a more fundamental issue. There is general agreement that early humans were living in intensely interacting groups, but since Isaac there has been surprisingly little speculation (among archaeologists) about the importance and mechanisms of cooperative behavior within these groups. This reticence to speculate extends beyond the first tool-using, meat-eating human ancestors. With some notable exceptions little voice is given to opinions about social modalities among later members of the genus Homo (however, see Farizy and David 1992; Foley and Lee 1989; Gamble 1986:3 43-393, 1993, 1996; Mithen 1996b). We know that at some point, probably later in the Pleistocene than Isaac initially believed, early humans or human ancestors came to share food and to cooperate in economic endeavors more regularly. But just when that transition occurred remains obscure at best. At least some of the reluctance to speculate is due to the scarcity of strong archaeological evidence. If the Plio-Pleistocene offers no undisputed evidence for the nature of social groups and interactions, then the later Pleistocene record is only marginally

derson and McGrew 1984; Boesch and Boesch 1994; Ferrari 1987; Garcia and Braza 1993; Perry and Rose 1994; Sept 1992b). Eventually, nonmovable resources surrounding these loci also came to be defended against other groups of hominids, and patterns of cooperative defense ultimately contributed to the use of focal places for a range of activities, including eating, sleeping, and the exchange of information (Rose and Mar-

the

studies

wider behavioral, ecological, and evolution-

among extant nonhuman primates (e.g., An-

in

evolutionary

twofold. The ability to determine the integrity and resolution of sites and, by extension, to make reliable inferences about specific early hominid activities has been strengthened by two decades of actualistic and experimental research. As noted above, an increasingly sophisticated awareness of both the discontinuous chronological resolution and the differential distribution of a range of activities across a paleolandscape has led to the realization that extant material evidence cannot easily support the kinds of fine-grained, synchronic behavioral reconstructions that paleoanthropologists once thought would shed light on ancient social relations. At the same time, the potential for placing hunting or scavenging activities in a

1996:308, 317). Increased group size and cohesion, as well as cooperative defense behaviors (alarm calls, mobbing, chasing, attacking), are typically found among primates who run a high risk of being preyed on, particularly those who live in or near open habitats (e.g., Boesch 1991; Cheney and Wrangham 1987). Like many others Rose and Marshall (1996:317-318) envision a landuse pattern based around focal points, probably trees (cf. Blumenschine et al. 1996;

the

legacy of the home base hypothesis for hu-

85

S. L.

KUHN

AND

richer. Still, as Gowlett (1996:160) has observed, the scarcity of archaeological evidence for home bases or structured use of space does not necessarily indicate that pair bonding, food sharing, or cooperative behaviors were absent among earlier hominids. It may be that we are simply unable to recognize the material correlates of such behaviors, either because our analytical approaches are insufficiently subtle or because the spatial

stance, generally based on variants of opti-

mality theory, seek to explain how external or internal conditions will affect the costs and

benefits accrued by an individual who per-

mits others to consume resources that she or

he has obtained. Models of mechanism, on the other hand, focus on how natural selec-

logical data are insufficiently precise.

tion could allow apparently “self sacrificing”

One pessimistic commentator has stated

behavioral strategies such as sharing to per-

flatly that early hominid social life is “unexcavatable” (Bahn 1990:75). In some sense we are in agreement with this view. The problem, however, is not that questions of social interaction among ancient hominids are inherently uninvestigable but that the phenomena of interest need to be rethought. The probability is vanishingly small that someone will find a “fossilized imprint” of Pleistocene social relations, much less a complete series of such cases that could be used to monitor the evolution of social relations over evolutionarily significant spans of time. Notions about changes in ancient patterns of social relations must be translated into expectations for generalized, long-term trends in behavior that are, after all, the most robust component of the Pleistocene record. The questions of most immediate relevance to Pleistocene social the-

sist and proliferate to the partial or complete exclusion of other strategies. When one is

concerned specifically with social interactions between individuals who may alter their

strategies for dealing with one another, evolutionary stable strategies (ESS) or game theory models are considered the most appropriate instruments for testing and exploring ideas about mechanism (Winterhalder 1997:123). A truly impressive range of formal models has been proposed to explain various forms of cooperation and sharing. Winterhalder discusses at least six different models of mechanism, including tolerated theft (Blurton-

Jones 1984, 1987) and its inverse—scrounging, marginal analysis (Winterhalder 1996b),

risk minimization (Cashdan 1985, Wiessner 1982), by-product cooperation or pseudoreciprocity (Mesterton-Gibbons and Dugatkin 1992), and “showing off” (Hawkes 1991, 1992). Several of these models, including tolerated theft, scrounging, and by-product coOperation, might not fit many social scientists’ definition of sharing as an intentional

orists should be (1) what is the basis of the

unique development of intensive sharing and cooperation

among

humans,

SARTHER

the development of intragroup exchange of resources, Wintethalder (1997) differentiates between “models of circumstance” and “models of mechanism.” Models of circum-

and chronological resolution of the archaeo-

economic

C.

(2)

how might that unique condition have evolved, and (3) what are the implications of these evolutionary scenarios for changes in archaeologically monitorable phenomena

act; they are, however, consistent with a more

In addressing the first question, concerning the evolution of uniquely human patterns of

generalized behavioral definition of resource transfer. In addition, a long series of game theoretic experiments (Axelrod and Dion 1988; Axelrod and Hamilton 1981; Boyd and Richerson 1992a; Nowak and Sigmund 1993) show how behaviors like sharing, which penalize an individual but benefit the group, can be successful even without group

of the formal theory regarding the evolution of sharing and cooperation. In a review of various theoretical models offered to explain

ness and diversity of the theoretical literature on altruism, cooperation, and sharing also raises a question of central importance to the

such as technology, foraging ecology, or land use? MIND

IF I HAVE AND

sharing,

SOME?

COOPERATION

COGNITION

it is instructive

to examine

some

selection (but cf. Wilson 1998). The very rich-

86

Foon,

LIES, AND

PALEOANTHROPOLOGY

argument at hand: if it is possible to model

the evolution of intragroup redistribution of resources in so many ways, why is it devel-

oped to such a high degree in humans yet comparatively poorly developed in our close

relatives? Given humans’ apparent ecological advantages, why do sharing and cooperation as general strategies seem to work better for us than for most other organisms?

A brief consideration of game theoretic models of ESS points to one possible answer.

based on more

extensive knowledge

of the

past histories of interacting individuals. Even relatively simple cooperative strategies— such as “TFT” (“tit for tat” or “do unto others as they have done unto you”)—can flour-

ish in contexts where individuals face many future interactions, recognize each other, and

recall past interactions (Axelrod and Dion 1988), all of which are typical of social or-

These formal models, often based on varia-

tions of the famous

145). It also seems that each generation of more successful strategies involves a slightly more complex response to past interactions,

“prisoner’s dilemma”

game, are used to address the question of how apparently altruistic behaviors that decrease individual fitness but benefit the group as a whole can evolve via individual selection and simple self-interested behavior. In some sense this is a classic “public goods” problem, asking how individuals who behave generously (to the benefit of the group as a

whole) can persist and thrive in interaction with selfish individuals (often referred to as

ganisms with long life spans. Ultimately, the

development of retaliatory behavior that punishes both cheaters and those who fail to

punish cheaters can allow virtually any form

of cooperation to thrive, provided that the penalty for being punished is sufficiently great and the cost to punishers is sufficiently low (Boyd and Richerson 1992a). In simple terms, the most successful “agents” in ESS

models seem to be getting more cognizant of

whom they are dealing with and better at keeping track of the tendencies of multiple individuals. Not coincidentally, a great deal of attention has also been devoted to the so-called Machiavellian intelligence hypothesis. This model challenges one of the dominant paradigms of the 1970s and early 1980s, the so-called ecological intelligence hypothesis (Dunbar 1996:5; cf. Foley 1995:166-167), which holds that higher intelligence evolved in the context of subsistence-based problem solving and reliance on patchy resources (e.g., Clutton-Brock and Harvey 1980; Gib-

free riders or cheaters) who act only to enhance their own immediate benefit. The models are usually structured so that mutual coOperation produces mutual fitness benefit, whereas cheating or selfish behavior by one of an interacting pair results in a benefit for the cheater but a penalty for the cheated cooperator. An interaction between two cheaters produces large penalties to both. A common goal is to develop strategies that allow cooperation to thrive while excluding selfish individuals or at least keeping them at a sufficiently low frequency so that they do not drive the “good guys,” and the population as a whole, to extinction. The history of game theoretical modeling and cooperative behavior seems to be one of escalating strategy complexity. No sooner is a successful form of cooperation developed than a more sophisticated way of cheating that succeeds in invading the population of cooperators appears. As it now stands one can even describe a kind of ecological succession of ESS that permits increasingly successful forms of cooperative behavior to invade and take over a population of selfish individuals (Sigmund 1993; Winterhalder 1997:

son 1986). Byrne and Whiten

(1988, 1992)

trace the development of and expand on the concept of a uniquely social intelligence. Identification of selective pressures favoring social intelligence in primates but not operating (or not operating to the same degree) in other species is a critical requirement of the hypothesis. Humphrey (1976), one of the first to address this topic, proposed that the problems inherent to group living favored individuals possessed of a special kind of intelligence that was and is still used in negotiating complex social situations. Competition over

resources such as food, mates, space, status, 4

S. L. KUHN

AND

and allies would have selected for the ability to infer the mental states of others in the group and to gather and maintain social knowledge about them.3 The ability to manipulate and deceive conspecifics confers selective advantages on individuals possessing a particularly high degree of intelligence. Byrne

and

Whiten

1992),

(1988,

de Waal

(1982), and others even present case studies in which extant nonhuman primates, particularly Pan and Papio species, engage in complex acts of tactical deception and social mafurthermore,

nipulation;

cases suggest cially learned. Humphrey (1988) believe the evolution

the

chimpanzee

that specific methods are so(1976) and Byrne and Whiten social intelligence is central to of human intelligence. Simi-

C. SARTHER

If group living, cooperation, and sharing are the ideal selective milieu for the evolution of some kinds of intelligence, then the reverse is true as well. One lesson of successive generations of ESS modeling is that the most successful social strategies are based on enhanced ability to anticipate the intentions of others and remember past actions. Perhaps, then, the relatively sophisticated development of sharing and economic cooperation among humans is underwritten in large part by humans’ unique cognitive abilities. Humans may not be the only primates capable of inferring the mental states of conspecifics. Chimpanzees appear to attribute distinct

motivations to other individuals in certain

situations (Cheney and Seyfarth 1990: 205-255). However, modern humans obvi-

ously have more elaborate ways of signaling intentions—falsely or honestly—than other

larly, Dunbar (1991, 1992, 1996; Aiello and Dunbar 1993) argues that the logistical

difficulties

of maintaining

social

organisms. It also seems likely that humans

relation-

are able to remember and keep track of the

ships within large groups led to the evolution

propensities of a larger number of individuals over a greater span of time, an obvious

of that uniquely human cognitive and behavioral trait, language. Certainly the existence

advantage in coping with large and diverse social groups.4 This is not to argue that hu-

and significance of social intelligence are sup-

ported by experiments in cognitive psychology that demonstrate that humans are partic-

ularly

adept

at

detecting

cheating

man

intelligence

renders

them

more

able

than other organisms to recognize the benefit

and

of sacrificing individual gains for the benefit of the group; all of the strategies discussed by

deception in social situations and that human reasoning abilities are most acute when the problem at hand involves social dynamics (Cosmides 1989; Gigerenzer and Hug 1992). However, these same authors contend that the social environment created by group living is “objectively more complex” than the

game theorists assume basic self interest. But

although formal models show that an organism does not have to be smart to share, the

high levels of economic cooperation among

human foragers suggest that a smarter organism may be much better at it.

ecological environment, a questionable claim (Kummer et al. 1997:158, 160-162). A more

Strategies of retaliatory behavior or pun-

ishment may be key, because they can help to maintain cooperative behavior under a variety of circumstances. Smith (1988:240) goes so far as to argue that collective goods and sharing-based economies cannot exist without mechanisms for monitoring and penalizing noncooperative individuals. Mechanisms for collecting and passing on information about past behavior, as well as for penalizing selfish individuals, are also highly developed among humans. The ethnographic literature on hunter-gatherers is full of examples of social mechanisms, ranging from gentle ridicule

moderate view is taken by Mithen (1996a) and others (e.g., Cosmides and Tooby 1992), who acknowledge the importance and early development of social intelligence but see it as one of several specialized cognitive domains. Foley (1995:171) rightly proposes that both ecological and social factors must be taken into account (see also Kummer et al. 1997): although brain enlargement might be a result of environmental or social complexity or both, large, metabolically expensive brains (cf. Wheeler and Aiello 1995) still require high-quality resources. 88

Foon,

to outright

ostracism,

credited

LIES, AND

with

PALEOANTHROPOLOGY

most basic assumptions as a basis for investigating how apparently complex phenomena such as cooperation may evolve from nothing more than self-interest. Very quickly, how-

main-

taining standards of generosity and reciprocity in human groups (Cashdan 1985; Kelly 1995:293-320 and refs. therein; Peterson 1993). Intergroup

communication

can also

ever, less-than-fundamental

effectively curtail lateral movement of cheating or free-riding individuals (Dugatkin 1992). An important dimension of many of these mechanisms is that they can be employed with relatively little cost or risk of injury to either the punisher or the individual punished, compared to actual physical retaliation. The important point here is that these low-cost strategies depend entirely on humans’ unique modes of communication. Many of the “models of circumstance” for the evolution of sharing also depend on, or are potentially enhanced by, high levels of cognitive development and symbolic communication. The most obvious example is Hawkes’s (1991, 1992) argument that the broad sharing of hunted game by males can be explained as a form of “showing off,” turning a nondefensible resource into status, individual

alliances,

mutual

goodwill,

behaviors

such

as punishment and retaliation have been integrated into the models. Although it is unlikely, indeed undesirable, that formal models will approach the complexity of real human actors operating in actual social situations,

there is always the danger that theoreticians will be diverted down conceptual cul-de-sacs, analytically interesting but empirically irrelevant lines of inquiry, as a consequence of an impoverished understanding of actual human behavior and tendencies. Furthermore, a great deal of the information about phenomena like cultural attitudes toward generosity and the ways that cooperative behavior is rewarded and reinforced is anecdotal rather than systematically collected. The most experienced students of human social interaction, social scientists, should be able to provide the knowledge needed to keep the more abstract

and

approaches

even reproductive opportunities. Obviously, the success of such strategies depends on the existence of symbolic social currencies that can somehow be obtained in exchange for gifts of meat or other resources and perhaps even “banked” in the form of prestige. Even the analysis of the marginal value theorem, a fundamentally economic approach to examining when and where an individual should give up shares of a large food package she or he is unable to utilize fully, can be affected by the existence of abstract “social currencies.” If the receiver of a gift automatically incurs some obligation to the giver in proportion to its value to her or him, then it is in the self-interest of the giver to distribute as widely as possible and to the individuals who have the greatest need (Winterhalder 1996b, 1997:128). This is one area in which traditional social-scientific approaches should have a great deal to contribute to our understanding of the evolution of sharing and economic

on course.

Egalitarianism, gen-

erosity, and cooperation are not part of the “natural” state of small-scale human societies; rather, they must be actively enforced (Kelly 1995:164-166, 296). One of many issues that could be productively researched on a cross-cultural basis is how sharing or other actions favoring the group as a whole are encouraged and regulated in small-scale egalitarian human groups. For example, is it correct to conclude that cheaters or free-riders are discouraged mainly through ridicule, public censure, and similar symbolic interactions, rather than through more direct physical confrontation? Is the tide of public opinion sufficient to keep most people in line in a small human group? A more difficult but no less important set of related questions concerns possible differences regarding how humans and nonhuman social organisms actually handle noncooperative individuals. AN EVOLUTIONARY

SCENARIO

We turn now to a brief and conspicuously coarse-grained scenario for the evolution of peculiarly human patterns of sharing and

cooperation in humans. Game theoretic ESS

models are intentionally simplistic, using the 89

S. L.

KUHN

AND

C.

SARTHER

ous to consume

economic cooperation. An important feature of this admittedly speculative model is that it does not rely ona single process or causal fac-

killed or scavenged

prey

where it was found, foraging hominids might

well have moved carcass parts to the nearest location where it was safe to consume them.

tor but on a cascade of evolutionary factors. No single model, formal or social scientific, circumstantial or mechanistic, would seem

As hominids

became

more

adept preda-

tors, and meat from large animals became a more important component of the diet, in-

adequate to take us from an imagined Pliocene hominid starting point to the tech-

tense interaction related to division of car-

casses could have led to more complex indi-

nologically sophisticated and socially elaborate societies of the Upper Paleolithic. We do

vidual strategies of cooperation or deception or both, resulting in a selective milieu favor-

not discuss the timing of the various transi-

tions in detail, nor do we consider the pos-

ing the next stages in the evolution of greater

ogy. This model is intended solely to illustrate the range and variety of evolutionary trans-

notes that group size alone is probably not sufficient to explain the unusual degree of en-

ogy, group size, and diet as a prelude to discussing archaeological criteria that might be used to monitor them. As many researchers have argued, the first steps toward intense food sharing were probably associated with hominids’ entry into a predatory niche during late Pliocene, sometime around two and a half to three million years ago. Whether these early human preda-

avellian intelligence model and that “complexity is not an automatic consequence of the number of individuals in a social group... but rather a function of the range and subtlety of the behavior expressed by its mem-

social intelligence and larger brains. Byrne

sible ramifications for human skeletal biol-

cephalization in humans based on the Machi-

formations likely to have occurred in physiol-

bers”

(Byrne

1996:174). More intensive in-

teraction connected with food sharing would have introduced new layers of complexity to social interactions, irrespective of group size.

scope of carcass consumption would have expanded rapidly. One likely “model of circumstance” is tolerated theft. Another is what has

At the same time, larger brains and larger bodies to support them could have led to changing life histories and more stable malefemale bonds (Foley 1995; Foley and Lee 1989; Kaplan 1996), a consequence of the greater costs and risks associated with providing for large-brained, slow-developing infants. This process might well have stabilized for long periods, but we know that there was directional selection for increasingly large brains and presumably more sophisticated cognitive abilities over the course of the Lower and Middle Pleistocene (for summaries of evidence, see Aiello 1996; Aiello

which alerting others to the presence of a resource allows the self-interested forager to obtain or hold onto that resource. For example, large groups of hominids probably could defend carcasses from scavenging competitors that would displace an individual forager. It is also important to emphasize that the earliest forms of food sharing need not have involved transport to a base camp or central place per se. Even if it was too danger-

ble and Davidson 1996:15 4-159). Episodes of destabilizing, directional selection could be attributed to stepwise increases in the degree of dependence on carnivory and to associated increases in the intensity of cooperative activities and food sharing within human groups. The shift toward the predatory side of the omnivorous hominid lifestyle might in turn be attributed to one or a combination of three related factors: 1) niche displacement

tors were hunters or scavengers is not as 1m-

portant as the fact that they would on occasion have had access to comparatively large packages of food in the form of animal carcasses. Food packages too large to be defended or consumed at once are likely to foster some form of sharing or at least communal feeding. Early on, we would expect sharing with offspring and closely related adults, but if larger social groups existed (e.g., for protection

been

called

against predators),

“by-product

mutualism,”

the

and Dunbar 1993; Mithen 1996a:1 1-12; No-

in

90

Foon,

LIES, AND

PALEOANTHROPOLOGY

resulting from coevolution with other predatory species (Brantingham 1998; Stiner 1994), 2) geographic expansion into arid and

tion of diet, these phenomena are not within the normal scope of paleoanthropological re-

fat-rich food in the development and mainte-

phenomena that has been most productive for investigators working with more recent

temperate zones after one million years ago or so, and 3) the need for high-quality,

nance of a large brain (Wheeler and Aiello

1995).

Most

of the

formal

models

discussed

above focus on large, difficult-to-defend food

packages. Yet hunter-gatherers routinely share even small game and gathered resources, albeit less widely (Altman and Peterson 1988; Kaplan and Hill 1985; Kelly 1995:

viduals within those groups. With the excepsearch.

One

approach

to examining

social

preliterate societies involves detailed analyses

of the arrangements of objects, features, constructions, and empty areas that gave struc-

ture to the space used by a

social unit (Petitt

1997). As argued above, detailed synchronic reconstruction will never be a strength of the

Pleistocene archaeological record. Moreover,

even if a remarkable confluence of favorable

167). In fact, the sharing of more easily obtained resources is arguably the most remark-

circumstances renders such a thing possible, we may not know what to do with the result:

among hunter-gatherers. First, these kinds of

currences may help to flesh out intuitive mod-

able

dimension

of economic

cooperation

foods are the dietary mainstays of other social primates, who nonetheless do not regularly share them. Second, such resources can

easily be defended or consumed by the collector, so simple models of tolerated theft or by-product mutualism do not apply. One possibility is that generalized strategies of punishment/retribution have simply expanded to encompass all types of resources. Such a scenario would be enhanced by central-place foraging, which makes concealment of food more difficult (Winterhalder 1997:162). Alternatively, generalized food sharing might have developed as a more economical way of establishing and maintaining interindividual bonds within large social groups (compared to grooming), perhaps as a precursor to complex verbal communication and eventually symbolic language. In this respect it is interesting that observed instances of sharing of monkey carcasses among male common chimpanzees seem to be associated with the formation of social alliances (e.g., King

1994:65-70; McGrew

Stanford 1998). SOME

ARCHAEOLOGICAL

1992:106-113;

IMPLICATIONS

Discussions of sharing, social organization, and cognitive evolution implicate a variety of processes and behavioral phenomena, ranging from diet to the sizes of ancient social groups to patterns of interaction among indi-

comparatively unique “high resolution” ocels of short-term dynamics, but they are not easily integrated into scenarios of evolution-

ary change. These cautions simply mean that

attention might be profitably redirected toward avenues of inquiry that take advantage of the strength of the Paleolithic record, particularly broad geographic patterning and long-term evolutionary trends in technology and human ecology. Steele (1996; Steele and Gamble 1998; see also Féblot-Augustins 1993 and Gamble 1996) has made the case that changes in the sizes of early hominid social groups should be paralleled by expansion or contraction of the sizes of territories occupied and, further, that

territory sizes can in turn be reconstructed through an examination of patterns of raw material transport. The underlying assumption is that, holding trophic level constant, territory size increases as a direct function of the number of individuals in a bounded social unit. Although this approach is intuitively attractive, it is not without ambiguities. The extent to which movement of raw materials is detectable is highly dependent on the sizes and locations of distinctive geological sources: for example, if raw materials are homogeneous over broad areas, it often looks archaeologically like no one carried anything anywhere. Moreover, the distances artifacts are carried depend as much on the distances between activity sites, artifact use lives, and

S. L. KUHN

AND

C. SARTHER

the propensity to plan around future contingencies as they do on actual range sizes (Kuhn 1995). If artifacts have short use lives, as appears to have been the case for most Lower Paleolithic tools, they will not necessarily travel as far as their makers. This is not to say that studies of raw material transport distances are without value but simply to point out that they are likely to result in minimum estimates of territory sizes. One of the most promising approaches to deriving expectations about hominid social organization based on common archaeological data focuses on the relationship among cognition, sociality, and technology. The

stant, changes in group size and the social transmission of technological information might well be linked to changes in group

through observation, imitation, and learning takes place within a social matrix (Boyd and

million years, with only minor alterations. In east Asia the pace of change over the course

tions must affect the broad scale dynamics of

reflected by both the rapid succession of novel artifact forms and more strongly delin-

transmission

of technological

structure and patterns of cooperation and

sharing. Rates of change in Paleolithic technologies may also be directly linked to patterns of cooperation and sharing. For the first two million years or more, the technologies of human ancestors appear by all measures to have been remarkably stable over some broad intervals of time, changing no more rapidly than the physical characteristics of the toolmakers themselves. The basic technological structures of the Oldowan and Acheulean stone

tool “industries” persisted for at least one

information

Richerson 1985; Ingold 1993a; Reynolds 1993), and the configuration of social rela-

of the Pleistocene may have been even more gradual. Rates of technological change, as

technological change. Mithen (1996b) has attempted to link assemblage variation and

eated

across

group dynamics at a large scale, arguing that

patterns space,

of synchronous

appear

to have

variation

accelerated

markedly only in the last 30,000 to 40,000 years, perhaps beginning with the appear-

larger group sizes would imply more similar learning experiences for diverse individuals

in large groups, presumably because more in-

ance of the Upper Paleolithic (Foley 1991; Isaac 1972b). This stasis in earlier technologies, although remarkable, does not reflect a funda-

pecially able technologists. He goes on to assert that the redundant and highly stable

clear that premodern hominids were capable of coping with a wide range of conditions.

and thus lead to greater standardization of

technological behaviors. Likewise, overall levels of technological skill should be higher

dividuals have access to the innovations of es-

mental

inflexibility

of behavior.

It is quite

biface forms of the Acheulean mark the first

They survived and even prospered in environments quite different from the ones in which

archaeological evidence for social learning and that differences in group sizes associated with open and closed environments during the Middle Pleistocene of England could account for the alternation between apparently “unstandardized” Clactonian and more regularized Acheulean assemblages. Of course, this model does not account for the interaction among individuals within groups of different sizes, and there are many examples of large groups of animals that do not cooperate economically. In fact, Mithen’s model assumes that the distribution of food supplies in open versus closed environments alone would explain the degree of aggregation. This implies that, holding environment con-

they had evolved, colonizing much of the land area inhabited by humans today long be-

fore the appearance

thought

to define

of that suite of traits

anatomical

modernity.

They were also more than capable of adjusting the ways they made stone tools to the diverse raw materials they encountered and to the short-term tactical demands of keeping themselves supplied with usable tools in uncertain environments (e.g. Geneste 1986; Kuhn 1995; Wengler 1990). It is the designs of implements that remain curiously static for long periods or time within the Middle and Lower Paleolithic. More specifically, it is the design of implements related to food procure-

92

Foop,

ment—especially

conspicuously

LIES, AND

weapons—that

monotonous

is

PALEOANTHROPOLOGY

most

social and material advantages to individual

in the earlier

sharers, as well as to the group as a whole.

time ranges (Kuhn and Stiner 1988). Hunting

Such conditions of open-ended demand (and

technology is an arena in which we know that artifact design has many direct functional im-

return) may ensure that there are long-term advantages for an individual to become more

where artifact design seems to have been particularly changeable during the Upper Paleo-

even when food is not especially scarce in the environment relative to absolute consumer

plications (e.g., papers in Knecht 1997) and

efficient or faster at harvesting

lithic and later time periods (Straus 1990,

demand. Clearly, the repeated development or adoption of more effective tools for getting food would be one obvious evolutionary consequence of such a situation (Kuhn and Stiner 1998). In contrast, the monotony of the earliest forms of food procurement technology—monotony crosscutting known or probable vari-

1993). This is not the place to develop a syn-

thetic overview of potential Upper Paleolithic

weapons

(see instead

Knecht

1991,

resources,

1993,

1997; Larsen Peterkin 1993; Straus 1990, 1993). However, it is safe to say that both the diversity of implements for game procurement and the variety of raw materials from which they were made within the comparatively brief span of the Upper Paleolithic greatly exceed what even the most enthusiastic observer would identify for Middle and Lower Paleolithic hunting technologies (e.g., Shea 1993, 1998). Assuming that form directly affected functional efficiency in some manner, rapid generation and adoption of novel forms of weaponry over long periods of time would suggest that there was a fairly constant selective advantage to improving the effectiveness or efficiency of foraging. Rapidly shifting environments could have led to such patterns, but we know that climate was fairly unstable through much of the Pleistocene, not just after the appearance of the Upper Paleolithic. In the case of implements devoted to food procurement, another source of constant pressure toward adopting novel, more effective or efficient technologies might relate to patterns of sharing and the sizes of consumer groups. Individual requirements for nourishment have probably not changed significantly over the last five hundred thousand years or so, but of course adult humans typically do not forage just for themselves. Among recent foragers food sharing networks are (or were) usually quite flexible, almost open-ended. No matter how much people collect or kill, there is someone to eat it, either immediately or in the future (where storage technology exists). As many of the works cited above show, broad patterns of sharing can provide both

ation in diet—may be showing us that the po-

tential payoffs for obtaining surpluses of food were not as open-ended among archaic humans as they are among modern peoples. We do not mean to suggest that earlier forms of human did not share food, or at least some

kinds of food. However, if the products of foraging were regularly distributed only within small groups—for example, a female and her children, a mated pair and offspring, or a group of close allies—then there would have been less general benefit to increasing the effectiveness of techniques for harvesting food resources in bulk so long as those resources remained at least moderately abundant in the environment. It is instructive to look beyond a strictly human/hominid context with regard to the question of sharing as a stimulus for producing and adopting novel varieties of food procurement technology. As discussed in detail above, nonhuman primates do not regularly share food except in a narrow range of social situations involving only a small percentage

of the total diet. However, for pregnant or lactating females a certain degree of sharing with offspring is obligatory. Anecdotal information suggests that many “advances” in primate foraging, including those involving the use of tools or techniques for processing food, originate with or are most readily and widely adopted by females (e.g., Boesch and Boesch 1981, 1984; McGrew 1992:88—106), the individuals for whom the reproductive 23

S. L. KUHN

AND

resources. Experimental and ethnoarchaeological research should be directed toward developing criteria for recognizing whether carcasses were butchered for distribution or consumed gradually (e.g., O’Connell 1997), although we are not optimistic about the resolution or general applicability of methods developed in this way. Where sites have a high degree of spatial integrity, the technique of

benefits of sharing are greatest and most direct. This evidence supports the contention that expanding consumer demand through sharing may place a greater premium on foraging efficiency by individuals, providing conditions ideal for the spread of new techniques of food acquisition. Moreover, these observations have interesting implications for the sources of such innovations as do ap-

refitting bones

pear in the early record. Another relevant dimension of the technofacilities,

as

well

as

tools

carcasses

(Enloe

1992;

dow on the fate of a large, nondefensible

package. As a result of the high-profile de-

invested in artifacts of all types, including feaand

and

Enloe and David 1992) provides a direct win-

logical record is the amount of time and labor tures

C. SARTHER

bates over hunting and scavenging among early hominids, taphonomists and zooarch-

and

weapons. The production of elaborate technological aids to foraging or other work carloading” of effort. On one hand this requires

aeologists have tended to focus on the largebodied, impressive game species. As has been argued here, however, sharing “defensible”

high degree of cooperation and coordination of activities among members of a social

spective. As plant remains are seldom preserved in Pleistocene sites, addressing this

ries with it a significant amount of “front-

resources that come in small packages is even more remarkable from an evolutionary per-

a certain degree of foresight on the part of toolmakers. More important, it requires a

question argues for closer attention to small

group. If an individual is to invest significant

game and especially “gatherable” (sessile or slow-moving) species. Of particular interest is the transport of small package resources to central places for consumption. Although small game have been ignored by many Pleistocene specialists, some pioneering archaeological (Stiner 1993a; Stiner et al. 1999; Villaverde et al. 1996) and ethnoarchacological

amounts of time and labor in the production of elaborate items of technology, the potential benefits from which might not be realized for days, weeks, or even years, that individual

must be free to divert this time and labor from more immediately pressing concerns such as getting food or shelter. If someone is able to devote many hours or even days to the manu-

(Hudson 1991) research has been carried out.

facture of artifacts, then someone else must

Some of the archaeological criteria described here—especially rates of change in hunting tools and the degree of investment in technology—may suggest that full-blown patterns of food sharing evolved fairly recently, perhaps even after the appearance of the Upper Paleolithic. On the other hand, criteria such as artifact standardization and perhaps changes in territory size may point to important changes occurring much earlier, well back in the Pleistocene. These results are not necessarily contradictory. As the hypothetical model outlined in the previous section implies, there is no reason to assume that the widespread sharing and economic cooperation characteristic of modern huntergatherers developed as a consequence of a single process or that all aspects of hominid

be carrying at least part of the load with respect to collecting and processing other resources, as well as sharing the fruits of their foraging. Actualistic (experimental) assessments of the time and energy needed to produce the range of known artifact forms at particular intervals in human history may reveal how much time individual toolmakers were able to divert from more immediate re-

quirements. Sharp increases in the amount of

time and energy invested in artifact production may in turn signal important changes in the economic “safety nets” afforded by the social networks of ancient humans and human ancestors. Of course, the faunal record also has the potential to provide both direct and indirect evidence for sharing and redistribution of 94

Foopb,

LIES, AND

PALEOANTHROPOLOGY

sociality—from group size to patterns of in-

Paleoanthropologists

teraction to pathways of information flow— evolved in perfect unison. Evolutionary changes almost certainly occurred in fits and starts, with a perceptible lack of synchronization among the social, ecological, and cognitive domains. Moreover, there were probably precursors to forms of behaviors we would recognize today as sharing, such that tolerated theft of large, indefensible resources came first, eventually to be supplanted by more generally reciprocal economic relations involving a variety of resources. Likewise, food transport might have occurred first as a means of seeking safety or securing some processing advantage (Rose and Marshal 1996; Stiner 1993b) and only later have become integrated into a broader complex of behavior related to resource sharing at a central place. Finally, it is important to emphasize that none of the criteria we have suggested for monitoring changes over time in the sizes of and degree of economic cooperation within hominid social groups would by themselves

have long been taken

with the notion that the appearance of sharing-based economies, and_ especially male provisioning of dependent offspring, marks an important, perhaps central, step in hominization. Almost three decades of research on these issues have resulted in important advances in archaeological methods and a vastly more sophisticated understanding of the formation of long-term archaeological records. Somewhat ironically, this same research has produced considerable pessimism concerning the potential for learning about hominid social relations in the remote periods of the Pleistocene. One limitation of past investigations of Pleistocene hominid social-

ity based on archaeological evidence is that

phenomena such as the evolution of cooperative economies and male provisioning are treated as historical events rather than as the outcomes of complex evolutionary processes. A wide range of formal models from evolutionary biology that show how different types of cooperation might evolve under particular circumstances provide one set of tools for unpacking the complexity of what is referred to as sharing. Many of these models in turn invoke causal mechanisms and external circumstances that may have more robust archaeological consequences than does the cooperative behavior itself. A constant concern with any formal model is that it is more than a metaphor, that it replicates important dynamics of the real world in some necessarily simplified way. In keeping with the theme of building bridges introduced at the beginning of this volume, we argue that this issue creates potentially fertile ground for mutual exchange between formalized economic and evolutionary approaches to cooperation on one hand and more traditional social-scientific approaches to the study of human action on the other. Key to the enterprise is an understanding of the general mechanisms by which cooperative relationships are established, monitored, and maintained in human societies. After all, with the possible exception of certain bonds formed during infancy, relationships between people are remarkably fluid and flexible.

permit one to reconstruct social organization

at any particular interval in the past. Looking for “inflection points,” instantaneous evidence for change in trends in artifact standardization, rates of technological innova-

tion, investment in tool manufacture, or sizes

of group ranges is a fundamentally imprecise endeavor (Gowlett 1997). Such an approach may help identify differences between more or less prolonged intervals of time, but it will never permit one to pinpoint the moment that a shift occurred. On the other hand, if the goal is to monitor and explain the evolution of patterns of economic and social interaction peculiar to modern humans, then it is less crucial that we identify the specific social configuration that obtained at any particular interval in the evolutionary past. Social relations may well be unexcavatable, but evidence for evolutionary changes in them is everywhere. CONCLUSION

Human societies involve levels of economic cooperation and exchange of resources unparalleled among closely related organisms. 95

S. L. KUHN

AND 2.

Moreover, natural selection does not act on

the relationships themselves but on the mechanisms by which relationships are created and reaffirmed—or at least the cognitive and of cooperation

(1992a).

(Premack 1988), and “higher order intention-

to com-

ality” (Dennett 1983): “I know that you know

plement the many theories of power and coercion in circulation today. 1.

The resultant literature is notoriously vast. For more in-depth discussion of these research programs and extensive bibliographies see, for example, Rose and Marshall (1996) and Sept An ability variously known as “mind reading” (Byrne and Whiten 1991), “theory of mind”

biological underpinnings of those mechanisms. In essence, what we need are social-

scientific theories

C. SARTHER

that I know...”

An alternative evolutionary trajectory, followed by the eusocial insects, is to genetically program social roles. Under these circum-

Notes See Gowlett (1996:155-157) for a thorough consideration of the intellectual context of

stances, an individual’s

“motives”

are deter-

mined by his or her genetic and phenotypic identity and are thus easily apprehended on contact.

Isaac’s model.

96

7

On What People Make of Places A Behavioral Cartography Maria

The management

of American

NIEVES

ZEDENO

Indian cul-

confront the potential impact their research

opment of social theory may appear at first the oddest of bedfellows. Yet for the last century an enormous body of governmentfunded research has been devoted to integrating the prehistory, ethnohistory, and ethnography of aboriginal land and resource use practices for political and economic reasons. In the products of this research one can track the development of social theory, its impact on the political history of land management in North America, and, interactively, the stimulus that anthropology and archaeology have received from public policy. In the products of this research one can also uncover fallacy, stereotype, contradiction, and painfully gained intellectual growth. At atime when the proliferation of themes, theoretical schools, and political stands has contributed to atomize anthropological fields

spective I will reflect on a topic of current concern in social theory: how people build social environments in the process of “appro-

tural resources on public lands and the devel-

may have on people’s lives. From this per-

priating nature” (after Ingold 1986).

This essay attempts to show how anthropological notions of American Indian land and resource use have been tailored to fit political demands at different times and how anthropologists have repeatedly redressed theoretical and methodological approaches to aboriginal human-nature relations both to respond to and influence the outcome of those demands. Recently, this highly interactive process has furthered the development of various trends in the conceptualization of

human-nature relations; this paper explores and elaborates on one such trend—the landscape approach.

and alienate academia, research devoted to

HUMAN-NATURE RELATIONS IN BEHAVIORAL PERSPECTIVE

the preservation of American Indian cultural resources furnishes a stimulating arena for examining the interplay of public policy and social theory. In applied research the intimate relationship between politics and anthropology takes on a material quality that scholarly discourse alone cannot invoke. I suggest that applied research plays an important role in the development of social theory because it opens new areas of inquiry, provides a laboratory for testing the commensurability of theory and practice, and forces scientists to

In the broadest sense I am concerned with understanding how people build social environments through interactions with nature. The concept of “built environment” is generally associated with the modification of the earth’s surface by means of constructing facilities—houses, streets, public plazas, temples. A great deal of research on cultural geography and landscape architecture is devoted to exploring ways the built environment constrains or enhances social interaction (e.g., 97

MarRiA

N.

ZEDENO

slightly different angle, to explore the sequences of interactions that (a) transform a place or a localized resource into a category of material culture, which is called here landmark and (b) link single landmarks into an integrated network or landscape. To develop this outline two basic concepts must first be decoupled: place and space. Of the plethora of approaches available in the literature of place and space—some as ancient as the Greek philosophers (Casey 1996)—I follow the more tangible usage of both concepts in the history of land management in the United States and briefly review its relationship to anthropology’s conceptualization of Ameri-

Cosgrove 1984; Jackson 1984; Norton 1989; Roberts 1996; Tuan 1977). Numerous stud-

ies of prehistoric and historic architecture tackle social issues from a variety of processual and postprocessual perspectives (e.g., Bender 1993; Hirsch and O’Hanlon 1995; Lipe and Hegmon 1989; McGuire and Schiffer 1983; Pearson and Richards 1994; Tilley 1994). In social theory the well-known work of Foucault (1977) specifically addresses, through an analysis of the panopticon, the use of buildings as social tools. Giddens’s (1984) structuration theory articulates the position that architecture arranges social interaction in recurrent patterns and thus gives

can Indian land tenure, past and present.

form to social structure. In at least three recent works Foucault’s and Giddens’s theories of social space(s) are cogently implemented in

Specifically, two landmark interventions of professional anthropologists serve to contextualize, historically, contemporary tives of place and space.

archaeological research to explain power relations in prehistoric state formation (Nielsen

perspec-

The landscape approach is then presented

1995; Smith 1996) and changes in social structure in intermediate societies (Ferguson 1996). Although expressing different theoretical perspectives, these studies share the

and its links to both social theory and the

politics of land and resource management

are discussed. Last, a behavioral alternative to the conceptualization and application of landscape in archaeology and ethnography

notion that social space is the product of in-

teractions between people and the material world and is thus amenable to archaeological inquiry. The conventional study of architecture provides a solid conceptual and methodological framework for addressing diversity and change in social behavior. This framework, however, is not directly applicable to social groups who, rather than modifying the natural landscape in any radical or readily ob-

is outlined.

The

approach

favored

here,

called “a behavioral cartography,” employs

the conceptual tools afforded by behavioral archaeology in the “mapping” of places and connections among

places. As material cul-

ture, places have properties, performance characteristics, and life histories that must be brought to light before we can understand how social environments are built from nature.

servable way, built their social environment

around the extraction and appropriation of localized natural resources—plants, animals, minerals, and landforms. I suggest that a less confining battery of conceptual and methodological tools is needed to incorporate such

TENURE, AND

PROPERTY,

BOUNDED

SPACE

The idea of “bounded space,” as conceived in geopolitical maps, implies territoriality or ownership of a portion of the earth’s surface and everything that lies within its boundaries (Tuan 1977:4). In the United States bounded spaces—states, counties, cities, public parks, reservations, and private estates—define and

built environments into a general theory of

social space(s). This essay investigates the landscape approach as an alternative view to the construction of social environments. “On What People Make of Places” attempts to outline some of the processes through which places and resources are incorporated into social life or, viewed from a

constrain

the interactions

between

citizens

and the resources contained in the national territory. Exclusive property over land and 98

On

WHAT

PEOPLE

resources is a basic right actively enforced since the adoption of the Constitution in 1787 (Rose 1994:71). European and EuroAmerican exercise of these rights over land

for when setting the land survey system was

tenures are intimately tied to American polit-

ical history. This tie can be traced to the founding fathers’ adoption of Enlightment philosopher John Locke’s notions of natural law and property (Williamson 1990:246). Thomas Jefferson’s “agrarian dream” was filled with visions of a wilderness tamed by the yeoman, the independent, prosperous farmer he had seen in Europe. The future of the country rested, in his view, on the development of a robust agrarian economy at the landowners

(Gates

gistic

into parcels of 160 acres (Worster 1994:12).

in

many

instances

to a relative degree, some semblance of their aboriginal land and resource use practices. But the Dawes Act of 1887, which authorized the U.S. president to allot reservation lands to individual Indians and gave the government discretionary power to open surplus reservation land to Western colonization (McDon-

nell r991:2; Otis 1973), sought to change drastically and forcibly aboriginal tenure systems—it promoted permanent settlement and the adoption of Western farming systems (Leupp 1909). One of the most detrimental effects of the Dawes Act was the imposition of a “space-bound,” nuclear, and intensive

and, almost a century later, the Homestead

Act (1862), settlers owned parcels of land. In practice, however, parcels did not materialize the introduction

were

cies, American Indians were able to maintain,

Settlers looking for a plot to farm would be referred to the grid, regardless of whether there was water or arable land in the parcel chosen. Nonetheless, the system was kept in place and served to establish property claims over portions of the national territory. According to the 1785 Land Ordinance

later, with

territories

shared by more than one tribe. Although tribes did have loosely bounded territories and enforced use rights, effective tenures were not contiguous, homogeneous, clearly bounded spaces (White 1991; Wishart 1985). As a result settlers encroached on what they perceived to be unoccupied lands (see Cass 1975; Schoolcraft 1852-1857), eventually leading to conflict, war, land cessions, relocation, and the creation of Indian reservations. Until the institution of assimilation poli-

1973:

12). But the taming of the wilderness was not an easy task; lands had to be explored and made accessible to Euro-American settlers. In 1785 the Land Ordinance established a national land survey—the Township and Range System—intended to transfer public lands to the private domain. The ordinance divided the country from the Appalachians to the Pacific Ocean into a rigid grid of square parcels, one mile on the side, each subdivided

until much

PLACES

that the wilderness was not an uninhabited world but the home of native groups whose conceptualization of land did not fit the notion of bounded space as a discrete geopolitical entity. Seasonal and logistic mobility, practiced by numerous tribes but most prevalent among big-game hunters, proved to be a resource use strategy completely incompatible with Euro-American settlement. Ownership claims to and exclusive use of water, surface, and underground resources within the parcels contrasted sharply with aboriginal strategies that favored extensive use of large territories, each containing numerous placespecific tenures over intensive exploitation of nuclear areas (Zedefio 1997). These vast lo-

233-286). Yet the systematic partition of the continent into “ownable” polygons dates to the first opening of the Western wilderness to settlement. Exclusive property rights, agrarian development, and the notion of space-bound

of private

OF

neighbors, or resort to physical aggression— and their neighbors were, for the most part, Indians. What the founding fathers and, later, the federal government did not account

and resources has occurred since the establishment of the colonies (Williams 1990:

hands

MAKE

of

barbed-wire fencing. Thus, to exercise their rights settlers had to maintain effective use of the land, honor verbal agreements with their

99

MaRIiAN.

tenure system that was in fundamental contradiction with the “place-bound,” extensive, and mobile system that hunter-gatherer and some farmer groups had maintained until the partition of the reservations (Zedeno 1997:71; Zedeno et al. 1997:124). But in the

eyes of the law there were savages with too

much unused land and settlers who needed land desperately and who were also needed to civilize the countryside, so the land was redis-

tributed accordingly.

The provisions of the Dawes Act were reversed with the passage of the Indian Reorganization Act in 1934; this Indian “New Deal” of the 1930s promoted self-governance, cultural pluralism, and tribal—rather than individual—land tenure. Yet it could not eradicate the fundamental contradictions that the Western space-bound notions of land tenure had created.

Early Anthropology and Indian Land Tenure The only well-known and widely acknowledged role of anthropology in the politics of land management dates to the assimilation period, when scholars hired by the Bureau of American Ethnology (BAE) or by other government and private institutions set about to

preserve the lifeways of the vanishing Indians by gathering as much information on their traditions as possible. During that period L.H. Morgan’s social evolutionary framework became the official worldview of the budding American anthropology (Hinsley 1981). In an interactive process government-funded anthropologists, attempting to preserve vanishing Indian lifeways, fed social evolutionary constructs to the government. Policy makers in turn acted on that framework to institute a space-bound notion of land tenure intended to destroy what anthropologists so keenly sought to preserve (Hoxie 1984). Nowhere is this interaction between nineteenth-century anthropologists and land

ZEDENO

West into civilization a new system of land tenure had to be instituted, even if traditional Indian practices were to be eradicated or the Indians removed (Powell 1878:18). Wittingly

or not, nineteenth-century scholars transferred to government officials what they had learned was the key to successful assimilation: by divesting people of their freedom to organize themselves about the land in the traditional way, they disarmed their social compass or the ability to ground their social inter-

actions in place.

THE NOTION

As early as acknowledge tween Indian land use and

OF PLACE-BOUND

TENURES

1905 anthropologists began the grotesque differences and Euro-American notions the lack of understanding of

to beof di-

versity in Indian tenure systems (Henshaw 1905:106). Yet it is difficult to sort out from

the period’s writings what the Indians considered aboriginal, what anthropologists ob-

served was aboriginal, and what the aborig-

ines adopted in an effort to fit better into their new country. Most likely this blurring of information results from the active involvement of anthropologists in government affairs, particularly Indian Affairs (e.g., Powell and Ingalls 1874; compare Kelly 1964:26; Stoffle and Evans 1976:184). It is equally challenging to sort out reconstructions of Indian tenure systems from the BAE documents because they are tainted with efforts to fit them into L. H. Morgan’s social evolutionary scheme, thus subsuming diversity under an all-inclusive classificatory system based on Western principles of land tenure. There are, however, a number of individual

reports and privately published monographs that, because of their strong historical emphasis and attention to ethnographic detail, constitute invaluable sources of information on aboriginal practices (e.g., Fewkes 1900; Fletcher 1885; Mallery 1894; Mindeleff 1989; Stephen

1936; for archaeological use

policy officials clearer than in Powell’s most famous political piece, Report on Lands of

of BAE sources see Cameron 1999; Walker 1998; papers in Zedefnio 1999).

which he explicitly stated that to bring the

took

The Arid Region of the United States, in

The bulk of ethnographic research that

100

place

between

the

establishment

of

On

WHAT

PEOPLE

MAKE

OF PLACES

the reservations and the reorganization of American Indian tenure and governance in 1934 did not address the effect of changing

to award compensation for land losses. At the opposite extreme anthropologists built improbable arguments about the sociopolitical

Indian groups. Sadly, invaluable information on persistence and change that could have

effort to gain for them the right to compensation (e.g., Dobyns and Euler 1970; compare Manners 1974). Underlying these debates were traces of the old Social Darwinism that had permeated the intellectual and political

tenure systems on the social organization of

organization of hunter-gatherer tribes in an

been obtained from this time period, when

native societies were struggling to accommodate the newly imposed system while attempting

to

preserve

what

they

could

of

understanding

their traditional practices, was lost. With the widespread adoption of ahistorical functionalism, cultural relativism, and the search for “the primitive,” this information was buried even further. It was not until 1946, when the

of American

Indian

society

since the late nineteenth century. But between extreme positions expert witnesses produced

some of the most brilliant research on land

and resource use ever written, as, for ex-

ample,

Schroeder

(1974)

on

the

Yavapai,

Aschmann (1974) on the Western Apache, Wishart (1985) on the Pawnee, Stewart (1966) on the Shoshone, and Hickerson

Indian Claims Commission was established,

that information on pre- and postreservation land tenure systems was brought to

(1974) on the Great Lakes Chippewa.

light. It is seldom acknowledged, much less

taught in the classroom, that for the better part of four decades ethnographers, historians, and archaeologists devoted a great deal of effort to gathering information on Indian land tenure practices either to support or deny the right of tribal groups to receive compensation based on the extent and content of past land use (U.S. Congress 1946). As either tribal or government expert witnesses, these scholars faced an_ ethical dilemma: on the one hand, archaeological and ethnohistoric research indicated that prereservation land tenure systems rarely involved nuclear, intensive, mutually exclusive territorial exploitation of localized resources, carried from time immemorial—a prerequisite for receiving compensation for lost land. On the other hand, by denying that a mutually exclusive “space-bound” tenure system existed, expert witnesses prevented the tribes from being compensated for their losses (Washburn 1985:27). In a process that pitted colleague against colleague and student against teacher, anthropologists engaged in bitter debates about American Indian traditional lifeways. Julian Steward, for example, fought the Land Claims Commission for its fallacious premises but never came forward with a positive alternative for establishing valid criteria

The extent to which the land claims research affected modern anthropological theory and how deeply the old stereotypes persisted or changed through this research, remains to be determined. I suspect there is an intimate relationship between the 1940s-to1950s expansion of cultural ecology and multilineal evolutionism in American anthropology and the research on land and resource use. For example, Alfred Kroeber’s “conversion” from Boasian particularist to cultural ecologist resulted from his involvement on tribal claims in California (Beals 1985:142). Many other anthropologists embraced cultural ecology and used the vast lands claims database on land and resource use to argue for the dependency of human subsistence on environmental factors, particularly in the

West

(e.g.,

Beals

and

Hester

1956,

1958;

Kroeber 1963). To a great extent the research focus on subsistence was rooted in the premises set by the land claims process; other elements, such as ritual-based land use and land-based social relations, were unevenly documented and seldom brought to bear on legal decisions regarding the nature and extent of land and resource use, especially of hunter-gatherers. I suggest that the emphasis on ecology and adaptation that characterized anthropological and archaeological research

IOI

MarRIiaN.

until fairly recently may have resulted to an

ZEDENO

1993).

As

Kelly

(1998)

suggests,

we

must

extent from this bias. This research continued for the most part to reify the space-bound notion of American Indian land and resource use that had ruled over decisions made by the U.S. Claims Court in earlier years.

seek new and productive ways to do archaeology given the current political climate, and I hope to contribute a few ideas toward this goal. Currently, American Indian land and re-

Current Legislation and Its Effect on Research Trends Since the passage of the National Historical

of convergence between land management politics and social theory. Central to the process of complying with preservation policy is the reassessment of land and resource

can Indian cultural resources has opened new Opportunities to revisit old research tenets regarding aboriginal land and resource use. Yet it is difficult to estimate, without the perspective afforded by the passage of time, the exact relationship between current concerns for

tal difference between land claims and current compliance research; land and resource use is not conceived only as subsistence economy but as a practice that extends to all realms of social life that, therefore, cannot be explained solely through a functional-

States and other countries where aboriginal populations thrive, and the adoption of ideas posed by contemporary anthropologists, particularly postprocessual archaeologists. In the American Southwest, for example, the re-

transcends political boundaries to incorporate the broadest possible range of land and resource uses and emphasizes the preeminence of place over space. One of the most intriguing but hardly unexpected twists— given the present intellectual climate in anthropology—in compliance research is the appearance of eclectic and postmodern approaches to human-nature relations. This trend is evident in the application of a landscape approach to the reconstruction or recreation, if you will, of aboriginal land tenure systems. Following is a brief summary of the intellectual roots of this approach and its applications.

source use issues are once again at the point

Preservation Act (NHPA) of 1966, a body of federal legislation aiming to preserve Ameri-

preserving

use practices. However, there is a fundamen-

aboriginal culture in the United

ecological

turn of migration and the sudden interest in

ritual and symbol coincide at least timewise with American Indian discourse on cultural

preservation and religious freedom.

Archaeological involvement in the development of public policy that directly affects our research endeavors has never before been explicitly acknowledged as it is today. These critical changes in theory and practice are placing new demands that we must strive to meet, without surrendering commitment to analytical standards

(Brumfiel

1992). But a

framework.

UNPACKING

historical evaluation of the effects of contemporary American Indian legislation on archaeological thought and vice versa will have to await at least another generation or two of scholars. Nonetheless, it is painfully obvious that the current legislation will affect profoundly the way we do archaeology. Although some traditionalists see in these changes the 666 of Armageddon (e.g., Clark 1996), those of us immersed in compliancedriven research find instead a golden field of untapped possibilities for theoretical and methodological advance (e.g., Dongoske et al. 1997; Mills and Ferguson 1998; Spector

THE

This

research

LANDSCAPE

also

APPROACH

There are numerous ways to describe a landscape approach to land and resource use. At

its most basic, landscape provides a frame of

reference for understanding relations at various scales. It work of nineteenth-century phers, such as Alexander von has

humanistic

ancestry

human-nature is rooted in the human geograHumboldt, and

(Cosgrove

1984;

Smith 1996). The landscape concept was originally introduced to American geography by Carl Sauer (1925). Yet modern landscape research is highly diversified, incorporating elements of geography, ecology, biology, history, architecture, and anthropology (Jack102

On

WHat

PEOPLE

son 1984; Jarvis 1998; Norton 1989). These

intellectual sources, in turn, address spatial,

historical, biological, psychological, social, and symbolic aspects of human-nature

re-

lations (Whittlesey 1998:18). In theory each source should offer a complementary view of

these relations; in practice divergent episte-

mological principles upheld by scholars (e.g., materialism vs. phenomenology) often ren-

der such views incompatible.

Interest in the landscape approach waxed and waned in geography and ecology for at

least 50 years since Sauer’s landmark work. During the 1960s and 1970s landscape studies fell from grace mainly because landscapes were not easily delimited or measured; the concept “landscape,” as understood in clas-

sic human geography, had strong visual, aesthetic, and even moral connotations that defied the boundaries of the scientific method (Allen and Hoekstra 1992:47; Cosgrove 1984:9). Quantitative research that focused on spatial analysis of natural and human ecosystems was favored instead for its more precise and objective results (Soja 1971; but see J. B. Jackson’s magazine, Landscape). This quantitative approach to spatial organization had a strong influence on processual archaeology (e.g., Clarke 1972; Hodder and Orton 1976).

The landscape approach returned to human geography partly as a result of Tuan’s (1977) qualitative study of place and space, Cosgrove’s (1984) sociohistorical research, and Jackson’s (1984) analysis of the built environment. Landscape studies, particularly those addressing the dwellings of ordinary people, have since gained popularity in geography and other disciplines as well (Whittlesey

1998:18).

Contemporary

geographers

MAKE

OF

influence of geography and architecture on archaeological landscape studies is most

prominent in historical archaeology and pub-

lic history (e.g., Deetz 1990; Ingold 1993b; Yamin and Metheny 1996). It further appears in legislation specifically written to docu-

ment,

preserve,

and

interpret

historical

scenery in urban and rural America (Copps

1995; Low 1987; McClelland et al. 1990). Landscape ecology has contributed a bio-

logical perspective on human-nature relations. In functional ecology landscape is

used to explain the spatial structure of contiguous resources; it is useful as an analytical unit because of its independence from prede-

termined scales (Allen and Hoekstra 1992). In evolutionary ecology, on the other hand,

historical relationships between humans and resources take primacy over spatial structures (Levins and Lewontin 1985). The ecological-evolutionary landscape perspective is represented in regulatory legislation applied to public parks and wilderness preserves, where the role of human-nature relations on environmental history and diversity is seen as an integral component of ecosys-

tem management Moote 1999:20).

strategies

(Cortner

and

A different historical view of landscape is found in historical materialism. Traditional Marxism conceived the natural world as a stage on which human history was enacted (Lansing 1991:10). A more recent formulation of landscape in neo-Marxism states that landscapes are determined by the interplay of forces of production, relations of production, historically contingent structures and processes of nature, and human cognition and valuing of natural structures and

processes

stress that cultural, socioeconomic, and political forces transform nature; the resulting hu-

PLACES

(Patterson

1994:57).

In this view

and at the same time acts as a mnemonic device people use to recall past interactions and to inform and legitimize their present re-

societies with different modes of production leave different landscape signatures, thus restating the social evolutionist notion of human-nature relations: humans _progressively increase control over nature as the complexity of their power relations increases

scape architects take this perspective further into an analysis of the visual impact of the built environments on social behavior. The

tionary and Marxist views of landscape). In their critique of environmental and economic determinism, postmodern social

man landscape organizes spatial interaction

lations

(Basso

1996;

Norton

1989).

Land-

(see Whittlesey

103

1998 for a review of evolu-

MarRIA

N.

theorists have offered an alternative land-

scape definition: landscape is a socially constructed environment, wherein humans shape

ZEDENO

ceptive, metaphoric, commonsensical meaning that makes passerby, bureaucrat, scientist,

postmodernist,

and

environmentalist

and to the natural world. Radically different from classic and modern Marxism, ecology,

alike forget for a moment all their differences and nod their heads in agreement. Landscape, as contained in the visual field of the

landscape perspectives see nature not as a world apart from humans but as a socially

understanding, a superficial image, if you will, of the surrounding world. And there lies

and

reshape

their

relations

to

each

other

geography, and anthropology, postmodern and meaningfully

constituted world

(Jarvis

1998). Anthropological applications of the

average individual, evokes a cross-cultural

the danger of applying it uncritically. Anyone can impose his or her own “image” over a

postmodern landscape approach are far from homogeneous, however; they range from co-

people’s land and resource use practices and dictate from that perspective how and why

(Bradley 1993; Smith 1996; Thomas 1996) to exuberant phenomenological recreations of

systematically

gent

analysis

of

the

social

environment

those resources were, are, or ought to be related to each other, without explicitly and identifying

the

criteria

that

past human-nature relations (Crumley and

guided the formulation of that image—after

Despite the potential that the landscape

jective. And, of course, American Indian land

Marquardt 1987; Tilley 1994).

all, perception and metaphor are entirely sub-

concept has for describing and explaining human-nature relations, there has been little

and resource use studies have not evaded this trend.

et al. 1997:126). Perhaps because of the dis-

Landscape in American Indian Land and Resource Use Studies

concerted effort to systematize its application in regional field-based studies (Zedefio

parate theoretical sources, definitions and applications of landscape vary wildly, and as

Throughout the first two decades following the passage of the National Historic Preser-

chaeologists have appropriated landscape to analyze settlement patterns with newly available technology but have yet to thoroughly address the theoretical implications of the concept. At the opposite end, landscape has been used by cultural anthropologists and archaeologists idiosyncratically and metaphorically, serving only as a means for grossly categorizing social, political, and ideological relationships between people and the environment. “Landscapes of power,” “landscapes of imprisonment,” and “engendered landscapes” are but a few examples of the landscape metaphor (Baker and Biger 1992; Bender 1993; Hirsch and O’Hanlon 1995). In fact, one of the standing criticisms of these landscape studies is that they do not outline the sufficient and necessary conditions that define a landscape (Lawrence-

legally and intellectually defined as material culture—artifacts, features, archaeological sites, crafts—and state and federal laws and agency regulations were thus geared toward the preservation and management of these

Whittlesey

(1998:19)

Zuniga 1996:916).

notes,

processual

ar-

The state of contemporary landscape literature at large underscores the main weakness of the approach as applied thus far: its per-

vation Act

(1966),

cultural resources

were

materials. However, amendments to the 1966

act and to the American Indian Religious Freedom Act of 1978, among other pertinent legislation passed by Congress and by the executive office since 1980, require consultation with American Indian groups who once inhabited lands now managed by the state and federal governments. Ensuing interactions among agencies, tribes, and anthropologists have greatly expanded the definition of cultural resources to incorporate a far broader set of natural and cultural materials,

features, and places. It is in this context that the landscape approach was introduced into American Indian cultural preservation programs (National Park Service 1994). One of the first steps toward expanding the notion of cultural resources in field stud104

On

WHAT

PEOPLE

MAKE

ies entailed the integration of ethnographic, archaeological, and ethnohistorical research in the evaluation of traditional cultural prop-

OF

PLACES

as mandated by the legislation (Evans and Roberts 1999). This framework must (1) in-

tegrate local and professional understandings of land and resource use in a single coherent package, (2) render American Indian views

erties or TCPs. TCP is a term introduced by

the National Register Bulletin 38 (Parker and King c. 1990) to designate places of cultural

intelligible to the public land manager, and (3) temper romantic and deceptive notions of “Indian land” with a more realistic and historically accurate depiction of traditional and contemporary native knowledge and practice. Given the complex history of Indian-

and religious significance for American Indians that may or may not have associated ma-

terial culture (e.g., a landform, a stand of old

growth). It quickly became clear, however, that these properties could not be evaluated

as discrete units but must be evaluated as part

government struggle for control over land

that ecological models used by anthropolo-

formation

of a much larger whole (Kelley and Francis 1994; Stoffle et al. 1997). It was clear, too,

and resources, the political and scientific importance of systematically retrieving the inneeded

to implement

legislation

gists and archaeologists to analyze regional systems were inadequate for capturing the

on American Indian cultural resource preservation cannot be underestimated.

whole. While applied researchers were pondering the practical challenges of identifying landscapes in the field, a new body of literature on landscapes appeared; driven either by policy compliance or ecological research, this literature targeted explicitly American Indian, native, or aboriginal landscapes. “Sacred ge-

scape approach to describe and explain human-nature relations, and _ specifically American Indian land and resource use issues, requires that we first define landscape materially by systematically describing its

formal, spatial, and historical content of that

ography”

(Walker

1991),

“spiritual

geo-

graphy” (Griffith 1992), “sacred landscapes” (Carmichael et al. 1994), and “symbolic landscapes” (Grieder and Garkovich 1994) are but a few of the terms currently employed to describe the relationships among Indians, land, and resources. Approaches range from scholarly essay to magazine editorial, and from ethnographically/archaeologically informed monograph to Sunday newspaper supplement. This wave of trendy publications—some romantic, some insightful—recapture the importance of place-bound approaches to American Indian land and resource use (e.g., Feld and Basso 1996). Unfortunately, many of them combine aspects of geography, anthropology, ecology, and postmodern social theory in a politically and ecologically correct, but conceptually eclectic, framework. It has been left to the applied researcher to build a theoretically informed and methodologically sound framework for identifying and evaluating American Indian landscapes

In my view the application of the land-

material components and identifying rela-

tionships among them. This view is not derived from an epiphenomenal understanding of the social or the symbolic, nor is it fashioned to accommodate the limitations of the archaeological record. Simply, a material approach to landscape is rooted in the conviction that all realms of human-nature relations, from cognitive mapping of the surroundings (Austin 1994:61) to permanent modification of the earth’s surface, to symbolic representation of power, are ultimately mediated by and derive from concrete interactions between people and the material world. This conviction constitutes the basis for rethinking the landscape concept from a behavioral perspective. A BEHAVIORAL

CARTOGRAPHY

To construct a behavioral landscape framework, I begin by briefly describing a passerby’s view of landscape. When we look into the horizon, we see an arrangement of objects. We can describe the kinds of things we see on the basis of their formal properties and estimate the spatial relationships (as in distance) among them. In other words we can 105

MarRiAN.

“round up” a set of objects contained in a space arbitrarily determined by our visual field. Furthermore, we can develop an opinion, perception, or feeling about the particular arrangement of objects after we have de-

termined what is inside that field. But to understand why objects are where they are, and why people move from one object to another, we must abandon the boundaries of the geometrical frame imposed by our visual field and focus on the location, characteristics, and order of each object relative to all others. There are two fundamentally different ways for describing and explaining a landscape: (a) delimiting an arbitrary space and then focusing on what is inside that space or (b) focusing on one specific object and describing, progressively, its relations with other objects. The

first, or space-bound

notion,

is exem-

ZLEDENO

tems, it is necessary to understand first the organizational relationships among places that were differentially used during the operation of past systems. This frequently cited but seldom fully applied idea (but see Ebert 1992) pointed rightly at the need to rethink space-

bound approaches to the study of humannature relations. Unfortunately, an outmoded functional and systemic emphasis separates Binford’s place-bound approach from social and symbolic landscape studies (Whittlesey 1998:19).

There

is, however,

an important

element in Binford’s work that fits well with the notion that people relate first and foremost to places: that places are not only external, passive, and fixed resource locations but also a human product. In principle, a place may be defined as the

discrete locus of human behavior. Places be-

plified in spatial studies of human-nature relations in modern geography, functional

come a material culture category by virtue of transformation through human action. This

regions, areas, communities, and single settlements. Perhaps the most serious drawback

mark”

ecology, and in archaeological definitions of of space-bound notions is that with the passage of time the arbitrary character of space 1s eventually forgotten, thus leading to the construction of essentialist typological arguments about the content of space. The second, or place-bound notion, appears in human ecological analysis of contextually

specific resource use (Vayda 1983) and in postmodern analyses of the primacy of place over space in human-nature relations (Casey 1996; Smith 1996). Distantly related as they may seem, both

contextual ecological analysis and postmodern analysis converge at one crucial point. For humans space is not a tabula rasa that is progressively filled with locations as they accumulate perceptions and _ experiences

through interactions but vice versa. Place is

what defines people’s relations to each other and to the material world—the proverbial widening of horizons starts at one place and extends from there. In archaeology the earliest application

of a place-bound

notion

to

land and resource use is found in L. Binford’s (1982)

“The Archaeology

of Place,” where

he stated that to understand past cultural sys-

material category is what I have called “landin an earlier work

(Zedefo

et al.

1997:125). Landmarks are locational markers that indicate places where interactions

and activities occurred and may include stationary and physically unmodified features of the natural landscape, such as rock formations, tree stands, and waterholes, or features of human manufacture, such as buildings, trails, burial mounds, and petroglyphs. What

makes a landmark significant from a behav-

ioral standpoint is that it constitutes a piece

of history of land and resource use practices. The process of transformation of a place

into a landmark may involve a series of activities and interactions that crosscut various

realms of social life, from subsistence to rit-

ual, and accumulate through time. Thus, a landmark has a life history that develops from multiple experiences lived at a particular

place.

During

fieldwork

with

western

Shoshone and southern Paiute elders in Nevada (Zedefnio et al. 1998), for example, my colleagues and I came across landmarks that have complex life histories: a landmark’s life history may begin as a place where a memorable encounter between hunter and prey occurred; and, through repeated use of the place and recalling of the encounter, it 106

On

WHAT

PEOPLE

MAKE

may become an arrow-making workshop, a initiation rites, and a vision questing site. A

landmark may also be a “juncture” at which place and resource are drawn together by human interaction. To illustrate, a caribou kill

can be characterized as having three basic di-

mensions: (1) formal: the physical character-

istics of landmarks; (2) relational: the interac-

site is a landmark where the mobile hunter and the mobile prey converged, regardless of the route each took to or from there. Through time, landmarks may signify multiple junctures, depending on the number and nature of activities in which they participated. Although acknowledging the fact that there may be landmarks bearing single interaction referents (or what we used to call “limited activity sites,” “isolated artifacts,”

tive links (economic, social, ritual) that through the movement of people connect landmarks to one another; and (3) historical:

the sequential links that result from successive uses of places (Zedeno et al. 1997:126).

In an interactive process landscapes are not only a product of human behavior, but they also define and constrain human behavior. The most eloquent portrayal of landscape as a network of places connected by interactions and activities may be found in map-making cross-culturally. Ojibwa bark maps, for example, are not only locational or spatial descriptions of geographic features but also graphic accounts of all sorts of activities carried out at different places along a particular

“other,” or “unknown”), the reconstruction

tarian”

artifact (e.g., a metate,

a pot, or a

pick-up truck) not only transforms the artifact into a ritual or a symbolic object but also transforms the place of discard into a ritual landmark (Walker 1998). Landmarks may be viewed as “pages” in the history of land and resource use, but they alone do not represent the full spectrum of land and resource use interactions and activities. Landscapes, on the other hand, contain the spatial, historical, and social dimensions of human-nature relations. Landscape may be defined as the web of interactions between people and landmarks. Through multiple interactions among people and between people and resources, landmarks become progres-

PLACES

sively linked to one another, forming an aggregate. This aggregate, however, is not simply a sum but a network. As such, landscape

locus of propitiatory ceremonies, a place for

of transformation processes expands the knowledge on landmark variability to include multiple interaction loci, whose contextual information would be lost in a functional typology. Mutually exclusive functional typologies, whether extracted from informants (after much verbal wrestling) or devised by the anthropologist, rarely convey the historical depth and contextual wealth encoded in behavioral transformations of place. As pointed out by S. Mithen (1990) in Thoughtful Foragers, people may move fluidly from one realm of social life to another. This fluidity is not limited to places but also involves other material culture categories, including portable objects. For example, the act of discarding ritually a “utili-

OF

trail (Densmore

1979:181).

Similarly, map-

like petroglyphs in southern Nevada were used historically by Paiute and Shoshones to point out connections between landmarks located at some distance from one another and that could not be captured in a single visual

field (Zedefio et al. 1998:60).

Landscapes may not be bounded, but they are finite. A landscape extends only as far as people’s experience, gained through direct interaction with landmarks and resources, can reach. However, beyond the landscape experienced firsthand there is yet another realm of experience that is acquired indirectly, through interactions with other people. Humans depend to a certain extent on knowledge of possibly available places and resources beyond their interactive network. What is significant about knowledge of an outward network or “mythical landscape,” however imperfect this knowledge may be, is that it informs a series of land use behavlors such as migration decisions, maintenance of property rights over distant places

and resources, or exercise of knowledge- and access-based power relations (Myers 1991).

Knowledge of distant places and resources may be passed on intergenerationally and 107

MarRiAN.

refer to ancestors who originated there. Among historical and contemporary American Indians experiential and mythical landscapes are closely tied by the political history of land and resource use; mythical landscapes

generally refer to those lands and resources

located in aboriginal territories that were lost

to Euro-American colonization. Most of these lands and resources are no longer acces-

sible to direct use but remain present in oral

history.

bow, and extracts a shaped piece of wood, leaving a permanent scar on the trunk. Thus the bow tree is a stationary natural resource used in the manufacture of hunting bows, an archaeological feature insofar as it has been

permanently modified by the bow maker, and a landmark in that it contains informa-

tion ona specific interaction or activity. Place

transformations resulting from a given inter-

action may become a place’s properties for subsequent interactions; the bow tree may be

sought after as a place suitable for manufac-

THE MAPPING OF LANDMARKS LANDSCAPES

In conventional

ZEDENO

archaeological

AND

fashion one

may be tempted to construct predictive mod-

els to explain variability in the nature and

content of formal and relational properties of landscapes. But because these landscapes are

created through transformational processes, their specific or contextual characteristics only come forward through the reconstruction of sequences of activities and interactions that led to the integration of multiple human and natural elements. Behavioral archaeology provides three useful concepts for mapping landmarks and landscapes: performance characteristics, life histories, and formation processes. Properties and Performance Characteristics A landmark study has two basic components: (a) identifying a place’s forma! properties and performance characteristics in order to evaluate its potential for attracting people to carry out activities there and (b) reconstructing the sequences of interactions and activities that transformed a place into a landmark. Properties may be divided into natural and cultural. Natural properties include topography and presence of one or more resources. Cultural properties refer to place modifications through human use. Both natural and cultural properties may combine in a single polythetic unit; such is the case of the “bow tree.” A bow maker

peels off the tree bark, outlines the bow

shape directly on the tree trunk, carves sets of parallel notches along the length of the

turing

hunting

tools,

initiating

a

young

hunter, depositing propitiatory offerings, or

simply stopping to rest on the way to another place.

Performance

characteristics are interac-

tion-specific capabilities that make a place uniquely suited for certain activities (adapted

from Schiffer and Skibo 1997:30). Location, direction, view, acoustics, distance, and ex-

posure to the elements are examples in point. Performance characteristics generally correspond to properties—topography may enhance or diminish a place’s visual or acoustic performance. A behavioral cartography rests on the premise that the selection and use of a place is driven by performance in the same sense that the design and execution of an artifact or feature is driven by performance (Schiffer and Skibo 1997:29). Certain performance characteristics may have more weight than others in the place selection process, thus driving the users to make compromises to obtain trade-offs. For example, the selection of a place suitable for building a hunting blind among a set of topographically similar places may be guided by the hunter’s need to be downwind from the prey, even if visual or acoustic performance diminishes with this choice. Or an individual may sacrifice proximity when isolation is needed for a specific activity, such as doctoring. The ability to balance compromise and trade-off to meet performance requirements depends on available information (technological, social, political, ritual, mythical) about a place and, more generally, accumulated knowledge about land and resource use. Thus, fine-tuning the placeselection process derives from both firsthand

108

On

WHaT

PEOPLE

MAKE

experience and ability to act on relevant in-

OF

PLACES

become the focus of social activities such as

formation, however limited it may be.

seasonal gatherings, festivals, and life-cycle rites; these landmarks accrue significance as they get used time after time, eventually becoming well-known place/resource markers and spatial/historical reference points (e.g., Steward 1938:98). In contrast, there are landmarks characterized by seclusion and are often marked as such; knowledge and use of

Life Histories It quickly becomes clear that, although iden-

tifying and evaluating properties and perfor-

mance characteristics of landmarks bearing

single interaction referents may be relatively straightforward, the difficulty of undertaking such a task on multiple-referent landmarks could increase proportionally to complexity

these landmarks may be restricted to certain

individuals or sectors of the society, and they

and variability of use histories. At this point the life history approach used in behavioral

may be used only once or twice in an individual’s lifetime. For example, during vision

defined here as the cycle of formation, use,

used to paint natural orifices on rocks with

archaeology comes to bear. Life history is

questing journeys, Shoshone warrior hunters

and transformation of a landmark (adapted from Schiffer 1987:13). Elsewhere I have ex-

black paint and stare at the painted holes for days until the vision came. To this day a

amined and illustrated at length the applica-

Shoshone woman will hesitate to approach

Here,

(Zedeno et al. 1998:111-112). Similarly, first menses initiation sites are gender-specific

tion of this approach to reconstructing the formation of territorial units (Zedefio 1997). therefore,

I briefly elaborate

on

any black-painted hole and its surroundings

the

concept. The reconstruction of a landmark’s life history begins with the isolation of evidence for an activity or interaction, given specific properties and performance characteristics, and subsequently traces the historical progression of interactions that occurred there. The success of this task rests on finding evidence (material, archival, oral) of transformation resulting from subsequent interactions, much in the way the archaeologist would reconstruct a structure’s stratigraphic sequence from foundation to abandonment and disintegration or reuse. Life history reconstructions are based on the premise that a given interaction or activity may modify existing properties and performance characteristics of a landmark at any point in its use history. As noted above, people who have the flexibility to move back and forth from one realm of social life to another tend to modify landmarks repeatedly. These behaviors result in complex landmark life histories that defy simple functional models of land and resource use. Certain landmarks—particularly those that are highly visible, like rock formations, or that contain critical resources, like springs—tend to attract attention and often

landmarks to which males have little or no access (Arnold et al. 1997). Both types of

landmarks indicate group- and individual-

oriented interactions, respectively. These examples also illustrate the importance of evaluating both stable and changing place properties and performance characteristics to further understand behavioral variability encoded in landmarks’ life histories. Mapping Landscapes A contextual characterization of landscape begins at a specific landmark, and progressively reconstructs formal, relational, and historical links with other landmarks. The departing criteria for the selection of the first landmark may be formal or relational. The former criterion focuses on natural and cultural properties and follows with an examination of potential interactions, whereas the latter first isolates a given interaction or activity (e.g., plant collecting) and follows with an evaluation of corresponding properties and performance characteristics. The choice of criteria depends to a great measure on the questions being asked about land and resource use and on the nature of the material or archival records, but both are valid and should produce equivalent results. 109

MarRIAN.

Progressively linking landmarks can be done formally or relationally; the focus varies from physical and morphological to resource- and activity-specific, respectively. Regardless of the focus or criterion chosen, the crucial strategy is to expand the links in all possible or feasible directions from the first landmark. The rationale behind this strategy is counterintuitive; whereas the average passerby’s idea of a landscape is visually confined to a horizon-like continuum, the actual or effective use pattern of a landscape often extends well beyond that continuum and is asymmetrical, punctuated, and

sometimes logic-defying (Western logic, that

is). Therefore, a systematic investigation of relevant place properties and a mapping of associated resources may be needed before links become apparent. Another critical as-

pect to consider when reconstructing formal and relational links is that landmarks may

not

always

constitute

readily

identifiable

and classifiable “sites.” Landmarks

left by

hunter-gatherer groups are often ambiguous and difficult to interpret in strictly functional

terms. For example, the “rock alignment” 1s

ZEDENO

homes [Zedeno 1997]). It follows that a land-

scape may never contain strictly contemporaneous landmarks at any point in its use history because people tend to cumulatively link old and new landmarks as they experience them directly or indirectly. Perhaps the greatest challenge in reconstructing landscape life histories archaeologically lies in acknowledging that people link archaeological sites into their systemic landscapes and in developing appropriate strategies for mapping these land use behaviors. An approach to formation processes that is not limited to taphonomical and geoarchaeological reconstructions of skeletal landscapes but that takes into account the dynamic interplay of human behavior and natural processes may be the key for addressing this challenge. It is at this juncture that archaeology, ethnohistory, and contemporary ethnography

may be integrated to provide complementary perspectives on the social, historical, and material dimensions of land and resource use. There is a clear need to expand or modify conventional perspectives of what constitutes a site, a settlement, and a built environment

West that has ambiguous functional referents and therefore is often underplayed in inter-

(e.g., Rossignol and Wandsnider 1992; Smith 1996; Teltser 1995a; Whittlesey 1998) and to refine identification methods accordingly.

alignments may be linked with other landmarks, such as trails, rock art, springs, game

ethnographic and historical information on land and resource use in the design and imple-

a relatively common feature in the American

pretations of the built environment; yet rock

routes, or lookout points. It is only after examining

the

web

landmarks—drafting

of potentially a

matrix

of

linked

specific

properties, performance characteristics, and activities that link one landmark to another (see Schiffer and Skibo 1997:20)—that one may begin to outline the logic behind a particular landscape.

Landscapes, too, have life histories that il-

lustrate variability in transformational processes and changes in landmark-resource networks. Landscape life histories often extend over hundreds of years and contain landmarks of varying age and complexity; links between such landmarks are also complex and may vary from actual reuse of places to incorporation into oral history (e.g., persistent places [Schlanger 1992] or ancestral

Refinement

may

include

incorporating

mentation of field surveys (e.g., Lange 1998).

Another important aspect of piecing together

information on landscapes is the application of adequate methods and techniques. Avail-

able tools such as GIS do help make a cumbersome task simpler by allowing the integration of layered databases on landmarks and resources and helping visualize the logic of landscape networks. A more theoretically informed and parsimonious use of technology is critical for succeeding in this endeavor. SUMMARY

I set out to write this essay with two broad goals in mind: increase our knowledge of the intricate relationships between building social theory and practicing anthropology, and demonstrate how applied research can furIIo

On

WHAT

PEOPLE

MAKE

nish a useful context for refining archaeological theory and method. This goal does not constitute a one-way road but an interactive

sources—plants,

cal relationships are explicitly addressed. In this specific case social theoretical perspectives on human-nature relations can profit from compliance-driven research on cultural The brief historical overview of American Indian land management and its political and intellectual ties with anthropology contextualizes, in practice, divergent views of the social construction of place versus space. These two abstract concepts, which have vexed philosophers and theorists since ancient times, have historically defined and polarized our understanding of what constitutes effective use of land and resources and how that use shapes social relations. Anthropologists have been devoted, for the most part, to the study of ancient and extant non-Western societies but have applied an essentially West-

their formal, relational, and historical links, it also feeds on other information sources,

concepts of land and resource use is the first

step to bridge this difference. Despite early efforts by archaeologists such as Binford (1982) and more recent attempts at redefining archaeological units of land-use analysis (e.g., Ebert 1992; Rossignol and Wandsnider 1992), the application of place-bound concepts to the study of humanland interactions is still in its infancy. The time is ripe, therefore, to develop broader and more fitting frameworks for explaining the social and material dimensions of the built environment than the ones we have produced thus far. Landscape provides a framework for explaining human-nature relations that conventional analyses of the built environment, generally embedded in space-bound notions of architectural design, do not offer. Landinformation

on how

and

of landscape but grounds them in place. This approach seeks to gain a deeper understanding of the relationships between concrete human-nature interactions and the built environment by first isolating their material manifestations. Although behavioral archaeology provides useful concepts for mapping these interactions and decoding

and present. As Casey (1996:15) “For the anthropologist, Space for the native, Place; and the difno means trivial.” Revising basic

conveys

minerals,

a priori perceptive or cognitive dimensions

ern perspective to explain human-nature re-

scape

animals,

landforms. Gender-based or ritual-based social relations, for example, could have developed from appropriative activities; these in turn might have led to differential access to certain places and resources. A focus on “place-making,” as Basso (1996) would say, allows one a more complete understanding of these and other processes through which people construct their social world. In outlining a behavioral cartography | have attempted to incorporate the concept of place into an archaeological conceptualization of landscape. A behavioral cartography is primarily materialist; it does not dismiss

resource preservation and vice versa.

past out, first; is by

PLACES

shaped, in turn, define and constrain further interactions. Traditional mobile societies built their social environments around the extraction and appropriation of places and re-

exercise wherein both theory and practice can mutually benefit insofar as their histori-

lations points comes ference

OF

such as history and ethnography, to develop relevant material correlates of land and resource use. For a behavioral archaeologist every context evidencing interactions and activities involving humans and the material world presents a profitable opportunity for theoretical and methodological advance; current compliance-driven research offers such an opportunity. 1.

Notes A detailed review of the legal, political, and practical complexities of modern American Indian tenure systems that derived from the 1887 and 1934 acts is beyond the scope of this paper; for further reference please consult Castile and Bee (1992); Coulter and Tullberg (1984); (1991);

people

Hauptman (1984); McDonnell McGuire (1988); Sutton (1991);

Wilkinson (1987).

build environments through complex interactions with nature and how the networks so Itt

8

Strange Attractors Feminist Theory, Nonlinear Systems Theory, and Their Implications for Archaeological Theory SUZANNE

M. SPENCER- WOOD

This chapter is concerned with the bridging of two very different theories increasingly evidenced in archaeology: feminist theory and

nonlinear systems theory. These theories represent two schools of thought that cannot be merged or synthesized into any single grand theory. Neither theory can be subsumed by the other. Each has a distinct worldview and is concerned with different subjects, issues, questions, and methods. Nonetheless, a few bridges do connect these models. The ways that each model theorizes historical relationships, as well as other complementary similarities and differences, provide foundations for dialogue and possibly even exchanges of conceptual principles or method. In addition, feminist theory and nonlinear systems theory have links to other archaeological theories. As a means of stimulating more detailed construction of bridges among these various models, this chapter points to areas of particular promise for future bridge builders. On superficial examination feminist theory and nonlinear systems theory appear very different, possibly even opposed to each other. Nonlinear systems theory descends from the same scientific systems theory that was applied in archaeology as a cultural model by Clark (1964) and as processual cul-

ture theory by Clarke (1978). In contrast, feminist theory has critiqued the androcentric biases in science and archaeological systems theory models (e.g., Code 1991; Spencer-Wood

1992:101-103; Wylie 1991).

However, this essay shows how at deeper levels these theories are in some ways similar

and related revolutions in epistemology and worldview. Schiffer (1996) has demonstrated the

value of comparing different theories that share some common ground. Without minimizing the many fundamental differences between feminist theory and nonlinear systems theory, comparing them shows how these theories resonate with complementary themes and how they support each other in important

ways.

Comparison _ provides

deeper understanding of both theories. Furthermore, they are in some ways historically related. Similarities in the worldviews of these theories suggest that each might in some ways be enhanced by complementary aspects of the other. This paper explores which aspects of each theory might be useful to the other. My research in feminist theory since the late 1960s and in nonlinear systems theory since 1983 has made me aware of complementary aspects of these theories. My training in and teaching of the “new” archaeology (1968-1987) and subsequent research in postprocessual archaeology has further led me to analyze relationships between these theoretical approaches and feminist theory and nonlinear systems theory. This explo-

ration of historical relationships and concep-

tual bridges among theories follows previous papers emphasizing the distinctions between T12

STRANGE

ATTRACTORS

mainstream archaeological theories and feminist theory (Conkey and Spector 1984; Nelson 1997; Spencer-Wood 19914, 1992, 1993;

Wylie

1991)

and

between

mainstream

complementary aspects of feminist theory. The fifth section suggests aspects of nonlinear systems theory that might be useful in feminist theory. The sixth section explores implications of feminist theory and nonlinear systems theory for other archaeological theories. The last section explores connections among nonlinear systems theory, feminist theory, and other archaeological theories.

ar-

chaeological theories and nonlinear systems theory (Spencer-Wood 1989a, 1989b, 1990, 1996).

An inclusive feminist approach permitted me to see connections between feminist theory and nonlinear systems theory. An inclusive feminist approach uses “and/and thinking” to find connections obscured by binary either/or thinking, which views theories in opposition to each other. In inclusive feminist theory, for example, black/white dichotomies that have been constructed within and between theories are perceived as continua that include all the shades of gray between black and white (Spencer-Wood 1993:129-130,

FEMINIST THEORY

Feminist theory has been revolutionary in revealing how a complex self-reinforcing structure of androcentrism permeates West-

ern culture, including science and archaeological theory and method. Feminist theory critiques and corrects biases created by a male-centered view that valorizes everything associated with men and devalues everything associated with women (Spencer-Wood

1999a:166-167). This chapter explores connections between apparently unrelated theo-

ries within a larger epistemological continuum. A shared Western cultural foundation provides a basis for finding historical relationships between feminist theory and nonlinear systems theory. At the most general level of similarity both nonlinear systems theory and feminist theory are broad multidisciplinary theories representing revolutionary shifts in worldview. Both have been applied to archaeology after having been applied in a number of other disciplines. Both have critiqued traditional scientific worldviews and have created more inclusive epistemologies. Further, both theories are concerned with questions and processes relevant at a variety of societal levels, crosscutting the usual segregation of social theory by societal level (see Schiffer’s introduction). These general similarities make it possible to further compare these theories in the sections below. The first two sections briefly summarize postmodern feminist theory and nonlinear systems theory. The third section explores bridges between nonlinear systems theory and feminist theory, including historical relationships, similar critiques, and worldviews. The fourth section suggests possible expansions of nonlinear systems theory based on

1992:98).

Androcentrism culturally constructs men as public, dominant, powerful, active, cultural, rational, and orderly. Women are constructed as domestic, subordinate, powerless, passive, natural, irrational or emotional, and

chaotic. Men are viewed as powerful social agents and the makers of history. Women are devalued as only domestic and therefore insignificant to history (Spencer- Wood 1992:

99).

Feminists have challenged claims of the universal validity of such culturally constructed gender stereotypes by revealing previously overlooked evidence that women have acted as powerful social agents and makers of history through both their public and domestic roles (Casey et al. 1998; Claas-

sen 1991; du Cros and Smith 1993; Gero and Conkey 1991; McEwan 1991; Scott 1994; Spencer-Wood 19914, 1991b, 1992, 19944, 1999a; Starbuck 1994; Sweely 1999; Walde and Willows 1991; Wright 1997; Yentsch

1991). Feminist research has shown that the

women in many cultures who held a variety of public roles, including queens, chiefs, warriors, traders, and priests, cannot be dismissed as insignificant exceptions (e.g., Nelson 1997:131-149; Spencer-Wood 19914;

Troccoli 1999). 113

SUZANNE

M.

Feminists have further critiqued the construction of cultures in widely assumed un-

gendered structural dichotomies that are de-

rived from historically constructed gender dichotomies (Spencer-Wood 1993:129-131). For instance, the widely assumed opposition between culture and nature is often not

recognized

as a gender-based _ historically

situated dichotomy. Some historical archaeologists have researched the Georgian mindset—in which ordered culture dominates and rationalizes chaotic nature—without taking into account the underlying gender ideology of the dominance of cultural, intellectual man over natural, emotional woman (e.g., Leone 1988:236,

242).

Deetz

(1988)

constructed

the historical evolution of Anglo-American culture in sets of structural dichotomies that progress from the chaos of multicolored houses and natural earth-toned pottery to orderly white houses and white ceramics expressing cultural dominance over nature. Deetz did not recognize the gender and racial stereotypes underlying the ideology valorizing the dominance of cultural orderly whiteness over earlier chaotic multiple colors and natural earth tones. At a deeper epistemological level Deetz (1988) justified his dualistic reconstruction of the past by the structuralist statement that binary thinking is universal and genetically programmed. The feminist critique of gender dichotomies has been extended to the underlying male bias in binary either/or structuralist thinking (Spencer-Wood 1993:129-130, 1999a:165). In contrast, postmodern femi-

nists have developed inclusive and/and thinking as nonwhite, working-class, and gay feminists have brought to light the diversity of women’s voices, viewpoints, identities, ideologies, roles, behaviors, and relationships. Postmodern feminist theory is concerned with gaining insights into social complexities stemming from multiple intersections among gender, race, ethnicity, class, religion, and other social dimensions. Postmodern feminist theory has grown beyond the earlier white middle-class essentialist characterizations of women and the Marxist-feminist definition of women as a homogeneous class

SPENCER-

WOOD

in Opposition

to men

as a class (Donovan

1985; Nicholson 1990; Spencer-Wood 1992: 105-106). In an early feminist article Conkey and Spector (1984) expanded the critique of gender stereotypes to consider the related processual emphasis on male-associated large-scale external (public) causes of culture change and the neglect of female-associated small-scale

internal (domestic) variations in behavior. Wylie (1991) critiqued Binford’s (1983a) contention that large-scale cultural institu-

tions completely determine individual behavior and his denial of the importance of smallscale internal cultural dimensions such as gender and ethnicity. Further, Wylie showed that gender cannot be subsumed within the larger social subsystem of the systems theory model of culture. Although one early thread of processual archaeology sought to connect variation in the actions of individuals with artifact style variations (e.g., Hill and Gunn 1977), Gero (1985) and Nelson (1997: 49-50) have pointed out how archaeology continues to assign higher status to maleassociated large-scale constructions of the past and to devalue female-associated smallscale research. Feminists have critiqued the androcentric exclusion of women

as drivers of the major

large-scale process of hominid evolution, which has been constructed as solely the result of men’s important hunting or scavenging roles. Feminists have shown the importance of women’s roles in hominid evolution, including child rearing, gathering, and

participation in hunting either directly or through providing environmental knowledge (Nelson 1997:70-73; Spencer-Wood 1999a:165). Feminists have also critiqued the male-biased overemphasis on competition and survival of the fittest as the sole force of evolution. Feminists have pointed out that the female-associated principle of cooperation in groups is necessary for human survival (Gross and Averill 1983; Tanner 1981:3, 6, 23). A major concern of feminists has been the

theorizing

of

gendered

power

dynamics.

Postmodern feminists have proposed a vari114

STRANGE

ATTRACTORS

ety of inclusive models to go beyond hierar-

predicting large-scale orderly patterns and processes from external causes, dismissing internal systemic small-scale chaotic or random

chical conceptions of “power over” (Shanks

and Tilley 1992:129-130) and the dialectical dichotomy of domination and resistance.

variation as irrelevant “noise.” Third, the Newtonian scientific paradigm mechanistically considers phenomena as if they were

Much feminist research has been concerned with the important issues of women’s resis-

tance to male dominance and cultural constraints on women’s power. More recently Ehrenreich et al. (1995) have proposed a more expanded conception of a heterarchy of powers, including not only hierarchical powers but also unranked lateral forms of power. Such “powers with” include noncoercive processes such as cooperation, affilia-

tion, dialogue,

inspiration, persuasion,

clocks that could be understood

among their parts. It is assumed that complex “noisy” phenomena at some point also could be reduced to linear relationships among their orderly parts (Briggs and Peat 1989; Stewart 1989).

In the paradigm of Newtonian classical determinism, the unique evolutionary trajectory of a system under given conditions could be prescribed for all time by linear differential equations, which do not consider any random external input. Deterministic science has been concerned with linear relationships, such as correlations and regressions. In linear relationships large-scale causes produce large-scale effects, whereas small-scale perturbations have only small, insignificant effects. Experiments and observations are thought to be precisely repeatable because the effect of small measurement errors is assumed to be small. The traditional scientific goal is to identify the linear relationships among large-scale causes of orderly recurring patterns. Linear differential equations involving rates of change have been used to describe simple relationships such as the elliptical periodic movement of two bodies in space

em-

powerment, and negotiation (Spencer-Wood 1994a:191-198, 1999b:179).

Feminist theory has replaced structuralist thinking in dichotomies with more inclusive models. For instance, Kent (1999) has replaced the opposition of egalitarian and hierarchical societies with a continuum model of the range of complex gendered power dynamics in societies between and including these two extremes. Other culturally constructed oppositions that can be viewed as continua include masculine vs. feminine gender identity, black vs. white races, and objectivity vs. subjectivity (Spencer-Wood 1993: 129-131). At the level of epistemology, Wylie (1992) has taken an inclusive approach to move beyond the binary either/or argument about whether data are theory laden. Further, feminist theorists have expanded beyond the culturally constructed opposition between objectivity and subjectivity by proposing alternative models that relate these forms of reasoning through dialogue, merging, or in a continuum (Code 1991; Spencer-Wood 1993:129-131; Wylie 1989). NONLINEAR

SYSTEMS

(e.g., the orbit of the earth around the sun). It

is assumed that more complex “noisy” patterns and processes could also eventually be described by more elaborate linear differential equations. However, attempts to use linear differential equations to describe rela-

tionships in the movements

in space among

three bodies have been unsuccessful (Briggs and Peat 1989:21-23; Ruelle 1991:26-31;

THEORY

Nonlinear systems theory has challenged the following assumptions and tenets of the dominant Newtonian scientific paradigm. First, the scientific goal of extracting orderly regularities and ahistorical laws from nature is based on the fundamentally flawed Western dichotomy of order vs. chaos. Second, traditional science has been concerned with

and pre-

dicted from analyzing regular relationships

Stewart 1989:9-15, 66-72).

Nonlinear systems theory is creating a revolution within the paradigm of classical determinism. Fundamentally, nonlinear systems theory has shown that unpredictable random variation can be produced in a deterministic system. This contradictory statement 115

SUZANNE

M.

is the most succinct definition of the scientific

meaning of the term chaos in nonlinear systems theory. In this definition chaos is not simply random irrelevant noise. Instead, nonlinear systems theory has shown that apparently random variation is inextricably connected to orderly recurring patterns in many natural and cultural patterns and processes. This holistic theoretical standpoint has shown that order and disorder are both essential interrelated parts of real-world pat-

terns and processes. Chaos can be generated

from orderly patterns and vice versa. Different combinations of irregular variation and

SPENCER-WOOD

rately enough for an experiment to be pre-

cisely repeated, overturning the Newtonian clockwork worldview that the whole is the sum of its parts. A wide variety of complex natural and cultural processes and patterns have been more realistically modeled with nonlinear differential equations than is possible with linear dif-

ferential equations. Nonlinear processes 1in-

clude the weather, heart rhythms, kidney functioning, epidemics, predator-prey popu-

lation relationships, population growth and

migration, the growth of cities and political movements, the outbreak of war, the growth

of related system states, from equilibrium through instability to chaos. Analyzing only

of settlement, and some decision-making processes and economic patterns, including the stock market (Allen 1988; Allen and San-

plete understanding and to partial approxi-

and Peat 1989:57, 138; Chen 1988; Dechert

orderly recurring patterns form a spectrum

orderly recurring patterns leads to incom-

mations of reality. Irregular and random variations must be included to more realistically

model processes of change (Stewart 1989:17, 22).

Nonlinear

particularly

systems theory is concerned

with

fluctuating

systems

and

irregular growth processes that combine or-

derly patterns and random variation. Nonlinear systems are “sensitive to initial conditions.” This means that internal small-scale irregular or random variations can lead relatively quickly to large-scale changes. Nonlinear differential equations are used to describe interactions between a system’s given starting

state, small-scale variations called “initial conditions,” and systemic responses involving complex feedback processes. Although the starting state of the system involves necessary factors permitting change, those factors are not sufficient to cause change without specific initial conditions. Nonlinear differential equations address changes in external variables only through their effects or constraints on internal system variables (Stewart

1989:17, 268). In systems “sensitive to initial

conditions” a small variation is iterated through interrelated positive and negative feedbacks, generating motion so complex that it appears random (Briggs and Peat 1989:24; Stewart 1989:16, 21). Further, the initial conditions cannot be measured accu-

glier 1981; Aron and Schwartz 1984; Briggs

1996; Glance and Huberman 1997; Glass 1997; Guastello 1995; Jensen et al. 1985; Mandelbrot 1982; Milton and Belair 1990;

Mosekilde et al. 1988, 1989; Pool 1989a, 1989b, 1989c; Rasmussen et al. 1985; Reiner

et al. 1988; Saperstein 1984; Saperstein and Mayer-Kress 1988; Schaffer and Kot 1986; Sterman 1989; Sturis and Mosekilde 1988).

Most relevant to archaeology are models of the growth of political movements and regional and urban development (e.g., Pumain 1997). Of particular interest are models of self-organization processes in the growth of market areas and centers from the interaction between random occurrence of entrepreneurs and average dynamics of supply and demand (Allen 1997:50—-53; Sanglier and Allen 1989). Prior to nonlinear systems theory, LeBlanc (1969) used historical data to demonstrate the importance of individual idiosyncratic behavior in the location and spread of early textile mills that were related to the development of urban areas in New England. The use of nonlinear differential equations in archaeology has created more realistic models and simulations of past complex patterns and processes. However, the patterns generated by equations do not yet completely match available data (Rosen 1997). Zubrow’s (1984) early nonlinear models of the growth of a Mayan settlement and chipping

116

STRANGE

patterns on stone tools appear generally realistic but were not tested against data. Models generated by nonlinear differential equations do not precisely match the archaeological

data for the geographical spread of a historically documented epidemic (Zubrow 1997:

250) or for discontinuous urban evolution in

the late La Tene (van der Leeuw and McGlade

1997:366—-367). In other cases there are in-

sufficient data to test computer-generated nonlinear models (e.g., Mithen 1997:212).

ATTRACTORS

gists have found that natural populations grow and decline within self-regulated limit cycles that can be modeled with nonlinear differential equations. In studies of many species the mechanism found for self-regula-

tion of stable population size is a rising or falling birthrate as a feedback response to population

density

(Augros

and

Stanciu

1987:124-129; Briggs and Peat 1989:37—-41; Pool 1989b; Toro and Aracil 1988).

tems theory to gain greater insight into some archaeological patterns and concepts. For in-

The importance of the Darwinian mechanisms of mutation, external natural selection, and gradualism is being challenged by paleontological and biological evidence that

Minoan domestic spaces combined regular recurring patterns and irregular variation.

small-scale agency, cooperation, symbiosis, and self-organization that create order from

Archaeologists have applied nonlinear sys-

stance, Hitchcock (1997) has analyzed how

Doonan (1997) used a dynamic model to redefine the emergence of complex social systems in Archaic Italy as products of the complex interaction of a number of variable behaviors rather than the result of any singular cause. I have related small-scale temporal variations in gravestone frequency to chaotic processes such as wars and _ epidemics (Spencer-Wood 1989b, 1996).

Interest in the processes by which order is created from disorder or chaos has led to the development of evolutionary theories of selforganization through cooperation and symbiosis. These theories challenge the overemphasis of Darwinian evolution on intra- and interspecies competition in adapting to external environmental constraints. Innovative evolutionary theory has researched the survival advantages resulting from initially random behavior that led to cooperation and self-organization in organisms ranging from slime molds to termites and humans in growing settlements (Allen and Sanglier 1978, 1981; Prigogine and Stengers 1984:181189). Margulis and Sagan (1986) have argued from previously overlooked evidence that oxygen-breathing plants and animals evolved from the symbiosis of nonoxygenbreathing and oxygen-breathing bacterial organisms.

Further, contrary to Darwin’s premise that unlimited growth occurs in natural populations unless it is externally checked, biolo-

points

instead

to

mechanisms

of

internal

chaotic situations generated by the recurring mass

extinctions

(Augros

and

of punctuated

Stanciu

1987).

equilibria

Further,

resis-

tance to entropy is seen as a larger evolutionary driving force than the environmental selective forces requiring adaptation in Darwinian evolution (Goerner 1994:78). Optimization models overlook the possibility that inertia may lead to systemic malfunctioning and the possibility of systemic transformations that change the definition of the evolutionary problem (Prigogine and Stengers 1984:207). Concerning Darwinian evolution, Stephen Jay Gould (1980:120) has stated, “The synthetic theory...as a general proposition, is effectively dead, despite its persistence as textbook orthodoxy.” BRIDGES BETWEEN NONLINEAR SYSTEMS THEORY AND FEMINIST THEORY

This section discusses historical relationships and complementary similarities and differences between feminist theory and nonlinear systems theory. Similarities reinforce both theories. The possibility of expanding the horizons of each theory with complementary aspects of the other is also discussed. Revolutionary Theories Both nonlinear systems theory and feminist theory have asked new questions and developed methods to reveal evidence of kinds 117

SUZANNE

M.

SPENCER- WOOD

of processes and agents of major cultural change that had been previously considered insignificant. Further, both theories have developed inclusive worldviews that have revolutionary implications for scientific epistemology and methodology. They both critique some fundamental assumptions in science and social theory that are historically derived from gender dichotomies critiqued by feminists. The historical critique of gender biases in Western culture by feminist theory deepens our understanding of critiques of traditional science advanced in nonlinear systems theory.

deprived many innocent women of their basic human right to life. This example shows the perniciousness of the cultural construction of men as rational, objective, and orderly in opposition to emotional, subjective, chaotic women. The scientific valorizing of objectivity and order as opposed to devaluing subjective reasoning and chaos is as flawed as the gender dichotomy on which these oppositions are based. This revealing feminist research on the horrific origins of the Western cultural valorization of the (male) orderly objective scientific method sheds light on the dominant male-generated scientific view of the world and nature as a mechanical clock that can be understood by taking it apart. Nonlinear systems theory critiques the limitations of this

Historic Gendered Foundations of Scientific Dichotomies The nonlinear systems theory critique of the order vs. chaos dichotomy is related to the feminist critique of the structuralist opposition between objectivity and subjectivity

mechanistic reductionist worldview without knowledge of its male-gendered historically

because these culturally constructed opposi-

situated cultural construction. Yet this cri-

tions were historically derived from interre-

tique in nonlinear systems theory is closely re-

beginning

ing of science, objectivity, and order. Feminist theory could easily expand its cri-

lated gender dichotomies. Feminist research has revealed that at the of

the

enlightenment

lated to feminist critiques of the male gender-

Francis

Bacon (1561-1626) and René Descartes (1596-1650) followed the dominant gender ideology, valorizing orderly objective reasoning as male and devaluing subjective reason-

tique of the gendered construction of science to encompass the order versus chaos dichotomy and the scientific view of the world as a mechanical clock that can be understood by

ing as irrational, chaotic, and female (Bordo 1986; Keller 1985:53-54). Men’s objective

dismantling it. Nonlinear systems theory could deepen its understanding of scientific

dictable order by rationalizing Mother Na-

tion from gender dichotomies. Further, non-

science

was

valorized

for producing

pre-

ture’s chaos.

Merchant (1980:168-172) shows that Ba-

con portrayed the purpose of men’s scientific method as the conquest and subjugation of female-identified chaotic nature by maleidentified rational objectivity. She reveals that Bacon made an analogy between men’s interrogation of witches and the scientific method of interrogating nature through torture with mechanical devices in order to extract the truth. Thus the birth of the supposedly objective scientific method justified superstitious religious men in torturing and

burning women accused of witchcraft. An analogy was made between scientific objec-

tivity and what we now see was men’s irra-

tional emotional behavior that unjustifiably

oppositions with knowledge of their deriva-

linear systems theory’s critiques could be expanded to include the feminist critique of the scientific opposition of objective versus subjective reasoning. The limited scientific critique in nonlinear systems theory is no doubt related to its development within the hard sciences of physics and mathematics, which feminists have shown most strongly conflate actual scientific practice with the ideology of absolute objectivity (Code 1991). Critiques of the Valorization of Large-Scale External Causes of Orderly Patterns and the Devaluation of Internal Small-Scale Variation Nonlinear systems theory has criticized the scientific focus on external causes of orderly 118

STRANGE

ATTRACTORS

patterns and the dismissal of internal small-

hierarchical dichotomies. Nonlinear systems

scale variation as insignificant noise. Simi-

theory has shown that order and chaos do not

larly, feminist archaeologists have critiqued archaeology’s continued focus on large-scale

constitute a simple dichotomy; rather, they

occupy two ends of a spectrum of system states that combine order and chaos to vary-

orderly recurring patterns, such as cultural

norms, and the frequent dismissal of individual deviations from norms as insignificant “exceptions” (e.g., Spencer- Wood 1992:105). Further, feminists have critiqued the valorization of large-scale external drivers of major

ing degrees (Stewart 1989:22). Similarly, an inclusive feminist approach critiques cultur-

ally constructed

cultural change and the devaluation of inter-

nal cultural factors, such as gender and ethnicity (Conkey and Spector 1984; Wylie 1991). The scientific focus on large-scale external variables and the neglect of internal small-scale variables has gendered roots in the dichotomy valorizing men’s external large-scale public history and devaluing women’s domestic/internal small-scale work.

varying degrees

to

(Spencer-Wood

1993:129-

chaos are related ona continuous spectrum in real-world processes and systems could be expanded to include the feminist construction of reasoning as a continuum that includes the full range of combinations of objectivity and subjectivity. Feminist theory could be expanded to consider the continuum

Competition and Neglect of Cooperation in Darwinian

between order and chaos developed in nonlinear systems theory.

Evolutionary Theory Both nonlinear systems theory and feminist theory have critiqued the overemphasis on competition for survival in Darwinian evolutionary theory, and both have shown that cooperation can also provide powerful evolutionary advantages. Feminist theory and nonlinear systems theory could each be enriched by research and cooperation in the other field. Feminist theory could be

Relating Diversity, Variation, and Large-Scale Orderly Patterns Both nonlinear systems theory and postmodernist feminist theory have taken a holistic inclusive approach to show that individual and small-scale variation are as essential as orderly patterns in accurately describing realworld patterns and processes. Feminist and/and thinking is concerned with the full range of diversity in gender ideologies, roles, behaviors, and relationships, including not only cultural norms but also individual deviations. The inclusive feminist concern with both diversity and norms is similar to the nonlinear systems theory concern for how small-scale irregular variation and orderly re-

expanded to consider the importance of non-

Inclusive Worldviews and Continuum Models Inclusive perspectives in both feminist theory and nonlinear systems theory have led to breaking down boundaries in gender-based

dichotomies

131). The inclusive approach of nonlinear systems theory in researching how order and

Critiques of Overemphasis on

linear processes such as symbiosis and selfOrganization in human sociocultural evolution. Nonlinear systems theory could be expanded to include feminist critiques of the focus on male competition in early hominid evolution. Nonlinear systems theory has already begun to enhance our understanding of cultural as well as biological evolution (e.g., Augros and Stanciu 1987; Goerner 1994: 94-97; Laszlo 1987:87-109).

gendered

model social dimensions as continua, including the range of scientific reasoning processes combining subjectivity and objectivity to

curring patterns interrelate across scales in

complex real-world patterns and processes. Diversity among individuals and groups is the source of small-scale irregular variation. Thus nonlinear systems theory is congruent with an inclusive feminist approach that analyzes the diversity in small-group and individual behaviors. Postmodern feminist research has been inclusively concerned with the relationships between small-scale variations or 119

SUZANNE

M.

SPENCER-WOOD

diversity in gendered behaviors and other small- and large-scale internal and external polythetic groups and factors, from classes, races, and ethnic groups to cultural contact and exchange (e.g., Scott 1994; SpencerWood

1992:105, 1994a,

1999a). Nonlinear

systems theory could be expanded to include diversity and gender systems, whereas feminist theory could be expanded by considering the variation in cultural phenomena resulting from social diversity. Small-Scale Social Agency in Large-Scale Change Postmodern feminist theory and nonlinear systems theory are both concerned with indi-

by the larger cultural environment. Nonlinear equations model variation in internal variables—from population migration to settlement growth—that are affected by environmental factors (Stewart 1989:268). How-

ever, nonlinear equations do not consider the effects of random external inputs that are often considered in feminist research. Nonlinear systems theory could be expanded by adopting an approach more similar to the broader feminist historical approach that researches the external constraints imposed on historic women, as well as opportunities provided by innovations such as the American Declaration of Independence.

although

nonlinear systems theory uses the language of small-scale variation. Both theories have

POSSIBLE EXPANSIONS OF NONLINEAR SYSTEMS THEORY SUGGESTED BY COMPLEMENTARY ASPECTS OF

syncratic behaviors are sufficient conditions to shape large-scale change. Further, both

The similarities between nonlinear systems theory and postmodern feminist theory rep-

deviations

larger nonoverlapping areas in each theory. This section suggests some areas of feminist

vidual found

and

small-group

that individual

agency,

and

small-scale idio-

theories are concerned with how individual and

small-group

variations

or

from normative behavior can be produced within an orderly cultural system. The differ-

ences are that nonlinear systems theory fo-

FEMINIST

THEORY

resent points of intersection that are parts of theory that might be useful in expanding the

scope of nonlinear systems theory. In fact, some expansion of nonlinear sys-

cuses on how the system generates random variation and responds to it, whereas feminist

tems theory in the directions recommended

tural processes such as migration and the growth of cities. Feminist theory might be en-

position between male-associated culture (mind) and female-associated nature (body).

theory focuses on how individuals and social groups use their cultural context to create change, as well as cultural responses to such change. Nonlinear systems theory could be enhanced by applying the feminist theorizing of social agency to nonlinear equations modeling the self-organization involved in cul-

riched by considering nonlinear social agency, in which individual actions indirectly produce unexpected large effects.

External Constraints Both nonlinear systems theory and feminist theory are concerned with external environmental constraints on internal system variables. Feminists have researched the constraints imposed on women by the cultures they live in, as well as the constraints imposed

above is already occurring. It is probably no accident that a woman has extended the nonlinear systems theory critique of the order vs. chaos dichotomy to the related opposition of mind vs. body (Goerner 1994:7), although, surprisingly, she didn’t connect this dichotomy to its gendered source in the opThe related apparently ungendered saying “mind over matter” masks the deeply gendered scientific worldview established by Bacon and Descartes in which male orderly rafemale natural dominates tional mind chaotic matter. Nonlinear systems theory could also follow the lead of feminist theory by expanding critiques of specific oppositions to critique the structuralist claim that binary either/or thinking is innate and universal and that it

I20

STRANGE

ATTRACTORS

justifies the cultural construction of reality in sets of dichotomies. The gendered origins of related scientific dichotomies critiqued both in nonlinear systems theory and in feminist theory further argue that nonlinear systems theory could also be extended to critique structuralism in general. Nonlinear systems theory could be enhanced by adopting more of the awareness of historical situatedness in feminist theory. Nonlinear equations model “chaotic” patterns and processes of change from any given starting system state and variable initial conditions without considering their historical context. Nonlinear differential equations do not model the sources of variation. Nonlinear systems theory and models could be improved with a more historical perspective, just as feminists have increased our understanding of scientific dichotomies by researching their gendered origins. POSSIBLE EXPANSIONS

race, has been modeled with nonlinear systems theory. The cold war was clearly a case of highly sensitive initial conditions because the small individual action of pushing a button would have started a major war (Saperstein 1984; 1988).

and

Mayer-Kress

At the methodological level the similar concern for diversity in postmodern feminist theory and for small-scale variation in nonlinear systems theory has led me to suggest that feminist theory adopt the nonlinear “fuzzy logic” method of “fuzzy sets” (Kosko 1993; Spencer-Wood 1993:129-130, 19942: 177-178). In fuzzy sets or groups members share some but not all behaviors, experiences, or characteristics, so the group is defined by a range of behaviors, experiences, and characteristics rather than by essentially identical criteria (Spencer-Wood 1992:105, 1993:130-131). Nelson (1997:176) has also suggested that feminists consider using the method of fuzzy sets. Nonlinear theoretical concepts such as fuzzy sets and nonlinear processes have been useful in expanding my feminist archaeological research on nineteenth-century women’s social reform institutions, although I have not used systemic language. Cultural systemic constraints help explain why these

OF

POSTMODERNIST FEMINIST THEORY SUGGESTED BY COMPLEMENTARY ASPECTS OF NONLINEAR SYSTEMS THEORY

Beyond intersections of similarity, feminist theory might benefit from considering some of the different perspectives and methods in nonlinear systems theory. This section suggests a few of the concepts and methods in nonlinear systems theory that might be useful to consider in feminist theory. Feminist theory might benefit from some aspects of the systemic approach of nonlinear systems theory. Feminist theoretical concepts might be enriched by considering the complex dynamics of chaotic processes. For instance, feminist theorizing of the process of constantly renegotiating power might benefit from considering its chaotic dynamics, in which a small individual movement away from the initial balance of power could languish under negative feedback or grow through positive feedback, disrupting the existing balance of power and creating disorder or chaos. In fact, the shifting balance of power in the cold war, measured by the arms

Saperstein

emancipatory public enterprises were cast in

terms of women’s normative domestic roles. Combining a feminist and a systemic approach revealed how the opportunities provided by American religious and democratic values were used by women’s movements to increase their status and power. The negative feedback against suffragettes, who directly confronted male dominance in the public landscape, was avoided by reform women, who claimed they were using women’s higher morality and domesticity to materially reform men’s supposedly sinful capitalistic

cities of stone (Spencer-Wood 1994a, 1994b).

Women’s

reform

movements

appeared

to

grow rapidly in a nonlinear fashion from

small local charitable organizations in churches to large-scale national and international organizations (Spencer-Wood 1991b: I2T

SUZANNE

M.

SPENCER-

237-241). Nonlinear systems theory enabled

me to identify the process by which these

movements spread geographically as discontinuous diffusion, a result of the irregular geography of the women’s networks. IMPLICATIONS AND

NONLINEAR

OTHER

OF FEMINIST THEORY SYSTEMS

ARCHAEOLOGICAL

THEORY

FOR

THEORIES

Feminist theory and nonlinear systems theory

each have fundamental implications for archaeological theory. Both theories have cri-

tiqued scientific dichotomies, and their inclusive theoretical approaches and methods

reveal that both theories are based on and/and thinking rather than the either/or thinking involved in structuralism. Nonlinear systems theory suggests a reconceptualization of culture as a nonlinear system (Spencer-Wood 1996). In archaeological systems theory cultural systems have been modeled as essentially stable and constantly returning to homeostatic equilibrium after small perturbations (Clarke 1978). In contrast, nonlinear systems theory has demonstrated that in many cultural processes random variation can produce cultural change relatively rapidly. Many cultural patterns and processes, such as style horizons, discontinuous diffusion, migrations, and the growth of settlements, can be better understood with a nonlinear theory of culture as sensitive to small-scale internal variations under certain kinds of systemic conditions. In cultural systems these small variations often correspond to the social agency of diverse individuals and small groups, as theorized by

postmodern feminists.

Processual archaeology adopted the dominant scientific systems theory paradigm of reducing cultures into subsystems and major variables with linear relationships such as correlations that were theorized to explain cultural change and evolution. Although reductionist analyses and linear relationships have general explanatory power, nonlinear systems theory has shown that these methods cannot adequately describe or predict the many complex cultural patterns and processes that are nonlinear. Nonlinear

WOOD

causal processes could be added to traditional linear methods of relational analysis such as correlations and regressions. Further, some individually theorized archaeological

nonlinear causal concepts, such as triggering

events and thresholds for change, could be more completely theorized within the larger

framework

of

nonlinear

(Spencer-Wood 1990, 1996).

systems

theory

Nonlinear systems theory and feminist theory both challenge still-widespread essen-

tializing practices of describing cultures, gender roles, classes, races, ethnic groups, or artifact types in monolithic norms, archetypes, or averages. What is lost in describing artifact types with averages and even one stan-

dard deviation are the further small variations in the tails of distribution curves that could lead to major change. Schiffer (1996: 656) has also recently pointed to the problem of dismissing variation as noise in the construction of typologies. Both nonlinear systems theory and feminist theory suggest the utility of inclusive methods, such as continua and fuzzy sets, that can more realistically describe the range of variation in a type or a social dimension. Nonlinear systems theory has proposed the construction of types as fuzzy sets that include variation within types. Fuzzy sets are quite similar to what Clarke (1978:3 5-37, 493) called polythetic groups (in systems theory terminology). This is no doubt because of the strong historical relationship between systems theory and nonlinear systems theory. Feminist theory and nonlinear systems theory both would expand evolutionary archaeology by adding the non-Darwinian internal evolutionary mechanisms of cooper-

ation, symbiosis, self-organization, and indi-

vidual and group agency. Some archaeologists have used nonlinear equations to gain increased understanding of the operational differences between models of biological and cultural evolution (e.g., King and Lindh 1997; Reynolds 1997).

Nonlinear systems theory has revealed that cooperative mechanisms are important in resisting the force of entropy, which is larger than particular selective environmenI22

STRANGE

ATTRACTORS

tal forces requiring adaptation (Goerner 1994:78). Understanding of the process of selection and the “replicative success” of arti-

fact types used in evolutionary archaeology (e.g., Dunnell 1980:38; O’Brien and Holland

1990) may be expanded by considering the complex interrelationships between positive

and negative feedbacks in nonlinear differen-

tial equations. However, Schiffer’s (1996: 646) critique of evolutionary archaeology’s

lack

of concern

with

sources

of variation

is also applicable to nonlinear differential equations. Evolutionary archaeology needs to consider the challenges to Darwinian evolution by nonlinear systems theory. Nonlinear differential equations have shown how many populations naturally grow and decline within limit cycles, without the possibility of unlimited growth posited by Darwin. Further, catastrophic and chaotic processes of change in mass extinctions result in punctuated equilibria rather than the gradualism postulated by Darwin (Augros and Stanciu 1987; Eldredge and Gould 1972). Finally, my research has shown how small-scale variation that increases our understanding of processes of change can be aggregated at larger scales to make change appear either gradual or more sudden, depending on temporal divisions. This shows how the evolutionary models of gradualism and punctuated equilibrium can both obscure nonlinear processes through which small-scale variations lead to large-scale change. In addition, the complexity of cultural initial conditions and system state are not susceptible to completely accurate measurement, so it is not possible to completely predict the course of cultural evolution (Spencer-Wood 1989a,

1989b, 1990, 1996). BRIDGES

TO PROCESSUALISM

AND

POSTPROCESSUALISM

This section explores possible bridges among aspects of feminist theory, nonlinear systems theory, processualism, and postprocessualism. Both feminism and nonlinear systems theory advocate relating the postprocessual concern for individual variation and social 123

agency to the processual concern for large-

scale orderly patterns. Behavioral archaeol-

ogy earlier initiated a concern with human

behavior in archaeological theory (Schiffer 1995b:148-152). Each of these theories has

in different ways urged greater analysis of individual behavioral diversity, which one

thread of processual archaeology has related

to artifact stylistic variations (Hill and Gunn

1977). This suggests that processual archaeology could respond to feminist critique by

considering individual social agency within processual theory. the

Feminist theory suggests a bridge between culturally

constructed

opposition

be-

tween the processual emphasis on objectivity and the postprocessual emphasis on subjec-

tivity and relativism. Similar to postprocessu-

alists, feminists have critiqued processual

claims to absolute objectivity. However, feminists have also critiqued postprocessual claims of complete epistemological subjectivity. Feminists have suggested more realistic models of the dialectic that interrelate subjective and objective reasoning processes in the actual practice of science (e.g., Code 1991; Harding 1991; Spencer-Wood 1992; Wylie 1989). Processualism has thoroughly theorized deductive reasoning processes in the scientific method but could further theorize subjective reasoning processes and their relationships to objective reasoning processes (cf. Watson et al. 1971). Feminists have also offered suggestions for bridging the argument over the extent to which data and observations are theory laden. Processualists have taken the scientific view that it is possible to make neutral objective observations independent of personal theoretical standpoint. Postprocessualists have taken the opposing view that all observations are theory laden. However, some leading postprocessualists have realized that data do offer “resistance” to different interpretations (e.g., Hodder 1992:172; Shanks and Tilley 1992:104). I agree with Wylie (1989, 1992) that data are both somewhat theory dependent and independent at the same time. An inclusive feminist perspective recognizes the unavoidable paradoxes of

SUZANNE

M.

SPENCER-WOOD

partially theory-laden data and the simultaneously subjective and objective nature of reasoning.

Hodder has pointed out that there is a strong bridge connecting postprocessualism and feminism, although there are also strong differences between these theories (Hodder 1992:84-88, 258-261). Postmodern femi-

nist theory preceded and strongly influenced the postprocessual concern for multiple voices, diversity, and individual social agency. However, feminism rejects the nihilistic relativism of the postprocessualist contention that androcentric voices and constructions of the past are as valid as feminist ones (Hodder 1991, 1992:87, 186-187; cri-

tiqued in Spencer-Wood

1992).

Feminists

have shown that androcentric constructions

tiquing androcentric biases in archaeological research feminists have in processual terms shown that some hypotheses can be disproved and other hypotheses have more support from the evidence. Feminists have shown that researchers who know it is impossible to be neutral or absolutely objective paradoxically come closer to this ideal by being aware of their own and others’ biases and theoretical standpoints (Spencer-Wood 1992:106; Wylie 1992:18).

Science and processual archaeology have ad-

vocated overt statements of assumptions and biases. Feminist theory has shown that dialectical critiques from different theoretical

standpoints are needed to reveal unconscious

deeply embedded cultural assumptions affecting archaeological interpretations of data

of the past are inaccurate because their biases create partial distorted constructions that are not adequately supported by evidence. As

and constructions of the past (Spencer- Wood

that there are limits to imposing theory on data, it comes closer to the feminist position

Without minimizing the differences between these theories, some similarities suggest po-

possible but that some can be shown to be biased in not considering some evidence.

ory, feminist theory, processualism, and postprocessualism. These are all revolutionary

ously ignored or overlooked evidence show-

questions and developed innovative models

postprocessualism

increasingly

recognizes

that many constructions of the past may be

1992, 1993).

CONCLUSION

tential bridges among nonlinear systems the-

Feminists ask questions that reveal previ-

theories that have asked thought-provoking

ing that androcentric interpretations are invalid. Feminists are similar to postproces-

and methods. These theories have been concerned with revealing and correcting differ-

sualists in accepting multiple constructions of

the past when interpretations are underdetermined by data. However, feminist research shows with evidence that some constructions of the past are inaccurate and biased (Spencer-Wood

1992). This is similar to the

scientific method, which holds that a hypothesis can only be disproved, not proved. Feminist theory and nonlinear systems theory share the processual scientific concern with eliminating biases. Harding (1991: rr1-118) has called the feminist critical approach of correcting biases “feminist empiricism” because it is concerned with doing better science. Feminist empiricism and nonlinear systems theory share with science the goal of creating the least-biased explanation that most consistently and coherently accounts for all the data available. In cri-

ent kinds of biases in order to create more

complete understandings of the past. Further, each of these theories has researched previ-

ously neglected questions and types of data in order to gain greater insight into cultural patterns and processes. Processualism brought to archaeology the scientific method and an increased interest in variation. Postprocessu-

alism critiqued the processual overemphasis on functionalism and objectivity. But postprocessualism overemphasized (at least initially) the subjectivity in archaeological interpretations, focusing on symbolic meanings of material culture that were underdetermined by evidence. Pushing further, nonlinear systems theory has revolutionized science by challenging some of its fundamental assumptions and demonstrating the critical significance of random variation in complex nat-

124

STRANGE

ATTRACTORS

ural and cultural patterns and processes. Both feminist theory and postprocessualism have asked related questions about the importance of diversity and individual and small-group social agency. Finally, inclusive feminist theory suggests that it may be possible to bridge the objectivity vs. subjectivity and the theory-laden data vs. independent data oppositions that have constituted major theoretical divisions between processualism

ology: Setting the Agenda” served as the catalyst for me to formalize thoughts about connections between nonlinear systems theory and feminist theory that I had mentioned only in passing in previous publications. I would like to thank other “knights of the roundtable” for their useful comments on this paper. Finally I would like to thank Carol McDavid for convincing me to use the phrase “and/and thinking” instead of my earlier

and postprocessualism.

term “both/and thinking” because the latter

could be misunderstood as dualistic.

Acknowledgments. Mike Schiffer’s invitation to the Foundations of Archaeological Inquiry Roundtable “Social Theory in Archae-

125

9

Evolutionary Archaeology Reconstructing and Explaining Historical Lineages MICHAEL

J. O’BRIEN

In the almost two decades since Dunnell (1980) published “Evolutionary Theory and Archaeology”—an article that Schiffer (1996:646) labeled as the “proximate roots” of evolutionary archaeology—an increasing number of archaeologists have begun to show an interest in using Darwinian

theory to examine and explain variation in the things they study. Of course, it could be argued that archaeologists have long been interested in applying evolutionism to the study of change in the material record. If this argument is valid, why single out the last two decades for special treatment? The answer is straightforward: although archaeologists have long been interested in evolution, there are significant differences between modern archaeological interest in evolution and that

of previous decades. The major difference is

that the kind of evolutionism under discussion here—Darwinian evolutionism—represents a signal departure from the kinds of evolutionism in vogue in Americanist archaeology during the late nineteenth century and throughout most of the twentieth century (Lyman and O’Brien 1997, 1998; Lyman et al. 1997; O’Brien and Lyman 2000a). Darwinian evolutionism traces its roots directly back to the publication of On the Origin of Species (Darwin 1859). It is a radical departure from the unilinear, progressive cultural evolutionism of Tylor (1871) and Morgan (1877) that figured so prominently in the writings of prehistorians of the late nine-

AND

R. LEE LYMAN

teenth century (O’Brien 1996a, 1996b). Likewise, it bears almost no resemblance to the

cultural evolutionism of White (1945, 1949,

1959a,

1959b),

Childe

(1951a,

1951b),

Steward (1955, 1956), and their followers (e.g., Sahlins and Service 1960), which played a comparably important role in the processual archaeology of the 1960s and 1970s. A major difference between these kinds of evolutionism and Darwinian evolutionism, at least as the difference applies in archaeology, is a shift in focus away from the evolution of culture per se and toward the evolution of particular and highly idiosyncratic cultural phenomena. It might be assumed that making such a shift would be easy, but it most decid-

edly is not. The difficulty stems from a deeply

rooted tradition in anthropology, and by ex-

tension archaeology, that views culture itself

and not cultural phenomena as the appropriate focus of study.

Americanist archaeologists now routinely

recognize a brand of archaeology labeled variously as evolutionary or selectionist, recognizing

further

that its proponents

claim

a

Darwinian intellectual heritage. Several recent articles and reviews (e.g., Boone and Smith 1998; Maschner 1998; Rosenberg 1990, 1994; Spencer 1997), however, have attempted to demonstrate that this claim 1s empty—that is, that evolutionary archaeology is based on faulty reading of the evolutionary-biology literature. We deny the veracity of such claims (Lyman and O’Brien 1998) 126

EVOLUTIONARY

ARCHAEOLOGY

and view them as little more than caricatures written by persons who themselves apparently have not read, or at least not under-

point with which at least the leading behav-

kind of evolutionary analysis discussed here on the faulty belief that selection ceased to be

this can occur only if we are able to sort out

stood, the applicable archaeological and biological literature. These authors reject the

an evolutionary mechanism relative to ho-

minid evolution once culture was acquired.

This is an indefensible position (Dennett 1995; Lerner 1959; Lyman and O’Brien

1998; Weiner 1994) but one that is widespread both in anthropology and biology (e.g, Angier 1997; Gould 1996; Huxley

1956; Simpson 1949). Other authors (e.g., Schiffer 1996) have argued that the basic tenets of evolutionary archaeology are not so

far removed from those underlying approaches such as behavioral archaeology and that the best of these often complementary approaches should work together. We applaud any effort to expand the scope of evolutionary archaeology, although there are epistemological problems involved in some areas of proposed expansion (O’Brien et al. 1998). Proponents of the evolutionary approach have not been as clear and concise as they should have been in setting forth what it is that makes evolutionary archaeology not only different from other paradigms but, more to the point, what makes it a different kind of approach from others in the intellectual marketplace. Darwinian evolutionary theory, especially in its modern (postSynthesis) expression, is not something that can be picked up through a quick and easy read. If it were, then evolutionary biology would not have witnessed the various intellectual difficulties with which it has grappled over the decades—problems such as the ontological status of species, the nature of adaptation, and so forth. Such conceptual problems are solvable, although resolution has occurred but slowly. If, as evolutionary archaeologists have argued (e.g., Dunnell 1989a; Leonard and Jones 1987; Lyman and O’Brien 1998; O’Brien 1996a, 1996c; O’Brien, ed. 1996; O’Brien and Holland 1990, 1992, 1995a, 1995b; O’Brien and Lyman 2000a, 2000b), artifacts are phenotypic features—a

ioral archaeologist (Schiffer 1996) agrees—

then it would appear that the archaeological

record should inform us about the evolutionary processes that took place in the past. But

such things as homologous from analogous features—a point made by Kroeber (1931)

seven decades ago. Such separation can be ac-

complished, although the process is tedious (Dunnell 1978; Lyman and O’Brien 1999; O’Brien and Holland 1990, 1992; O’Brien et

al. 1994; O’Brien and Lyman 2000, 2000b). The purpose of this essay is to lay out what the tenets of Darwinian evolutionism are and, perhaps more important, what they are not. We use the discussion to correct several false impressions of evolutionary archaeol-

ogy and to point out several future directions

that the approach might take. It is not our intention to provide a nuts-and-bolts discussion of evolutionary archaeology and the myriad issues that have at one time or another been examined; good bibliographies on these topics can be found in O’Brien (ed. 1996), O’Brien and Lyman (2000a), and Teltser (ed. 1995). In fact, we focus rather narrowly on one key topic that if left unaddressed will undermine serious efforts to apply Darwinian evolutionary theory archaeologically. That topic, which must be at the core of any effort labeled Darwinian, is history—specifically, the development of historical lineages at various scales. As evolutionary archaeology matures, and we are confronted with more and more so-called Darwinian explanations, it will be imperative that we lay out in precise terms why the examination of historical lineages is so critical. Previous failure to do so has made it appear that an ecological approach is commensurate with evolutionary archaeology. Although ecology plays an important role in evolutionary studies, without history the resulting narratives are functional accounts of how things work. They tell us nothing about how certain features changed, nor do they tell us why features turned out the way they did. A scientific evolutionism has to answer both functional and historical questions. 127

M. J. O’BRIEN ARCHAEOLOGY

AND

SCIENCE

From the time they take their first course work, archaeologists are constantly reminded that although they study “stuff,” that stuff was once part of vibrant cultural systems. They also are reminded, sometimes subtly, other times not so subtly, to keep their sights firmly on the “Indian behind the artifact,” as Braidwood (1959:79), echoing Taylor (1948), put it. Despite these reminders,

there are those anthropologists, such as Service (1964:364), who would also remind archaeologists of their proper place at the anthropological table: “The day-to-day work of an archaeologist (as an archaeologist, not as an archaeologist turned general anthropolo-

gist or philosopher) is to dig up the remains of peoples and their cultures, to map, measure,

describe, count, and so on, and in his report to make an interpretation of what life was

like ‘then.’” In other words, although archaeologists should remember that they are dig-

ging up the remains of cultures, they should not even begin to think about providing explanations for things being the way they are. Anthropologists, not archaeologists, explain.

Archaeologists in the 1960s began to chafe under such strictures, preferring to think that maybe Service and anthropologists with similar views were referring to archaeologists of a day gone by—the culture historians who had labored to put things in proper time-space sequences but who, it was thought, had showed little interest in understanding the processes that shaped culture and hence the archaeological record. The crop of archaeologists that came along in the 1960S was interested in such processes, and its members were bound and determined to make archaeology an equal partner in the study of culture. To do so required that archaeology be made more anthropological. There was an alternative—to make archaeology a humanity—but given that Americanist archaeologists had long seen themselves as scientists (Lyman and O’Brien 1997; Lyman et al. 1997; Willey and Sabloff 1993), this option had little appeal. But what was the best approach to take in attempting to make the archaeology of the

AND

R. L. LYMAN

1960s more scientific than its predecessor had been—in fact, to make it so scientific that

anthropologists such as Service would have to sit up and take notice? The answer was obvious: Turn the archaeological study of culture into a science, just as White (1949) had said could be done for anthropology in general. And because culture had evolved— White (1959a, 1959b) and his intellectual forebears such as Spencer, Tylor, and Morgan had demonstrated as much—evolutionism would naturally come along as part of the scientific package. The wedding between anthropology and archaeology that was supposed to create a scientific approach to understanding the archaeological record had actually been anticipated decades earlier (Dunnell 1982; Lyman and O’Brien 1997; Lyman et al. 1997; O’Brien 1996a, 1996b, 1996c), but the formal union did not take place until the early 1960s, when Binford

(1962) published “Archaeology as Anthro-

pology,” in which he made an eloquent case for uniting the two fields into a common program. He followed that article up three years later with

“Archaeological

Systematics and

the Study of Culture Process” (Binford 1965). At that point processual archaeology

was born—an offspring of White’s “culturology”—literally, the science of culture (White

1949).

The union between the two disciplines that took place in the 1960s, although hailed in some quarters as the replacement for the stale, old paradigm of culture history that had gripped Americanist archaeology since early in the century, actually vindicated numerous culture historians who had long assumed they were reconstructing past cultures (Dunnell 1982) and at the same time trying to understand something about the processes that had created those cultures. To be sure, some culture historians were little more than “trait chasers,” but many others understood what culture is and how to study its manifestations archaeologically. Ford, for example, the chief architect of culture history in the southeastern United States (O’Brien and Lyman 1998), kept culture clearly in focus throughout his career. In one of his earliest ar128

EVOLUTIONARY

ticles Ford (193 5:8) stated that culture “is in reality a set of ideas as to how things should be done and made.” He also pointed out, using wording typical of the period, that culture “is In a continuous state of evolutionary change since it is constantly influenced both by inventions from within and the introduction of new ideas from without the group” (Ford 193 5:8). And what about Binford’s call to make archaeology anthropological? This was nothing new; numerous culture historians had been calling for this for decades. Phillips (195 5:246-247), for example, one of Ford’s contemporaries, was the author of one of the more memorable lines in Americanist archaeology: “New World archaeology is anthropology or it is nothing”—a quote with which Binford (1962) had no quarrel. Phillips might have penned the quote, but he was speaking for many of his generation when he wrote it. Of course, no one ever asked why archaeology should be anything other than archaeology. Despite warnings raised by a few ethnographers during the early processualist days—one of whom (Harris 1968:360) even went so far as to encourage archaeologists to

“shrive yourselves of the notion that the units

which you seek to reconstruct must match the units in social organization which contemporary ethnographers have attempted to tell you exist”—archaeologists plunged headlong into the science of culture, employing anthropological concepts and units to understand such things as prehistoric marriage rules and the kinds of descent systems that

were in use in the past (e.g., Deetz 1965, 1968, 1970; Hill 1966, 1970; Longacre

1964, 1966, 1968). These exercises, although interesting, began to lose some of their appeal when chinks began to appear in the new archaeological armor. Archaeology thus found itself in a peculiar situation. The discipline, by aligning itself with ethnology, had wanted so desperately to become scientific, but the early studies were found to be flawed on both methodological and conceptual grounds (e.g., Allen and Richardson 1971; Dumond 1977; Friedrich 1970; Stanislawski 1972). For one thing, the exercises

ARCHAEOLOGY

were based on what reviewers found to be naive assumptions about the social structures that underlay such things as information transmission among female potters. Some of this criticism could be passed off as carping about details, but most processualists realized that they were going to have to work harder to provide solid, scientifically valid answers to questions that were being posed about the past. What about all the cultural regularities that ethnologists such as White said were out there waiting to be discovered? How could they be found? At issue, of course, was one big, underlying question: How did one go about making anthropological archaeology scientific? The answer was provided yet again by Binford, who suggested that archaeologists should look to the philosophy of science for answers—a

tack

that

he claimed

(Binford

1972:17) he had earlier been instructed to take by White. As a result processualists began reading the works of philosophers Hempel and Nagel, among others, and almost immediately the archaeological literature was replete with references to such phenomena as the hypothetico-deductive approach, the deductive-nomological model of explanation, and bridging arguments (see Fritz and Plog 1970). The most influential book dealing with science and archaeology written during the early processual period was Watson, LeBlanc, and Redman’s (1971; see also Watson et al. 1984) Explanation in Archeology: An Explicitly Scientific Approach, Watson and her coauthors were later joined in their quest to make archaeology scientific by several philosophers (e.g., Salmon 1975, 1982; Salmon and Salmon 1979), but it was Watson and her coauthors who first laid out a detailed blueprint for how to make archaeology an equal partner in the scientific enterprise. They argued strongly that, first, archaeology could and should be a science and, second, that the end product of any scientific endeavor had to be explanation. The means to achieving that end lay in strict adherence to the scientific method: “Archeologists should begin with clearly stated problems and then formulate 129

M. J. O’BRIEN

testable hypothetical solutions. The degree of confirmation of conclusions should be exhibited by describing fully the field and laboratory data and the reasoning used to sup-

port these conclusions. This is what we mean by an explicitly scientific archeological

method” (Watson et Watson, LeBlanc, on how they defined of their views on

al. 1984:129). and Redman were clear science, acquiring many science from Hempel’s

(1965) Aspects of Scientific Explanation, and

Other Essays in the Philosophy of Science: [S]cience is based on the working assumption or belief by scientists that past and

present regularities are pertinent to future events and that under similar circumstances similar phenomena will behave in the future as they have in the past and do in the present. This practical assumption

of the regularity or conformity of nature is the necessary foundation for all scien-

tific work. Scientific descriptions, explanations, and predictions all utilize lawlike generalizations hypothesized on the presumption that natural phenomena are

orderly....

The ultimate goal of any science is construction of an axiomatized theory such that observed regularities can be derived from a few basic laws as premises. Such

theories are used to explain past events and to predict future ones. Good theories lead to prediction of previously unsus-

pected regularities. Logical and mathematical axiomatic systems are essential as models of scientific theories, but no em-

pirical science has yet been completely axiomatized. As Hempel indicates, it may ultimately turn out for any science, or for all sciences, that the goal is actually unattainable. (Watson et al. 1984:5-6, 14)

The vision of science that Hempel espoused was a reintroduction to the philosophy of nineteenth-century empiricism, although the term usually applied to his view is logical positivism. It is difficult to argue with either Hempel’s notion of science and how it operates or with Watson, LeBlanc, and Redman’s

restatement of Hempel’s views. Science has to be based on the working assumption that

AND

R.L.

LYMAN

past and present regularities are pertinent to

future events and that under similar circumstances similar phenomena will behave in the future as they have in the past and do in the

present.

Without

this

assumption

science

would be little more than a chaotic enterprise best left to soothsayers and the like. But this

by no means suggests that there is only one

way that science views nature; what was lost on the processualists was that science is not some monolithic ontological and epistemological entity that prescribes one particular

formula for understanding the natural world. There are at least two vastly different ways

the natural world can be viewed, both of which are applicable to the study of why organisms change. Both are based on under-

standing regularities in the natural world, but one deals with regularities both in process and product, and the other views regularity only in terms of process. Under the latter, products cannot be “regular,” because they are, in terms of shape and composition, con-

tingent products between below. There

on how processes affected earlier in the ancestral line. The differences the two are explored in more detail are still a few philosophers around

who view science strictly in Hempelian terms, but by the middle of the 1970s it was becom-

ing clear that the model was dying. There

were attempts to keep it alive—for example

by linking it to Nagel’s (1961) bridging-law concept—but these also eventually died out except

among

archaeologists,

who

began

making bridges between the archaeological present and the archaeological past through such things as ethnographic analogy and ethnoarchaeology (see Fritz 1972). In other

words, archaeologists were using the present as an analog of the past—an approach that

had been of concern to some culture historians, who saw the dangers in conflating ho-

mologous and analogous structures. For ex-

ample, Kroeber (1931:151) remarked, “The fundamentally different evidential value of homologous and analogous similarities for determination of historical relationship, that is, genuine systematic or genetic relationship, has long been an axiom in biological science. 130

EVOLUTIONARY

ARCHAEOLOGY

The distinction has been much less clearly made in anthropology, and rarely explicitly,

and molecules act the way they do, but they do not work well for explaining how and why

but holds with equal force.” Despite such early warnings, processualists bad to resort

organisms came to be the way they are. A nitrogen atom, for example, is always a nitro-

to analogy; how else were they going to discover the laws that Hempel said were there—the very laws that, once discovered, led to the formulation of “axiomatized the-

gen atom, regardless of where it is in time or space, and there are deterministic laws that govern how nitrogen atoms interact with other atoms. Similarly, burning H, in O, al-

ory” and thus ultimately to explanation? Now archaeology could access the past

ways produces water. We know that it does today, just as we can bet that it will a million

through the present. All that had to be done

years from now. The safety of the bet resides

data sets and interpret the patterning in terms of modern or ethnographic analogs (O’Brien

the behaviors of atoms and in our understanding the various chemical-physical mech-

served behavior could serve as a guide to what to look for in the archaeological record.

Those kinds of laws apply to invariant properties of inanimate objects, but they do not

was to discover patterning in archaeological

in our knowing what the laws are that govern

1996a). Or, conversely, ethnographically ob-

anisms that carry out the dictates of the laws.

If one found enough corollaries between the

work on such things as the behavior of organ-

haps laws could be written to account for the similarity in pattern. Any slight deviations

Lest anyone think that this portrayal of processual archaeology as a law-based disci-

past and the ethnographic present, then per-

between or among patterns could be explained away in terms of slightly different “boundary conditions,” to use Hempel’s phrase, that had impinged on the creators of the past and present signatures. The end result of this exercise was scientific explanation—defined as interpretation through the discovery of laws. This is the reason that Watson (1986:452) directly equated archaeological interpretation with “describing and explaining the real past.” There are, however, several archaeologists who do not agree with this conflation of interpretation and explanation nor with the belief that the Hempelian view of science by itself can be applied to the study of organisms, including humans. No one has ever denied that essentialist laws—those governing chemical and physical phenomena—do not apply to organisms, but at the level that concerns archaeologists—behavior (why we do what we do) and the products of those behaviors (tools and the like)—they cannot play a deterministic role. They might play a probabilistic role, but at this point it is unclear what this means. How are we to assess the validity of any probabilistic claim? Deterministic laws may work well for explaining why physical things such as elements

isms (O’Brien and Holland 1990, 1995b).

pline is a straw man, note, for example, how Fritz and Plog (1970:405), in their widely cited article, “The Nature of Archaeological Explanation,” defined law: “A statement of relation between two or more variables which is true for all times and places” (emphasis added). Could such a statement be any more definite in terms of what the processualists were after? One might, as has been sug-

gested elsewhere (O’Brien

1996a), skirt the

issue and make a semantic distinction between “universal facts” and “laws”—Binford (1972:18) claimed that White once noted that “Julian Steward doesn’t know the difference between a universal fact and a

law”—but this obscures the real issue, which

is the purported existence of invariant laws that govern human behavior. If there are such laws, then the essentialist notion of science is quite adequate for archaeology. If there are not, then where do we look for explanation? We might begin by taking a closer look at Darwinian evolutionary theory. DARWINIAN

EVOLUTIONISM

As Eldredge (1995:10) points out, evolutionary biologists are fond of tracing their intellectual roots directly back to Darwin, especially to On the Origin of Species. Evolu131

M. J. O’BRIEN

AND

R. L. LYMAN

tionary archaeologists claim a similar, though broken, intellectual lineage. It is interesting, however, that most Darwinists rarely define what it is that makes their work Darwinian. This is certainly as true for archaeologists as it is for biologists. The words evolution, selection, adaptation, and drift appear regularly in evolution-based studies, but in delving into both the biological as well as the archaeological literature, one soon gets the feeling that there is considerable diversity of opinion not only on what the terms themselves mean (Feder 1997; see Keller and Lloyd’s [1992] attempt to ease the problem in evolutionary biology) but on what falls under and what falls outside the Darwinian umbrella. Darwin did not and could not have written a treatise that would withstand a cen-

have particularly well-adapted sets of characters will tend on average to give rise to the next generation. In the simplest of terms Darwinian evolutionism is a framework for explaining change as the differential persistence

changed. Critics might argue that the Modern Synthesis of the late 1930s and early 1940S undermined some of Darwin’s central tenets, but this is nonsense. Darwin produced a way of looking at the natural world that was radically different from what came before it, and the basic tenets of his theory remain unaltered. Hence, we might expect, at the very least, considerable agreement on

and method built around the subject of change. The subject of Darwinian evolutionism is change, not simply difference and similarity—a point with which biologists have long been familiar but that was first pointed

tury of biological progress

and remain

un-

what is and is not Darwinian, but recent discussions in archaeology (e.g., Boone and

Smith 1998; Lyman and O’Brien 1998; Maschner 1998; O’Brien and Lyman 1998; O’Brien et al. 1998) lead one to suspect this 1s not the case. So what is Darwinian evolution? What does any scientific endeavor have to entail to

of variation (Campbell 1970; Dunnell 1980;

Lewontin 1970, 1977). In classical ian terms differential persistence of was conditioned by environmental the relentless selector of variants their fitness

Darwin’s

First, evolution

overarching

particular evolutionism is a body of theory

out in archaeology

than

two

decades

by Dunnell

ago.

(1980)

less

In fact, Dunnell’s

(1980) discussion of Darwinian change is one

of the best encapsulations that has yet been produced in either biology or archaeology, although

his telegraphic

style tends to ob-

scure the full impact of his message: The continuity implied in the terms change and persistence bespeaks a fundamental assumption: the phenomena being

examined are historically and empirically related to one another (Alland 1973:3). It is also critically important to note that

evolution views change as a selective, and not as a transformational, process. Vari-

ability is conceived as discrete. Change is accomplished by alteration of the frequency of discrete variants rather than alterations in the form of a particular vari-

does indeed occur; second,

every group of organisms that now exists, once existed, or will ever exist sprang from a common ancestor, and all ultimately go back in time to a single origin of life; third, species multiply by splitting into separate species; fourth, evolution takes place through gradual variation is constantly being produced in every generation, and those individuals who

know

If the perpetual production of heritable

theory.

change within populations as opposed to through transformation; and fifth, heritable

now

variation—of whatever sort—is central to Darwinian evolutionism, then we are left with the inescapable conclusion that that

(1991) we can identify five components, or to

We

and have known since the third decade of the twentieth century that there is room for chance—drift—in the evolutionary process.

qualify as being Darwinian? Following Mayr

theories,

(adaptedness).

Darwinvariation effects— based on

ant. This characteristic places rather severe constraints on the application of evolutionary

theory,

although

perhaps

not as severe as it may appear on first reading. (Dunnell 1980:38)

Change has to be the central point of any study professing to be Darwinian. And as 132

EVOLUTIONARY

ARCHAEOLOGY

Mayr (1991:107) points out, it has to be a particular kind of change: “[D]uring and after the evolutionary synthesis [of the late

species now covers the globe, there is no opportunity for isolation—a mechanism he has long championed as being a major cause of

change under the influence of natural selection, and variational instead of transformational evolution. ... Any other use of the term Darwinism by a modern author is bound to be misleading.” | Philosophers of science (e.g., Hull 1983, 1985) have a difficult time trying to encircle Darwinian evolutionism—that is, defining what it is that unites modern followers of Darwin’s theories. One could argue, as Mayr (1991:104) has, that “unless a person is still an adherent of creationism and believes in the literal truth of every word in the Bible, every modern thinker—any modern person who has a worldview—is in the last analysis a Darwinian.” He goes on to point out that the “rejection of special creation, the inclusion of man into the realm of the living world (the elimination of the special position of man

could have cultural evolution and develop higher and better concepts” (Angier 1997: 10). Here Mayr sounds just a bit like White. In such a view the human phenotype is perceived as being so plastic that it can adapt to just about anything that nature throws at it. Evolutionism with respect to humans, then, becomes little more than attempting to understand how humans adjust to their environment, usually through reference to deviations from optimal responses. In other words, it becomes a hunt for adaptations. This is the heart and soul of so-called Darwinian ecological anthropology (e.g., Boone and Smith 1998; Smith 1991; Smith and Winterhalder 1992; Winterhalder and Smith 1981)—a field that tends to focus on the functional aspects of humans and to bypass the historical reasons for those functions (Lyman and O’Brien

of every enlightened modern person are ultimately based on the consequences of the theories contained in the Origin of Species.” Although evolutionary archaeologists would agree in principle with Mayr, they also would find it rather ironic that many evolutionary biologists, who obviously profess allegiance to central Darwinian tenets, would be excluded under this definition because they specifically have not eliminated the special position of humans relative to other animals.

The one element left out of these studies— something that cannot be left out of any study purporting to be Darwinian—is his-

1930s and early 1940s] the term ‘Darwinism’ unanimously meant adaptive evolutionary

versus the animals), and various other beliefs

Mayr, of course, is referring to creationism

when he refers to the “special position of man,” but it is true nonetheless that numerous Darwinian scholars (e.g., Gould 1996) believe that culture and all the things that go with it exempt humans from natural selection. Even Mayr himself slams the door shut on the role of selection vis-a-vis humans, stating recently, “I do not feel there’s any natural selection in any positive sense going on right now” (Angier 1997:10). Based on the context

of Mayr’s remarks, he presumably does not reject that selection once operated on humans; rather, he believes that because the

speciation (e.g., Mayr 1959). Interestingly, Mayr goes on to note, “Theoretically, we

1998; Vayda 199§a, 1995b).

tory. Darwinian evolutionary theory is a theory about how and why particular organisms

look as they do and behave as they do at particular times and in particular places. That, by definition, makes Darwinian evolutionism historical. As Gould (1986) put it, in Darwinian evolutionism, “history matters.” Selection, drift, gene flow, and all the other evolutionary processes are important factors

in Darwinian explanations, but without history they are simply processes and mechanisms with little to do and certainly nothing to produce.

Darwin

was a historian of na-

ture, always keeping one question in front of him: Why do things become something else?

His whole theory of the multiplication of

species, as with his theory of common descent, was a theory about history: All groups of organisms are descended from a common ancestor, and through time groups of them gradually split off to become new species. What is descent with modification through 133

M. J. O’BRIEN

natural selection—the embodiment of Darwin’s various theories—if it is not a historical statement? History implies the passage of time, and time is what is missing from many recent attempts to employ Darwinian evolutionism in anthropology and archaeology (O’Brien and Lyman 1999b). It is that missing element that is the focus of much of the subsequent discussion. DARWINIAN

EVOLUTIONISM

IN

ARCHAEOLOGY

One of the problems with applying Darwinian evolutionism archaeologically is that Darwin did not have the archaeological record in mind when he formulated his theory—that is, it was not written in archaeological terms (Dunnell 1995). But having said that, we cannot, as Rindos (1989:5) once pointed out, blame Darwin for not doing our

work for us. It is up to us to devise and systematically use methods and units that allow us to incorporate archaeological materials into evolutionism. Modest efforts have been made in this direction since the publication of Dunnell’s (1980) “Evolutionary Theory and Archaeology,” but progress has been slow. This should have been predictable to anyone who, after reading Dunnell’s article, looked ahead to what the future held for an evolutionary archaeology. The discipline was fragmented into any number of “isms,” with new paradigms springing up annually—an event welcomed with open arms by some archaeologists (e.g., Schiffer 1988) but certainly not by all. The term evolution, which saw consider-

able use during the heyday of processualism

as a result of the influence of White, Sahlins,

and Service through Binford, was rarely used in the 1980s except in a large and almost metaphorical sense. There were a few exceptions—Rindos’s (1984) The Origins of Agriculture: An Evolutionary Perspective and

Leonard and Jones’s (1987) “Elements of an

Inclusive ogy,” for between. situation dictable.

Evolutionary Model for Archaeolexample—but they were few and far But again, to anyone sizing up the in 1980, this should have been preIt is not easy to displace Whitean

AND

R. L. LYMAN

thinking, especially when what is replacing it is a theory about how and why organisms change over time, becoming in the process entirely new kinds of organisms. How, one might ask, does such a theory help us archaeologically, where the material record is inorganic? Stones and pots do not evolve—Brew (1946) pointed that out decades ago, effectively squelching any further attempts on the part of culture historians to incorporate biological evolutionary language into their analysis (Lyman and O’Brien 1997). Given such a perspective, which was discipline wide by 1980, the future should have looked grim indeed for any prospects of incorporating Darwinian evolutionism into archaeology. However, the past two decades have seen modest growth in literature on the subject. Modern efforts to sketch out an evolutionary archaeology now include dozens of articles

and papers,! as well as several books and dissertations,~ that either implicitly or explicitly hinge on extending the phenotype to include things found in the archaeological record. These attempts not only have been based ona better understanding of the complexities of evolution than the level of understanding demonstrated by culture historians, but most proponents also understand the need to avoid charges of reductionism—the distillation of theoretical principles from one field for use in another—which was the death knell of attempts made by cultural historians to incorporate evolutionism into archaeology. Evolutionary archaeologists might be pleased with the renewed attention Darwinian theory is currently receiving in the discipline, but they also realize that there are two related measures of success. The first has to do with how faithfully evolutionary theory 1s being applied and the second with how widely it is being applied as an explanatory framework. With regard to the first point, it is difficult to call evolutionary archaeology successful if what is being produced in its name does not follow logically from Darwinian evolutionary theory. Critics will immediately point out, however, that regardless of whether we are talking about biology or archaeology, Darwinian evolutionism itself is

134

EVOLUTIONARY

ARCHAEOLOGY

not some integrated, compact package of theory and approach but a much larger, um-

chaeological terms. Neither is it a cure-all for the myriad problems that confront evolution-

brella-like set of theories and approaches. There is truth in that statement. Fortunately,

one does not have to interpret Darwin’s writings to understand how his reasoning revolutionized the ways we view nature. On the Origin of Species is not Holy Scripture; hence, we do not have to search for deep meaning or worry about putting its words in the context of the time. Darwin’s explanation of descent with modification works just as well today as it did in the 1860s. Does this mean that Darwin knew of and identified all the mechanisms of evolution? No. If he had, there would have been no need for the Modern Synthesis in biology that occurred in the late 1930s and early 1940s and wedded the views of the paleontologists with those of the geneticists and neontologists. In other words, Darwin’s original theory has been refined and extended as we learn more about the natural world. But significantly, Darwin understood that variation was present in nature; he understood the rudiments

of transmission

and inheritance

(like pro-

duces like); and he understood, in the quaint

parlance of his day, how selection worked. Together those three things created the only engine that Darwinian evolution needed. Other components of the engine were identified later, but they did not change its basic design. Could it legitimately be argued that so much extension has been made to the original theory that Darwin himself would not recognize modern evolutionism? No. He would recognize it for the simple fact that his three basic

components—variation,

inheritance,

and selection—still form the centerpiece of the theory. The key point is that when archaeologists profess to do Darwinian archaeology, they should ask themselves, is it really Darwinian, or is it something else (Lyman and O’Brien 1998)? One has to be careful here, lest what is being said sound condescending and selfaggrandizing. No one has ever suggested that Darwinian theory is a cure-all for archaeological problems. In fact, it is not and cannot be a panacea because it was not written in ar-

ary biologists,

and

thus they make

use of

essentialist approaches—say, of the physicalchemical kind—in their everyday work. It is, however, the solution to archaeology’s evolutionary problems (O’Brien et al. 1998). With regard to the second point above, of what use is theory, regardless of how well it explains something, if no one bothers to employ it in real-world situations because either it is not understood or it appears that the theory is being applied in contrived situations? It should come as no surprise that archaeolo-

gists, as with scientists in general, have large stakes in the paradigms under which they operate, and hence they are not easily convinced that the manner in which the discipline has traditionally approached the material record is based on the wrong model of science. What about the theory being used to explain the nature of the record? Could it be inappropriate as well? Or, worse yet, is it possible that the “theory” really is not theory? Some archaeologists are so strongly wedded to one paradigm or another that they will never be convinced that Darwinian evolutionism has anything to offer. Then there are those who might be willing to listen to the argument but who are unwilling to accept the view that things in the material record are phenotypic in the same way that somatic features are. However, there are numerous archaeologists who probably can be convinced that evolutionism is the most powerful explanatory tool available for some phenomena. Acceptance, however, will hinge directly on a basic understanding of what evolutionary archaeology is and what it decidedly is not. We are going to have to be clear about such issues as what is meant by Darwinian evolution, what it is that evolves, how change is measured, and so on. As noted previously, even among those with an interest in evolutionary archaeology, there appears to be some confusion over basic premises as well as method. Part of the confusion, we suspect, resides in lack of familiarity on the part of archaeologists with the biological literature, part of it rests on the fact that classical evolutionary

135

M. J. O’BRIEN

AND

R. L. LYMAN

and Lyman

1999a, 2000b). Such a proposi-

theory does not begin to include all the terms needed to address archaeological phenomena (Dunnell 1995; Teltser 1995b, 1995c), and part of it is the result of a lack of clarity on the part of evolutionary archaeologists. This brings us to the crucial point of this

tion demands the study of immanent properties and processes and the construction of laws concerning them (Simpson 1963, 1970; see also O’Brien et al. 1998)—an endeavor to which behavioral archaeologists have made

the missing component in some of what to-

Skibo 1987, 1997; Schiffer et al. 1994; Skibo et al. 1989; Vaz Pinto et al. 1987)—as well as

essay: the role of history in evolution. This is

day is called evolutionism, and if evolution-

numerous

contributions

(e.g., Schiffer

and

ary archaeology is going to have an impact on the discipline at large, it is going to have to contend with this issue. The evolutionaryarchaeology literature has been relatively clear on the centrality of history in any

the construction of a set of units for measur-

the message has not been made as forcefully

mands that the uniqueness of historical contingencies and configurations be considered

Darwinian-based enterprise, but apparently

as it otherwise might have been. Evolutionary mechanisms such as selection and drift have

been afforded front-row status, as have topics such as adaptation, but discussions have

taken for granted that the reconstruction and

explanation of historical lineages is the long-term goal of evolutionary archaeology, just as it is for paleobiology. Perhaps we should no longer take this issue for granted and be more specific about the role of historical explanation in evolutionary archaeology. HISTORY

MATTERS

Evolution constitutes change in the composi-

tion of a population over time. In archaeology the population comprises objects, which are viewed as phenotypic features, and “it is the differential representation of variation at all scales among artifacts for which [evolutionary archaeology] seeks explanations” (Jones et al. 1995:28). Time is treated as a continuous, as opposed to a discontinuous, variable, and “change is conceived in terms of frequency changes in analytically discrete variants rather than [in terms of] the transfor-

mation of a variant” (Teltser 1995c:53). Such changes could be the result of natural selection—shifts in adaptational state—or they could be the result of drift (O’Brien and Holland 1992) or other mechanisms. The analytical challenge is to determine which is applicable in any given situation (Dunnell 1978; O’Brien and Holland 1990)—that is, to construct and explain artifact lineages (O’Brien

ing and describing a lineage’s fossil record

(Lyman and O’Brien 1999)—that 1s, for writ-

ing a historical (materialist) chronicle and narrative. Explaining the lineage involves the

writing of an evolutionary narrative and de(e.g., Beatty 1995; Simpson 1963, 1970). These points have been made repeatedly by

evolutionary biologists (e.g., Lewontin 1974; O’Hara 1988; Szalay and Bock 1991), as well as by evolutionary archaeologists (e.g., Dunnell 1980; Lyman et al. 1997; Lyman and O’Brien

1998;

O’Brien

and Holland

1992;

O’Brien and Lyman 2000a), but perhaps not strongly enough. Boone and Smith (1998:20) state that “evolutionary change and historical change are not the same thing.” By implication more than explication they maintain that (a) much

of culture history consists of a chronicle of “unintended and unselected consequences” (p. 20); (b) phenotypic adaptation can “produce adaptive changes without concurrent selection,” and such constitutes history but not evolution

(p. 21); and

(c) “phenotypic

adaptation in response to environmental conditions...is change” (p. 22). In other words, evolutionary change, on the one hand, is historical and encompasses only change that is driven by selection; history, on the other hand, encompasses nonselection-driven change within a plastic phenotype over time. From Boone and Smith’s discussion it appears that the only evolution—selection-driven change—that has occurred is the creation of a plastic phenotype—human culture—that can shift states as its human carriers/replicators deem necessary (Lyman and O’Brien 1998). The only role for Darwin’s natural selec-

136

EVOLUTIONARY

tion in such a scenario is that of a past pro-

ducer of phenotypes or cultures that were

adaptively plastic. Although there probably

are reasons to believe there is some truth in

this, it appears in some respects to be more a presumption regarding the archaeological

record and modern cultures than it is axiomatic. Certainly it has not been tested archaeologically. More important, how could it be? How can the plasticity of that part of the human phenotype we term culture be measured archaeologically? No one has answered this question satisfactorily, and hence a critical question is left begging. When did particular cultures, if we choose to use that typological construct, become plastic, and why did they become plastic in the particular times and places that they did? This leads directly to the relevance of history in any inquiry categorized as “Darwinian,” be it biological, anthropological, or archaeological. Few if any evolutionary biologists would argue that history is a noncritical component of what they do (see the chapters in Nitecki and Nitecki 1992). The plethora of statements to this effect, especially with respect to identifying particular character states as “adaptations,” continues to grow (e.g., Baum and Larson 1991; Brandon 1990; Burian 1992; Coddington 1988, 1990; Gould 1986; Gould and Vrba 1982; Leroi et al. 1994; Nitecki and Nitecki 1992; O’Hara 1988; Sober 1984a; Taylor 1987; WestEberhard 1992). There is a less explicit recognition of history in anthropology (e.g., Boyd and Richerson 1992b; Mace and Pagel 1994; Vayda 1995a, 1995b) and archaeology (e.g., O’Brien and Holland

1992), with one com-

mentator (Terrell 1999) going so far as to deny the essential role of selection in producing adaptations. Biologists would be surprised to learn this. Boyd and Richerson (1992b:179-180) acknowledge the importance of history, noting that “Darwinian theory is both scientific and historical. The history of any evolving lineage or culture is a sequence of unique, contingent events.” They grapple with the question of what makes change historical, as well as with the issue of how to make historical explanation scientific.

ARCHAEOLOGY

Evolutionary

archaeologists

should

agree

with Boyd and Richerson’s (1992b:201) con-

clusion that in “the biological and social domains,

‘science’ without

‘history’ leaves

many interesting phenomena unexplained, while ‘history’ without ‘science’ cannot pro-

duce an explanatory account of the past, only a listing of disconnected facts.” Further, both anthropologists and archaeologists should examine the exchange between White (1938, 1945) and Kroeber (1946) in this regard (see

Lyman and O’Brien [1997] for a discussion of

this exchange). Archaeology’s claims to unique status within the human sciences are its recognition of how deep and wide human dependence on artifacts is and its access to those portions of past phenotypes. Ethnographers, ethnologists, sociologists, and others who study humans are limited to living human organisms or written records. Only archaeologists have access to the entire time span of culture, however it is defined. Wissler, for one, could have

argued that the archaeological record was unnecessary to his use of the culture-area and

age-area notions to write history (see the re-

view in Kroeber 1931), but instead he employed that record to confirm directly his ideas about culture history (e.g., Wissler 1919). The significance of this observation is found in an instructive parallel in paleobiology. Modern biologists who undertake cladistic analyses might protest that the fossil record is unnecessary for determining the phylogenetic history of organisms, but this position is losing ground as paleobiologists have come to use the fossil record more frequently to help test cladistic hypotheses (e.g., Clyde and Fisher 1997; Donoghue et al. 1989; Fisher 1994; Forey 1992; Hitchin and Benton 1997; Huelsenbeck and Rannala 1997; Norell and Novacek

1992; Novacek

1992; Smith 1994; Szalay and Bock 1991; Wagner 1995). The important point here is that historical questions are the most obvious ones archaeologists can ask. This, of course,

does not

mean that such questions have to be asked. However, archaeologists should ask historical questions not only because they have 137

M. J. O’BRIEN

AND

access to data that provide a direct test of his-

later time... . Just as narrative sentences distinguish history from chronicle, evolu-

torical hypotheses (whether founded in cladistical analyses [e.g., Mace and Pagel 1994],

Wissler’s age-area notion, or some other model) but also because answers to historical

R. L. LYMAN

tionary narrative sentences distinguish evolutionary history from evolutionary chronicle.

questions are critical to gaining a complete understanding of cultural manifestations occupying particular time-space positions. To

Two critical points can be inferred from O’Hara’s discussion: (1) false or inaccurate chronicle cannot produce accurate history,

From an evolutionary perspective, to “ex-

are. Archaeologists such as Kidder, Rouse, and Ford, who were categorized as culture

causes evolution and to demonstrate the con-

chaeologists of the 1960s and 1970s, recog-

demonstrate why this is so we must be clear on what we mean by both history and explanation.

plain means to identify a mechanism that sequences of its operation” (Bell 1997:1). Importantly, the causes precede the consequence

or effect of the working of the mechanisms.

Two important mechanisms are selection and

drift (transmission), both of which are histor-

ical mechanisms. They operate constantly, at

some times more strongly or more rapidly than at others, creating the varying tempo of evolutionary change (e.g., Gould et al. 1987). So what is history other than the passage of time? O’Hara (1988:144), following philosophers of history, provides a useful discussion: [Glenerally speaking a chronicle is a description of a series of events, arranged in

chronological order but not accompanied by any causal statements, explanations,

or interpretations. A chronicle says simply that A happened, and then B happened, and then C happened. A history, in contrast

to

a chronicle,

contains

state-

ments about causal connections, explana-

tions, or interpretations. It does not say

simply that A happened before B and that

B happened before C, but rather that B happened because of A, and C happened because of B.... Phylogeny is the evolu-

tionary chronicle: the branched sequence of character change in organisms through

time. ...[H]istory, as distinct from chronicle, contains a class of statements called narrative sentences, and narrative sentences, which are essential to historical writing, will never appear in [chronicles]. A narrative sentence describes an event, taking place at a particular time, with reference to another event taking place at a

and (2) narrative sentences provide the explanations of why chronicles are the way they

historians and criticized by processual arnized these distinctions decades ago (Lyman and O’Brien 1997; Lyman et al. 1997; O’Brien and Lyman 1998, 2000a). However,

they could not escape the same problem that plagues evolutionary biology today—one

identified by O’Hara (1988) when he distinguished between the explanation of states

and the explanation of events of change. The former emerges from what others have

termed essentialist thinking; the latter consti-

tutes materialist thinking (Hull 1965; Mayr 1959; Sober 1980) and distinguishes Darwinian evolution as not only a different theory of change but a different kind of theory (e.g., Lewontin 1974). It demands, in short, a different ontological perspective. Culture historians tended not to recognize the existence of this perspective and attempted to explain the difference in culture states (types) in anthropological terms, not realizing that a more rigorous approach existed—one that would have entailed explaining events of change in Darwin’s materialist terms. Their struggles led ultimately to the fall from favor, first, of the utility of any Darwinian sort of evolution as a model for theory building (Lyman and O’Brien 1997) and, second, of virtually the entire culture-history paradigm (Lyman et al.

1997).

The second fall from favor opened the door for what came to be known variously as the “new,” or “processual,” archaeology of the 1960s and 1970s. The processualists were reacting to the tendency of culture historians to invoke simplistic “causes” such as invention, diffusion, and migration to explain vari138

EVOLUTIONARY

ARCHAEOLOGY

ation in the archaeological record. As a replacement for what they viewed as the nonscientific nature of culture history, the proces-

sulting in the production of details concerning how a state works—in short, functional

sualists called for adoption of Hempel’s (1942; see also Hempel 1965), covering-law

accounts. Such a perspective is extremely important, as we have stated repeatedly (Lyman and O’Brien 1998; O’Brien et al. 1998), but

had

particular state came to work the way it did,

model of science, just as evolutionary biology

in part done

a few

decades

by itself it can tell us nothing about why a

earlier

(O’Hara 1988). This was, as noted earlier, an inappropriate model to adopt. Its adoption

Save to say that it was inevitable. In other

words, history does not matter; in particular, lip service may be paid to historical contin-

meant that states, as opposed to events of change, were to be explained. Although the

gency, but it is not critical to the building of satisfying, or what Schiffer (1995a:34) re-

difference between the two was recognized in

evolutionary

biology

(Hull

1965;

Mayr

ferred to as “richly contextualized and audi-

1959; Sober 1980), this sticky ontological dilemma still plagues that field of inquiry, as

ence friendly,” narratives (see O’Brien et al. 1998; Rambo 1991). Perhaps an example will clarify the signifi-

evidenced by the continuing debate over the

ontological status of species (e.g., Cracraft 1983, 1987, 1989; Davis 1996; Ereshefsky 1989; Ghiselin 1974a, 1974b, 1981, 1987; Hull 1976; Kitcher 1984a, 1984b; Mayr

cance of the preceding discussion. Borrowing from O’Hara’s (1988) discussion, we might ask why flamingos are pink. This is a question of character state. Thus one could reply that flamingos are pink for the purpose of mate

Schwartz 1981; Sober 1984a, 1984b; Wilson 1996). What has the dilemma meant for archaeology? In the four-plus decades since Phillips (195 5:246-247) stated emphatically that archaeology must be anthropological, the discipline has spent an inordinate amount of energy trying to effect that transformation. Along with this effort have come no fewer

points out, this is purely a “functional, rational, or purposive” answer. There is nothing inherently wrong with such an answer, and such research can be a fruitful source of

1987, 1993; Mishler and Brandon 1987; Mishler and Donoghue 1982; O’Hara 1993;

recognition.

Morgan’s

stage

evolution (Carneiro 1973; Steward White 1959a, 1959b), (2) White’s (19 definition of culture as humankind’s somatic means of adaptation (Binford 205), and (3) Hempel’s covering-law

1956; 59b:8) extra1965: model

(O’Hara’s

Tylor’s,

[1988]

(e.g. Binford

and

state)

1968b;

model

Fritz and Plog

1970;

O’Hara

(1988:150)

but an evolutionist would emphasize the historical basis for flamingos being pink and thus would rethink and perhaps even rephrase the question from one concerning a state of being to one concerning continuous

change. To do so requires that we first establish that flamingos have not always been pink so that we could then say they became pink for a certain (historical) reason. We might

of cultural

Watson et al. 1971). Given archaeology’s desire to be anthropological, the first and third trouble areas reinforced the essentialist ontology that ethnographic states were to be explained, and the first and second dictated the kinds of explanation to be generated. The law-based character of the latter resulted in attention shifting from homologous—that is, historical—similarity both to analogous similarity and to the familiar functionalism, re-

as

hypotheses (e.g., Mitchell and Valone 1990),

than three problem areas: (1) a reincarnation

of Spencer’s,

But

note that some existing flamingos are more,

or less, pink than others. We could then search for reasons for the variation. Recalling that evolutionary change requires heritable variation, we might wonder if perhaps lighter-colored birds have different ancestors than darker ones do. Variation in today’s flamingos might, or might not, have an obvious purpose or function, but it surely has an evolutionary history. As Mayr (1961) once noted, to ascribe a function to the bird on his porch migrating south every autumn is one thing; to know the historical reasons for its doing so is something else entirely. 139

M. J. O’BRIEN We

in no way

AND

are suggesting that every

study of function has to pay homage to the historical pathways that led to the production of a certain tool, for example, but the joining of function and history becomes critical when the topic is adaptation. The underlying assumption of most anthropological studies of adaptation is that natural selection has shaped the observable plastic phenotype we call culture. Leaving aside the important question of when such a supposedly plastic phenotype first appeared in the hominid lineage (or if it appeared before that particular evolutionary branch began), there is value in studies that seek to measure fitness, or degree of adaptedness, among cultures. For example, there should be no argument with Bet-

tinger and Richerson’s (1996:224) statement that knowing the functional reason for why a

dog pants—to regulate body temperature—is important, but is it true that one need “not

R. L. LYMAN

time-space

position

it occupies

than

some

other kind of pottery does? If so, why? In other words, how does that particular state of pottery work in that context? Evolutionists (e.g., O’Brien et al. 1994) have joined forces with behaviorists (e.g., Skibo et al. 1989) in beginning to understand the nature of such historical contexts. These are, however, only the first questions that must be answered. Additionally, what is the selective environment in which that state occurred, and what were the selective environments that led to its appearance? In other words, what was the history that led to that kind of pottery’s selection? These are questions about the history of

change in pottery, and they are what makes

evolutionary archaeology evolutionary. Answering the questions regarding pottery state requires the use of immanent properties and processes; answering the questions regarding

long evolutionary history” and that such a history “is beside the point directly at hand”?

pottery change requires the use of configurational properties and processes. Bettinger and Richerson (1996:226) state that “given time’s ravages, few archaeologists

biologists (particularly paleobiologists) have come to appreciate: To be considered an

structing a ‘how actually’ explanation.” They appear to make two points. First, they are

question that this panting is the result of a To ignore history misses the critical point that

will ever be privileged to participate in con-

adaptation,

a_ history

stating that the historical chronicles and nar-

point elaborated at length in archaeology by O’Brien and Holland (1992). As Gould

evolutionary archaeologists construct are merely plausible stories. This, of course, is true, but it also is true that the accounts are

a trait must

have

demonstrating it was shaped by selection—a

(1995) put it, we need not bother studying the

fossil record—either paleontological or ar-

chaeological—if we can see all evolutionary processes and products in a petrie dish or among a group of college students. Evolutionary archaeologists would not deny, and in fact would emphatically endorse, the notion that showing that a particular phenotypic trait has a positive fitness value is critical. Behavioral archaeologists would agree with this point (Schiffer 1996). In archaeology this requires the measurement of mechanical properties of artifacts (Schiffer and Skibo’s [1987] performance characteristics) in a manner similar to the way one determines that a dog regulates its body temperature by panting (O’Brien and Holland 199 5a; O’Brien et al. 1994). Does a particular kind of pottery work better within the particular

ratives (in the sense of O’Hara

1988) that

theoretically informed and thus are testable.

Second, Bettinger and Richerson are arguing that the real story will never be known. Apparently, what they are doing in distinguishing “how possibly” from “how actually” explanations is suggesting that they find little satisfaction with the former, characterized by O’Hara (1988:149) as statements regarding “how a change may have taken place,” and would much prefer the latter, or how a change “did take place” (O’Hara 1988:150). Bettinger and Richerson’s point is that the latter is impossible to attain. Paleobiologists would not disagree, but neither would they throw up their hands in despair. We make two important points relative to Bettinger and Richerson’s (1996) discussion.

First,

I4O

in making

the

distinction

between

EVOLUTIONARY

ARCHAEOLOGY

“how possibly” and “how actually” explana-

tions, they reference Brandon (1990) but take his discussion out of context. Brandon’s (1990:176—184) point was that “how possibly” explanations are quite valuable and in many cases can only be epistemologically distinguished from explanations of the “how actually” sort. When a “how possibly” explanation both accounts for numerous observations and provides an empirically and logically coherent explanation, it attains the status of a “how actually” explanation, yet it remains testable in light of new evidence. Brandon acknowledges that we may never know when we have truly attained the latter, although he also states that “no one can fairly describe [such a ‘how possibly’ explanation] as merely an imaginative bit of story telling” (Brandon 1990:183). The second point relative to Bettinger and Richerson’s discussion is the nature of their claim that the archaeological record is a poorly preserved reflection of evolution. Excusing archaeologists from rigorous scientific standards because of the formational history of the archaeological record harks back to the decades-old lament over the imperfection of that record (e.g., Hawkes 1954; MacWhite 1956), although even in those days some individuals perceived this as a not-insurmountable problem (e.g., Bennett 1943, 1946). In the context of the last two or three decades these laments are founded largely on the belief that the archaeological record is a relatively static phenomenon and must be reconstituted, with great analytical care, into dynamic human behavior if our questions are to be answered (e.g., Boone and Smith 1998; Schiffer 1996). Paleobiologists have come far in their dealings with the quality of the fossil record since the explicit placement of taphonomic research under the umbrella of biological evolution (e.g., Benton and Storrs 1994).

Studies of formational processes of the archaeological record have made similar strides (e.g., Schiffer 1972b, 1983, 1985, 1987), although it has been suggested (Lyman and O’Brien 1998) that perhaps it is time for archaeologists to shift gears and bring themselves in line with paleobiological taphon-

omy and to stop complaining that the record

is a poor reflection of anything. CONCLUDING

REMARKS

With Dunnell’s work in the late 1970s archaeologists began to explore the inclusion of archaeological phenomena under a Darwinian umbrella, and in retrospect a fairly solid foundation has been laid. Attention has been paid not only to ontological and epistemological issues but also to purely methodological ones. Incorporation of archaeological materials into a Darwinian framework rests on the key tenet that those items were parts of previous phenotypes—what Dunnell (1989a: 44) referred to as the “hard parts of the behavioral segment of phenotypes.” Without this tenet the application of evolutionary theory to the understanding of how things found in the archaeological record came to be the way they are makes absolutely no sense. Further, without the belief that evolutionary processes, including selection, continue to work on humans, it also is senseless to apply evolutionary theory archaeologically. Evolutionary ecologists, at least some of them (e.g., Boone and Smith 1998), do not believe that selection still works on humans—a view shared by a number of eminent evolutionary biologists—Gould (1996) and Mayr (1982; see also Angier 1997), for example—and

so

for them evolutionary archaeology is an empty exercise. Evolutionary archaeologists are going to have to face the fact that not all their colleagues are going to be persuaded that the approach has something to offer. Focusing even more narrowly, fewer colleagues yet will be persuaded that, given the configurational nature of the archaeological record, Darwinian evolutionism offers the best hope of explaining it in scientific terms. No one likes to be preached to, and unfortunately much of the previous evolutionary archaeological literature has tended to ignore some of the rather careful work of those outside the confines of the evolutionary paradigm. For evolutionary archaeology to have the intended impact on the discipline, proponents have to begin paying attention to such things

141

M. J. O’BRIEN

as behavioral studies. As we have pointed out here and elsewhere (e.g., O’Brien and Holland 1995b), studies of specific prehistoric

AND

R. L. LYMAN

some people choose to believe that humans are immune from the process of selection and

of the

have been since the advent of culture, so be it. But for those who do not hold to such a belief, evolutionism offers a solid basis for develop-

tion, evolutionary archaeologists are going

chaeologically visible change. But such an ap-

activities

(Schiffer

1976)—the

heart

behavioral paradigm—are essential components of evolutionary archaeology. In addi-

to have to be clear about such things as what it is that evolves and the units that they use in structuring

the

detailed

kinds

of

analysis

called for in an evolutionary study (O’Brien and Lyman 2000a). But the most critical point to be stressed in future evolutionary studies is that evolution is based in history; in fact, at base level evolutionism is the unraveling of complex historical lineages—a point

that behavioral archaeologists have long realized (Schiffer 1976). The complexity can only be dealt with by sorting out analogous from homologous features, which itself is a historical concern. Someday soon, there hopefully will be no reason to continue the struggle of showing the applicability of Darwinian evolutionism to the archaeological record, and we can get on with developing and using appropriate methods to sort out artifact-lineage histories. These

methods,

in our opinion,

ing scientifically sound explanations for arproach cannot ignore component—history.

important

Acknowledgments. We greatly appreciate

numerous

conversations

on

evolutionism

held over the years with T. D. Holland, R. D.

Leonard, G. T. Jones, H. C. Wilson, H. Neff, and especially R. C. Dunnell and M. B. Schiffer. We thank Schiffer for his kind invitation to present our views at the 1997 Snowbird

Conference and for his comments on various drafts of this essay. E. J. O’Brien also made numerous suggestions regarding content and style of the paper. tr.

See,

Notes example, Braun

for

(1990);

Dunnell

(1978, 1980, 1982, 1985, 1988, 1989a, 1989b, 1992a, 1992b, 1994, 1995, 1996);

Feathers (1989, 1990a); Graves et al. (1990);

Jones et al. (1995); Leonard (1989); Leonard

and Jones (1987); Leonard and Reed (1993,

will come

1996); Lipo et al. (1997); Lyman and O’Brien

from paleobiology (Lyman and O’Brien 1998; O’Brien and Lyman 1999a, 2000A