ASHGATE
Christopher S. Beekman
William W. Baden
f
I
,
Nonlinear Models for
I
Archaeology and Anthropology
I
I
Continuing the Revolution
I,
1
i
t
J
,!セ@
セ@
セZ@
f
%
!
Edited by
ch}ustoperNbセ@
University of Colorado at Denver and Health Sciences Center, USA
WILLIAM W. BADEN
Indiana University - Purdue University Fort Wayne, USA
ASHGATE
© Christopher S. Beekman and William W. Baden 2005
All rights reserved. No part of this publication may bereproduced,stored in a retrieval
system, or transmitted in any form or by any means, electronic, mechanical,
photocopying, recordingor otherwise without the prior permission of the'publisher.
Christopher S. Beekman and William W. Baden have asserted their right under the
Copyright, Designs and Patents Act, 1988, to be identified as the editors of this work.
Published by
Ashgate Publishing Limited
Gower House
Croft Road
Aldershot
Hampshire GUll 3HR
England
Ashgate Publishing Company
Suite 420
10 1 Cherry Street
Burlington, VT 05401-4405
USA
Li
Li
Li
P,
A(
1
2
Ashgate website: http://www.ashgate.com
British Library Cataloguing in Publication Data
Nonlinear models for archaeology and anthropology :
continuing the revolution
1. Social archaeology 2. Nonlinear systems
1. Beekman, Christopher Stockard II. Baden, William W.
930.1'01
3
Library of Congress Cataloging-in-Publication Data
Nonlinear models for archaeology and anthropology: continuing the revolution / edited
by Christopher S, Beekman and William W. Baden.
p.cm.
Includes bibliographical references and index.
ISBN 0-7546-4319-0
1. Archaeology-Mathematical models. 2. Nonlinear systems. I. Beekman, Christopher
Stockard. II. Baden, William W.
5
4
6
7
CC80.6.N66 2005
930.101'51--dc22
2005024902
ISBN 0 7546 4319 0
1
J
Printed and bound by Athenaeum Press Ltd, '
Gateshead, Tyne & Wear.
Chapter 1
Continuing the Revolution
Christopher S. Beekman
University of Colorado at Denver and Health Sciences Center
William W. Baden
Indiana University-Purdue University Fort Wayne
Introduction
This volume presents theoretical perspectives and case studies extending nonlinear
systems theory (commonly called complexity, chaos theory, or dynamical systems)
to archaeological studies of politics and economy. Nonlinear systems theory has
taken the scientific world by storm in the past two decades with promises to sweep
away reductionistic approaches while retaining a scientific and experimental
perspective. This revolution has taken place even as the social sciences have
struggled to rebuild after the devastating critiques of Post-Modernism. The two
trends collided in the 1990s, resulting in a burst of publications in sociology,
political science, and anthropology that explored the possibilities that the new
science offered. Archaeological applications have developed more slowly, and
have not explored as widely. We assembled the contributors for this volume
because it appeared to us that nonlinear systems theory was being interpreted and
applied in a very limited way by archaeologists. In the course of discussions with
colleagues, we found different concepts of how analysis should proceed and
significantly different characterizations of what nonlinear dynamics might mean
for archaeology.
Is it a ‘Post-Modern Science’ attacking science from without? Is it a revolution
from within that redefines what science itself means? Does complexity constitute a
consistent theoretical package, or is it better understood as a cluster of methods
capable of being used by different theoretical schools? Nonlinear research
encompasses a more polyglot range of concepts than usually appreciated, and it is
very unclear whether we can truly make pronouncements beginning with ‘chaos
theory states…’. Continued exploration of these concepts is necessary. This
volume presents contributions by archaeologists from both sides of the Atlantic
and with different theoretical backgrounds seeking to continue the revolution and
extend nonlinear systems to new areas of archaeological inquiry, such as social and
political organization, and subsistence and political economy.
2
Nonlinear Models for Archaeology and Anthropology
Nonlinear Systems: A Brief History
Unlike many theoretical approaches used in archaeology, nonlinear systems theory
was not so much borrowed from other disciplines as it emerged more or less
independently from research in both the physical and social sciences. Research in
physics, chemistry, economics, meteorology, biology, and other fields led to a
recognition that many types of systems could not be studied using the intentionally
reductionistic approach that has been the guiding principle of science since Sir
Isaac Newton. In order to better understand energy, motion, disease, and other
familiar topics, traditional science has proceeded by breaking the problem down
into increasingly smaller parts that could be studied in isolation from the whole.
This allowed experimental manipulation of different variables so as to isolate the
effects of each one.
The successes of this approach have been numerous and need not be repeated
here. But as researchers have tackled more and more difficult problems, the
reductionistic model has become less useful. There has been a growing realization
that studies of closed systems are often unrealistic, and that assumptions of system
equilibrium must be left behind if we are to study more complex systems where the
interaction of many components leads to quite unexpected emergent properties.
Science as a whole has, in a sense, caught up with the kinds of systems that social
scientists have had to deal with for years - open, fluctuating, unpredictable, yet
patterned - and has begun to develop conceptual tools for understanding them.
The research that comes under the heading of nonlinear systems is diverse.
Related research findings in meteorology, mathematics, and geometry (e.g., Lorenz
1963; Mandelbrot 1977; May 1976) were baptized with the term Chaos theory in
the 1960s, encompassing deterministic systems (i.e., based on equations that
behave predictably over short periods of time) that nonetheless eventually become
unpredictable due to the equations’ sensitive dependence upon initial conditions. In
other words the relationships between parts are understood, but prediction fails at
the level of the entire system due to the complexity of all the interrelationships and
our inability to observe the initial conditions with sufficient precision. The
canonical example is Lorenz’s experiment with weather prediction. Lorenz
discovered that an extraordinarily small, seemingly insignificant, variation in the
initial input of one parameter (from 0.506127 to 0.506) into his computer program
resulted in an eventual divergence between two predictions that were expected to
be exactly identical (Gleick 1987:16). This variation related to the arbitrary
precision of his model’s input, i.e., what he could measure coupled with the
accuracy of his observations and the computer’s ability to maintain that precision.
From a statistical perspective these precision issues produce what appear to be
random fluctuations in the model’s trajectory. The mechanisms for the weather are
highly quantifiable and relatively well-known, but their sensitivity to our ability to
accurately measure or observe them means that predictions can only be made a
short distance into the future (up to the prediction horizon), after which the
Nonlinear Models for Archaeology & Anthropology: Continuing the Revolution
3
potential variations become too great. Despite the impossibility of ever adequately
measuring initial conditions (Heisenberg 1927; Nicolis 1995:58-61) or making
medium or long term predictions, recurring trends around topological attractors
allow retrodictive analysis and explanation. We can typically explain the weather
in hindsight, even if we are unable to effectively predict it. Another significant
element of classical chaos theory is the concept of bifurcations. Analysis may well
identify points in time when change is almost certain to occur, but it is often
impossible to determine which of the possible new trajectories will predominate
due to apparent elements of randomness. Historical pathways and contingency are
therefore deeply rooted elements of chaos theory, despite the presence of the often
misinterpreted term ‘deterministic’ which reminds us that even the most irregular
trajectory is defined by causal forces.
Another related field of inquiry that developed in mathematics is Catastrophe
theory (Poston and Stewart 1979; Thom 1975), largely a creation of the 1970s.
This is closely related to chaos, and largely focuses on how even systems with very
few variables can show surprising transformations and nonlinear trajectories
depending upon their interaction. Its abstract mathematical and topological nature
tended to prohibit its use in real world situations, where it became extremely
difficult to quantify any system in terms of just a few variables.
Complexity or Complex systems theory (Pines 1987; Cowan, et al. 1994)
emerged in the 1980s and shares many elements with the previous approaches, but
has de-emphasized deterministic chaos and instead focuses on how the mutual
interaction of many agents or variables can lead to unexpected emergent properties
within the broader system. Complex systems theorists therefore explicitly reject the
intentionally reductionistic perspective associated with the Newtonian approach to
science and have instead tried to develop methods to study entire systems. A major
element here has therefore been the imprecise but nonetheless powerful concept of
emergence, colloquially described as ‘the whole is more than the sum of its parts’.
That is, systems cannot be described through the aggregation of their components,
but require an entirely new description at the level of the system.
In a terminological shift, complexity researchers have appropriated the term
Chaos to contrast unpatterned systems (not strictly true in the earlier definitions of
chaos theory) as an opposing pole to Order – a situation of low potential for change
(Langton 1990). Complex systems (or Anti-Chaos in the most recent parlance –
Shermer 1995; Lansing 2003) lie in the boundary zone between these extremes,
within which there are forms of temporary stability that emerge through the
interaction of internal components. The complex systems approach has been
closely linked to the extensive use of computer simulation due to the amount of
data and number of variables being analyzed, and the field is significantly related
to issues of adaptive learning and artificial intelligence. The best known of these
simulations are the agent based models or artificial societies, in which simulators
have attempted to model phenomena ‘from the bottom up’, concentrating on
simple rules that nonetheless produce complex behaviour and emergent properties
at the level of the group (Langton 1988, 1994; Langton, et al. 1992; Epstein and
Axtell 1996).
4
Nonlinear Models for Archaeology and Anthropology
Much of the impetus for the coalescence of research into complex systems
theory came from the influential research of Ilya Prigogine into dissipative
structures and self-organization over the prior few decades (e.g., Glansdorff and
Prigogine 1971; Nicolis and Prigogine 1977, 1989; Prigogine, et al. 1972a, 1972b).
Briefly, his research recognized that pivotal physical laws such as the second law
of thermodynamics (the pessimistic law referring to the unavoidable loss of
energy) only worked in a closed system, and failed to accurately describe
behaviour in an open system in which energy could be introduced from the outside.
The introduction of energy produced perturbations to the system and caused it to
self-organize into complex patterns far from equilibrium. Common examples of
these dissipative systems from the physical world include fluid dynamics and other
phase transitions. But the perceptive Prigogine quickly saw the application to
biological evolution (particularly the emergence of animate forms from inert
matter) and even social change (popularized in Prigogine and Stengers 1984).
Prigogine’s own inconsistent use of the terms Chaos, Complexity, and
Stability theory to refer to his work reflects the rather loose approach to this
material that most practitioners have taken. For this reason, in this volume we
prefer the one aspect that all these approaches hold in common – their focus on
nonlinear systems. Less catchy perhaps, but more accurate and inclusive. By
nonlinear systems we mean those systemic environments in which the number of
variables or the relationships between them are so structured as to lead to quite
unpredicted and emergent properties. Analysis tends to be oriented towards
retrodictive explanation, not prediction, and the simplistic theoretical dichotomies
that archaeologists are so attracted to, such as materialism vs. idealism,
functionalism vs. conflict, agency vs. structure, or intention vs. adaptation, become
difficult to maintain. As opposed to the systems theory of decades past, nonlinear
systems are open and subject to varying degrees of stability. Rather like the much
touted landscape archaeology of recent years, nonlinear systems invite parallel
analysis at multiple scales (see McGlade 2003).
Many prominent social scientists have weighed in on the importance of
nonlinear studies for human societies (e.g., Adams 1988; Luhmann 1986, 1995;
Maturana and Varela 1980; Wallerstein 1997). Textbooks specifically for use in
the social sciences have appeared (Brown 1995; Byrne 1998; Guastello 1995;
Marion 1999), numerous collections of applications have been assembled (Albert
1995; Arthur and Arrow 1994; Bertuglia, et al. 1998; Bütz 1997; Ellis and Newton
2000; Eve, et al. 1997; Kiel and Elliott 1996; Leydesdorff and van den Besselaar
1994; Milanovic 1997, 2002; Prietula, et al. 1998; Richards 2000; Schieve and
Allen 1982), and a listing of individual studies would probably fill this
introduction, although we might single out a few (Allen 1997a; Artigiani 1989;
Gilbert 1995; Harvey and Reed 1994, 1996; Lansing 1991; Lansing and Kremer
1993; Leydesdorff 2001; Reed and Harvey 1992; Shermer 1995) as fruitfully
exploring the links between existing social theory and nonlinear approaches. But
the dynamism in nonlinear systems theory may perhaps have most to offer
archaeology, as the social science most deeply involved in the issue of social
change over time.
Nonlinear Models for Archaeology & Anthropology: Continuing the Revolution
5
Archaeology and Nonlinear Systems
The earliest interest in nonlinear modelling for archaeology was primarily amongst
those European and American researchers who had drifted away from the
ecological functionalism, or Processualism, of the 1960s and 1970s to explore the
utility of General Systems Theory (Maruyama 1963; von Bertalanffy 1951; Wiener
1948). Concepts broader than adaptation were considered, such as the role of
information in social change, and archaeologists such as David Clarke, Colin
Renfrew, Sander van der Leeuw, and Ezra Zubrow were of central importance,
both in the work they did and the work they inspired in their students (e.g., Clarke
1972, 1978; Renfrew and Cooke 1979; Van der Leeuw 1981a; Van der Leeuw and
McGlade 1997a; Van der Leeuw and Torrence 1989; Zubrow 1985). A genuine
problem remained in that the rare applications of nonlinear approaches at this time
were very simplistic and abstract (Renfrew’s application of catastrophe theory to
the Maya collapse used 3 variables), and it remained quite difficult to determine
exactly how to measure or even identify the important variables in a human
system. Hence variants such as catastrophe theory enjoyed an extremely brief
period of popularity (e.g., Renfrew 1978, 1979).
Ilya Prigogine was a distinct source from which the interest in nonlinear
systems spread to archaeology. Surely encouraged by his own wide range of
interests (see Prigogine 1998 to see what we mean), colleagues of his began to
apply dissipative structures to the social sciences (e.g., Schieve and Allen 1982;
Adams 1988). Eventually, the intellectual descendents of this line carried the
approach further to archaeology (e.g., McGlade 1990, 1999). Besides his work out
of Brussels, Prigogine also spent time at the University of Texas at Austin, and
Americans became introduced to his work more directly (perhaps first mentioned
by Blanton and Kowalewski 1982:15). This southern focus, European in origin,
independently influenced the early work of some of the contributors to this volume
(Baden 1987, 1995; Stone 1999).
Nonlinear systems approaches had an unclear relationship to other theoretical
schools in archaeology. It was the eclectic Jonathan Friedman (1982) who would
make reference to Prigogine’s work in the context of punctuated equilibrium and
other rapid transformational models from the physical and biological sciences as
part of a larger argument - that Processual archaeologists, in their desire to put their
discipline on a scientific footing, had chosen idealized and discredited equilibriumbased theories as their model of science. Nonlinear systems research thus went
from being an approach parallel to Processualism to an independent source of
critique.
But interest in these approaches may have failed to catch on more widely
because most of these applications focused on the analysis of entire systems (see
particularly the contributions in Van der Leeuw and McGlade 1997a), even as
archaeology as a discipline was turning away from macro-scale analysis.
Regardless of whether it caught on immediately or not, Prigogine’s work spawned
6
Nonlinear Models for Archaeology and Anthropology
some of the earliest direct archaeological applications of nonlinear systems since
Renfrew and Zubrow (Baden 1987; McGlade 1990). It is also members of this
loosely defined ‘Prigoginian’ line that will make attempts to confront the PostProcessual critique, by addressing issues of power and top-down political control
(McGlade 1997).
Significant changes occurred in the field in the late 1980s and 1990s, due to
two major related trends. The first was the rapid growth in computing power,
allowing the development of considerably more complex simulations, particularly
agent based modelling. The second was the development of institutional
infrastructure in the form of the Santa Fe Institute (SFI) and other research
organizations within or parallel to university settings. We suspect that agent based
modelling created a fascination for archaeologists that earlier system-wide analyses
of complexity could not, and it brought the scale of modelling down to the level of
individuals just as other theoretical approaches in archaeology were doing the
same. But we must also consider the training in technical methods that the new
infrastructure made possible. Few archaeologists leave graduate school with a
mathematical background beyond that required to carry out multivariate statistical
analyses, and interdisciplinary institutions such as SFI provide an entirely new
training opportunity. The simulations of Anasazi culture change by Timothy
Kohler, Jeffrey Dean, George Gumerman, and their colleagues are the best known
examples of archaeological collaborations to come out of SFI (Dean, et al. 2000;
Kohler and Van West 1996; Kohler, et al. 2000). But this highly prominent
organization (positively treated in Waldrop 1992 and less so in Helmreich 1998)
served as host to a series of other archaeologists over the 1990s and 2000s such as
Robert Hommon, Mark Lehner, Suzanne Spencer-Wood, Tony Wilkinson, and
Henry Wright (see also the contributions in Kohler and Gumerman 2000). Some
have worked on computer simulation while others have chosen to apply complex
systems more metaphorically, and examples of their work appears in this volume.
Certain prominent themes run through much of the research at SFI whether
practiced by archaeologists or by representatives of other disciplines. Human
societies, flocks of birds, the economy, sand piles, the brain, immune systems,
ecologies, ant colonies, and pre-organic molecular formations are all
conceptualized as Complex Adaptive Systems (CAS) sharing many features in
common (Lansing 2003). That is, much of the self-organization of these systems is
described as occurring under adaptive pressures towards greater fitness. Modelling
has often, though not exclusively, made use of computational mechanisms such as
modelling strings of cultural attributes through genetic algorithms (Holland 1998),
adaptive interaction through game theoretic exchanges (Axelrod 1984), the
creation of fitness landscapes to represent optimal and sub-optimal solutions
(Kauffman 1999), artificial life studies involving adaptive learning (Langton 1988,
1994; Langton, et al. 1992), and other tools with strong ties to some form of
Darwinian selection. Even when religious concepts or symbols are given an
equally important role in modelling behaviour or communication, they are often
operationalized again through programming methods that follow a selectionist
structure.
Nonlinear Models for Archaeology & Anthropology: Continuing the Revolution
7
Thus CAS is a distinctive approach that does not represent the entire gamut of
theories currently practiced in archaeology. That in turn suggests that there is only
room here for certain theoretical approaches to society. SFI has been very inclusive
in its efforts to cross traditional disciplinary boundaries. But research concentrating
on the similarities between complex systems may fail to appreciate the differences
between them, which can be just as significant for our understanding. Adaptivist
theory is perfectly legitimate as a theoretical approach, but the rather exclusive
focus on the concept suggests that perhaps other theoretical schools are not
compatible with CAS. This is an implication to be considered as we look at other
areas.
Research on human societies and nonlinear systems in Europe traces distinct
trajectories. Themes that run through social simulation work in western Europe by
researchers like Nigel Gilbert and Jim Doran (Gilbert and Conte 1995; Gilbert and
Troitzsch 1999; Gilbert and Doran 1994) include religion, belief systems,
irrationality, and social action. Social theorists such as Anthony Giddens and
Fernand Braudel are cited alongside Ilya Prigogine. This greater interest in the
social leads to applications that look quite different from those recently appearing
in the United States (e.g., Doran 2000; Doran, et al. 1994; various in Van der
Leeuw and McGlade 1997a) and greater weight to the notion that social systems
are qualitatively different from other complex systems. Archaeologists working
within this intellectual milieu have often chosen to use nonlinear concepts as
metaphors to frame a verbal analysis rather than develop computer simulations
(e.g., Bintliff 1997, 1999a, 2003).
The last five years have seen the range of practical applications of nonlinear
systems theory to archaeology continue to expand. Fractal analysis has been
applied to Maya and Central Plains settlement patterns (Blakeslee 2002; Brown
and Witschey 2003; Brown, et al. 2005) and urbanism at Teotihuacan (Oleschko, et
al. 2000), technological innovations have been examined through the metaphorical
application of dynamical systems and emergence (Roux 2003), self-organized
criticality has been used to analyze pottery style (Bentley and Maschner 2001), and
nonlinearity has been applied to human organization (Crumley 2001). Nonlinear
systems theory has been discussed (R. McC. Adams 2001) and applied (Yoffee
2005) in more metaphorical terms to Mesopotamian social evolution, and
theoretical links between Feminist theory and complexity have been outlined
(Spencer-Wood 2000). R. Alexander Bentley and Herbert Maschner have
published a diverse collection of articles (Bentley and Maschner 2003) examining
complex systems in archaeology and history. Their contributors examined
settlement patterns, chronology, and social networks, and came much closer than
previous volumes to what we wanted to assemble here.
Despite (or because of) its growing profile, nonlinear systems theory has had
its detractors. Helmreich (1998) examined the Santa Fe Institute from a
deconstructionist perspective, pointing out that white male heterosexuals make up
the majority of complex systems researchers and arguing that this has led to a
narrower view of social modelling than might otherwise be the case (e.g., use of
the metaphor of biological reproduction). Khalil (1995) argues that the work of
8
Nonlinear Models for Archaeology and Anthropology
Prigogine and others is inappropriate for human societies, where intention
distinguishes human organization from quasi-cyclical phenomena such as those
modelled from the natural world. Offhand comments by prominent archaeologists
such as Adam Smith (2003:104) and private communications with colleagues have
alluded to what many perceive as the lack of complexity in ‘complex’ systems
research. Much of this comes from theoretical perspectives other than the
selectionist approaches welcomed within prominent centres like SFI. But recently
there have also been critiques from within. McGlade (2003) has expressed concern
that nonlinear dynamics embraces a wider array of concepts, methods, and
implications than appreciated by some practitioners. Among other things, he
argues for the construction of multiple models of the same system from different
perspectives, for more attention to differences across scales, and for a greater
appreciation of the fact that we are constructing tools, not representations of
reality. Mark Lake (2005) has recently argued that the utility of computer
simulation of complex systems may be more limited than once thought. It may be
most useful for sensitivity analysis and specific methodological questions, but it
has produced far less in the area of theory. These critiques, especially from those in
the thick of current research, point up the need for a more detailed examination of
nonlinear systems’ applicability to the archaeological record, and a careful
appraisal of current approaches.
This Volume and the Contributors
A brief exposition of nonlinear studies in Thomas Kuhn’s (1970) terms brings us to
the central concern of this volume. Since at least Poincaré, studies in various
disciplines have been producing what Kuhn called anomalies, or cases left
unexplained by the dominant science paradigm emphasizing the isolation of
individual variables in closed experiments. Complexity and chaos were born as
self-referential entities when the commonalities among these anomalies were
recognized, and Kuhn’s stage of ‘revolutionary science’ followed with the creative
adoption of new approaches and the application of ideas to new disciplines. That
was in the late 1980s and early 1990s, and is best exemplified by the formation of
new institutes devoted to the study of nonlinear dynamics. Over the past 10 years,
however, complexity has become more established and less avant-garde, and
practitioners have focused more on Kuhn’s ‘normal science’, working out the
methodological aspects of how to model this or that phenomenon. The revolution
is evidently over.
This book has its origins in our conviction that the revolutionary and
exploratory period of nonlinear systems studies was too brief and incomplete, at
least in the realm of the social sciences (see also Turner 1997). The CAS approach
promulgated by SFI and other centres for complexity studies seems to imply a
distinct theoretical package different from existing schools of thought in
archaeology. The work by many in Europe on the other hand tends to suggest that
various different theoretical orientations might profitably use nonlinear tools for
Nonlinear Models for Archaeology & Anthropology: Continuing the Revolution
9
their own questions and topics of research interest. These are quite different
interpretations. This volume is intended to present the ideas of a variety of
researchers known for their contributions to other areas of archaeology, with the
hope that it may be possible to clarify what nonlinear systems signify for the
discipline. Is it a new paradigm that offers nothing to those archaeologists who
prefer studies of agency, hermeneutics, and multivocality, or does it cross-cut old
divisions in a manner that can potentially unify the scattered remnants of
archaeological thought left in the early 21st century?
It has recently become clear that practitioners sometimes have different ideas
of what complexity is, and as a result adopt highly distinct forms of discourse.
Some refer to it as a ‘Post-Modern’ science (Spencer-Wood 2000) because it
eschews determinism and embraces contingency, while others see it as simply a
development that unifies major trends within many of the sciences (see Price
1997). Some see nonlinear studies as a natural outgrowth of the computer
revolution, allowing analysts to work with vastly increased volumes of data, while
others see little need to adopt computer simulation to apply the theoretical
principles involved. Our goal for this volume is to continue the revolution in the
study of nonlinear systems by re-evaluating its theoretical bases and how they
relate to paradigms already in place within the discipline. To this end we have
assembled research from a broad variety of viewpoints as to what nonlinear
dynamics might mean for archaeology. The articles in this volume are meant to
bridge the gap between anthropological and archaeological theory and nonlinear
concepts. In some chapters the authors have chosen to work with archaeological
case studies to illustrate their points, but in others they have drawn upon
ethnographic, historical, or contemporary societies where the characteristic
ambiguities of archaeological data could be overcome. Regardless of the datasets
used, the contributors are all concerned with the importance of nonlinear systems
for archaeology.
The ability of humans to incorporate, integrate, and use information is the
focus of Robert Hommon’s investigation into the concepts behind CAS. Collective
behaviour is presented as a result of ‘schema-driven behaviour of interacting
agents’ where schema refers to internal, experience-based rules following GellMann’s use of the term (1992:10), but the definition could equally well have come
from current authors in social theory. Hommon is interested in how human
societies differ from other CAS by invoking processes of appropriatizing
(encouraging conformity) and ecaptation (altering the environment), and the
presence of stratified control hierarchies (unequal empowering of specific agents
or groups of agents). Although primarily drawing upon examples from modern and
recent human societies, he cites archaeological and historical evidence to
demonstrate the influence of these three processes in Hawaiian culture prior to
European contact.
Carole Crumley is also interested in human organization, but particularly those
forms that do not correspond to the stereotypical stepped hierarchy. Since the
interaction of components is one of the central elements of any complex systems
approach, she focuses her chapter on those interactions, in particular the concept of
10
Nonlinear Models for Archaeology and Anthropology
heterarchy. A heterarchy is a meshwork of systems whose elements are either
unranked or potentially ranked in a number of ways. The less rigid, more flexible
structure of heterarchical systems is an adaptive tool that can more effectively
combat ‘surprise’ within a society (e.g., environmental change, invasions,
epidemics, etc.). Using examples of ‘disorganized organizations’ from recent times
such as Al-Qaeda and the Anarchist movement, Crumley demonstrates that in
uncertain times the sharing of information through flexible authority structures
reduces risk by increasing available information to decision makers and
multiplying solutions.
The fact that the concept of agency, in varying degrees, runs through many of
these contributions reinforces the notion that the role of the individual is an
important element of nonlinear studies. Many aspects of the nonlinear approach
concentrate on mapping the emergent outcomes of the actions of individual agents,
whether those agents represent people, ducks, or water molecules. Christopher
Beekman builds upon prior suggestions by Nigel Gilbert that the authors of agent
based simulations of human societies need to pay greater attention to the social
actor and current social theory on agency. Prominent social theorists like Anthony
Giddens, Margaret Archer, Pierre Bourdieu, and others have all discussed agency
in different ways that have significant implications for agent based modelling.
Beekman argues that, far from pursuing an empirical, ‘bottom up’ agenda with
agent based simulation, modellers have unknowingly espoused particular theories
about how the individual relates to society and left others unconsidered. He also
argues that the empirical ideal that crops up repeatedly in enthusiastic portrayals of
complex systems may not be possible. While many authors have emphasized selfsimilarity across scales (drawn from fractal analysis) as a central element of
nonlinear studies, Beekman instead points to the work by Prigogine and by social
theorists arguing instead for phase transitions, thresholds, emergence, and distinct
scales of analysis with different rules. He finishes with an example drawn from
Late Formative-Classic period (200 B.C.-A.D. 550) western Mexico to illustrate
some of the complexities of collective agency.
Tammy Stone is also interested in the issue of quasi-group formation, or the
appearance of highly unstable social groups in middle-range societies. Social
institutions which aid in the flow and processing of information are argued to be
self-organizing (Van der Leeuw 1981b; Stone 1999). Thus changes in information
flow can have major repercussions on the structure of social groups which process
information. As the intensity of flow among closely interacting groups increases,
social complexity can differentially expand or collapse at spatially concentrated
loci on the social landscape, forming information vortices (after Van der Leeuw
1981b). Stone applies these insights to quasi-group or faction formation as they
occurred among the Hopi early in the 20th century. She contrasts the response at
Orayvi against that of other Hopi communities that did not fission. Stone’s
discussion is consistent with Prigogine’s argument that the current state of a
system is one of many possible outcomes ultimately self determined by its initial
conditions and by the response to subsequent disruptions in the flow of necessary
resources (here, information).
Nonlinear Models for Archaeology & Anthropology: Continuing the Revolution
11
William Baden examines the applicability of dissipative structures within
archaeological models of agricultural systems. Baden provides an historical
overview of anthropology’s adoption of thermodynamic concepts, positioning
Prigogine’s nonlinear paradigm within anthropology’s on-going theoretical use of
these principles. By applying Prigogine’s concept of systems far from equilibrium
to prehistoric cultures, he updates and operationalizes Leslie White’s (and his
followers’) original arguments linking the Second Law to cultural evolution. Using
Mississippian examples from the southeast United States, Baden demonstrates how
agriculture, as a dissipative structure, can be seen as an external source of negative
entropy consistent with the theoretical suggestions of Schrödinger and Prigogine.
Predictable disruptions in the flow of energy and their impact on social systems far
from equilibrium are correlated with observed phase transitions in the
archaeological record.
Tony Wilkinson represents the MASS project and its work on settlement
formation in northern Mesopotamia, and his paper discusses the impact of
simulation upon traditional landscape archaeology. Clearly influenced by Kohler’s
highly important simulations of the ancestral puebloans, the MASS project seeks to
model an entire city of a complex, stratified society. This is a tall order, but is an
excellent example of how agent based simulation is growing in ambition and
capability. The subject has required increased attention to the spatial organization
of subsistence and political economy. Wilkinson also clarifies how an agent based
approach can improve upon traditional methods such as site catchment analysis.
Instead of applying the fixed parameters of catchment analysis to a site, the MASS
project models the activities of individual agents from below. While reading this
article, one can easily envision future simulations that will model certain activities
and predict the presence, volume, and diversity of material residue likely to result
from those activities, providing a powerful tool for middle-range theory.
Stephen Lansing and Robert Axtell generously agreed to review the articles in
this collection and give us their thoughts on the application of complex systems
theory to archaeology. Lansing and Axtell have been at the centre of the maelstrom
for over 15 years, and they discuss the merits of the expanding perspective from
their unique vantage points in complexity research and agent based modelling.
Conclusion
The editors and contributors to this book do not seek to appropriate nonlinear
studies for a specific area within archaeology. The diverse backgrounds of our
contributors – Heterarchy, Social Agency, Information Theory, Factionalism,
Systems Ecology, Landscape studies – should make that point clearly. Each has
ties to one or the other of the theoretical lineages discussed above, and each shows
how nonlinear dynamics might be considered within their own field. There are
differing positions amongst the two editors as well, creating a dynamic tension that
we find stimulating and fruitful. The differences amongst all contributors suggest
that nonlinear systems models have provoked a curious realignment of theoretical
12
Nonlinear Models for Archaeology and Anthropology
traditions quite different from the outmoded Processualist vs. Post-Processualist
divide. We hope that this volume will widen the field of debate over the utility of
nonlinear systems for archaeology, and help other archaeologists evaluate the field
differently than they might if they were only to consider the high profile work
produced within some American institutions. Because, as Stone argues in her
chapter, perception of the approach is as important as the approach itself.
We end our introduction with an anecdote to bring this point home. The
editors originally organized the session that slowly and painfully inspired this
volume for the 2002 American Anthropological Association meetings in New
Orleans. At the time, there was a vocal ‘pro-Science’ splinter group that was
holding its own sessions down the street from the conference hotel, sessions that
had not been accepted into the official AAA programme because (according to the
dissidents) they were too scientific. The group circulated a list of AAA sessions
that did meet their criteria for a scientific approach, and our session was included.
This unsolicited testimonial was bestowed on the basis of no more information
than the title of the session – ‘Nonlinear Systems Approaches in Archaeology’ – all
that was available to the group when they made their pronouncement. Our session
participants showed mixed reactions to the news, and happily our audience
represented a gratifyingly diverse cross-section of our anthropological colleagues.
In the polarized aftermath of the debates within archaeology in the 1980s and
1990s, some archaeologists are willing to declare an approach valid or not on the
most superficial of evidence. We direct this book to those members of the field
who are dedicated to the work, and are still open to potentially effective and
innovative ways of analyzing human society, regardless of what one calls them.