The Organization and Evolution of the Hohokam Economy Agent-Based
Modeling of Exchange in the Phoenix Basin, Arizona, AD 200-1450
Chapter 1
A NEW PERSPECTIVE ON THE HOHOKAM ECONOMY
The late prehistoric peoples of central Arizona, known today as the Hohokam,
left behind tantalizing evidence of a culture markedly different from – and probably
more complex than – the small communities of O’odham farmers first encountered by
Europeans in the sixteenth century A.D. While archaeologists have worked for well over
a century to document Hohokam culture history, many important details about life in
the Sonoran Desert remain poorly understood. As archaeologists’ methodological and
theoretical toolkits for learning about the various social and environmental processes
that shaped the Hohokam improve each decade, certain interesting topics tend to be
revisited and re-evaluated through new lenses. A perennial favorite, and the focus of this
dissertation, is the organization of the Hohokam economy.
Archaeologically, the Hohokam of the lower Salt River Valley and the middle Gila
River Valley (known collectively as the Phoenix Basin, Figure 1.1) are relatively well
known thanks to decades of many and often large cultural resource management (CRM)
projects. The chronology and settlement patterns are reasonably well documented, and
recent years have seen a florescence of research oriented toward more subtle questions
(Dean 1991). For example, the organization of the Hohokam economy has been an
important focus of several established researchers (Abbott 2000, 2009; Abbott, ed.
2003; Abbott et al. 2007; Bayman 2001; Doyel 1991, 2000; Woodson 2011). Much of this
current research reflects a shift from an emphasis on external and regional trade
networks in the 1970’s and 1980’s towards an effort to understand the organization of
production and exchange within the Phoenix Basin in the 1990’s and 2000’s (e.g., Braun
and Plog 1982; Crown 1991a). There is good evidence that over the course of
1
,approximately seven hundred years some Hohokam adopted specialized production of
Figure 1.1. Map of the Phoenix Basin.
plain and decorated pottery, possibly organized at the community level (Abbott 2009;
Abbott et al. 2007). Likewise, prehistoric residents of the Phoenix Basin also practiced
extensive irrigation agriculture in the Salt and Gila River valleys for most of that time, a
strategy that may have complemented specialized production of pottery and other goods
(Bayman 2001; Watts in prep). Note that throughout this dissertation, I adopt the
convention of using the word "exchange" to refer to transactions that are more
reciprocally or culturally motivated, and the word "trade" is used when discussing
economically motivated transactions. This is consistent with previous use of the terms
(per Plog 1977; Stark 1993).
The research project described in this dissertation was structured around a
program of assessing several working hypotheses about the ceramic sector of the
Hohokam economy. This was accomplished using an agent-based modeling
2
,methodology known as pattern-oriented modeling (Gilbert and Troitzsch 2005; Grimm
et al. 2005; Railsback and Grimm 2012), the first application of this approach to the
prehistoric archaeology of the Southwest United States. The objective of the research
was to first identify a variety of economic models – abstract representations of the real-
world system theoretically drawn from different sources, including microeconomics,
mathematics (network/graph theory), and economic anthropology – that may explain
patterns of artifact distribution in the archaeological record. Next, the effort was turned
toward implementing those hypotheses as agent-based models, and finally assessing
whether or not any of the models were consistent with Hohokam ceramic datasets.
Importantly, the archaeological datasets included in this research were existing
ceramic data assembled from many different projects and contributed by multiple
analysts working in the Phoenix Basin. While I participated in either the field work or lab
analysis on a small number of these projects, the vast majority of the data integrated for
this research was made available by others. It is with deep gratitude I thank David
Abbott, Kathy Henderson, and Mary Ellen Walsh for sharing their datasets and
encouraging me to pursue bigger-picture questions that would be difficult to address
with any single isolated dataset. By revisiting and repurposing existing data, this
dissertation project was very specifically in the “preservationist” tradition, as no new
potentially-damaging fieldwork was needed.
The Problem
Underlying decades of research into prehistoric economies in the Southwest was
the assumption, implicit or explicit, that utilitarian goods were mostly produced by
households for their own use and occasionally exchanged down-the-line within kin
networks, while more-valuable items may have circulated in prestige goods economies
(Bayman 2001, 2003; Crown 1991). Recent analyses (Abbott 2009; Abbott et al. 2007)
led to the proposal that the Sedentary period Hohokam (ca. A.D. 1050-1175) provide an
3
, example of the adoption or evolution of marketplace trading of pottery. Abbott’s
argument that marketplace exchange of occurred in the late Preclassic was an important
acknowledgement that ethnographic and historic analogues from the region may not be
sufficient for understanding economic patterns observed in the archaeological record.
Fundamentally, the problem that framed the current dissertation research was an
extension of Abbott’s criticism. First, generally accepted models seemed to inadequately
explain patterns in the distribution of Hohokam pottery, which necessitated the
development of alternate models. Second, appropriate methodological toolkits for
systematically evaluating alternate models of the Hohokam economy needed to be
identified, developed, and applied.
While Abbott’s (Abbott et al. 2007) argument for periodic marketplaces is focused
on the Middle Sacaton phase, other evidence (Abbott 2009) hints that an efficient and
reliable economic mechanism to distribute pottery from specialist producers to consumers
must have been adopted by the Hohokam, possibly as early as A.D. 500 and almost
certainly by A.D. 650. That mechanism may have been a nascent market-based system, or
any of several competing hypotheses. The current research project, described in much
greater detail below, proceeded by first documenting spatial patterning in the distribution
of Hohokam pottery in the Phoenix Basin. Then computer models provided a virtual
laboratory to generate patterns that were eventually compared against the archaeological
dataset to assess which hypothesized economic systems (if any) may have more likely been
adopted by the Hohokam. Among the models that were tested, I included both naïve
models (e.g., random, complete, nearest neighbor, and scale-free exchange networks) and
theoretically informed models drawn from economic anthropology (e.g., reciprocal
kinship-based networks and centralized redistribution) and microeconomics (market-
based trade, including marketplaces). The agent-based modeling effort contributed to a
4
Modeling of Exchange in the Phoenix Basin, Arizona, AD 200-1450
Chapter 1
A NEW PERSPECTIVE ON THE HOHOKAM ECONOMY
The late prehistoric peoples of central Arizona, known today as the Hohokam,
left behind tantalizing evidence of a culture markedly different from – and probably
more complex than – the small communities of O’odham farmers first encountered by
Europeans in the sixteenth century A.D. While archaeologists have worked for well over
a century to document Hohokam culture history, many important details about life in
the Sonoran Desert remain poorly understood. As archaeologists’ methodological and
theoretical toolkits for learning about the various social and environmental processes
that shaped the Hohokam improve each decade, certain interesting topics tend to be
revisited and re-evaluated through new lenses. A perennial favorite, and the focus of this
dissertation, is the organization of the Hohokam economy.
Archaeologically, the Hohokam of the lower Salt River Valley and the middle Gila
River Valley (known collectively as the Phoenix Basin, Figure 1.1) are relatively well
known thanks to decades of many and often large cultural resource management (CRM)
projects. The chronology and settlement patterns are reasonably well documented, and
recent years have seen a florescence of research oriented toward more subtle questions
(Dean 1991). For example, the organization of the Hohokam economy has been an
important focus of several established researchers (Abbott 2000, 2009; Abbott, ed.
2003; Abbott et al. 2007; Bayman 2001; Doyel 1991, 2000; Woodson 2011). Much of this
current research reflects a shift from an emphasis on external and regional trade
networks in the 1970’s and 1980’s towards an effort to understand the organization of
production and exchange within the Phoenix Basin in the 1990’s and 2000’s (e.g., Braun
and Plog 1982; Crown 1991a). There is good evidence that over the course of
1
,approximately seven hundred years some Hohokam adopted specialized production of
Figure 1.1. Map of the Phoenix Basin.
plain and decorated pottery, possibly organized at the community level (Abbott 2009;
Abbott et al. 2007). Likewise, prehistoric residents of the Phoenix Basin also practiced
extensive irrigation agriculture in the Salt and Gila River valleys for most of that time, a
strategy that may have complemented specialized production of pottery and other goods
(Bayman 2001; Watts in prep). Note that throughout this dissertation, I adopt the
convention of using the word "exchange" to refer to transactions that are more
reciprocally or culturally motivated, and the word "trade" is used when discussing
economically motivated transactions. This is consistent with previous use of the terms
(per Plog 1977; Stark 1993).
The research project described in this dissertation was structured around a
program of assessing several working hypotheses about the ceramic sector of the
Hohokam economy. This was accomplished using an agent-based modeling
2
,methodology known as pattern-oriented modeling (Gilbert and Troitzsch 2005; Grimm
et al. 2005; Railsback and Grimm 2012), the first application of this approach to the
prehistoric archaeology of the Southwest United States. The objective of the research
was to first identify a variety of economic models – abstract representations of the real-
world system theoretically drawn from different sources, including microeconomics,
mathematics (network/graph theory), and economic anthropology – that may explain
patterns of artifact distribution in the archaeological record. Next, the effort was turned
toward implementing those hypotheses as agent-based models, and finally assessing
whether or not any of the models were consistent with Hohokam ceramic datasets.
Importantly, the archaeological datasets included in this research were existing
ceramic data assembled from many different projects and contributed by multiple
analysts working in the Phoenix Basin. While I participated in either the field work or lab
analysis on a small number of these projects, the vast majority of the data integrated for
this research was made available by others. It is with deep gratitude I thank David
Abbott, Kathy Henderson, and Mary Ellen Walsh for sharing their datasets and
encouraging me to pursue bigger-picture questions that would be difficult to address
with any single isolated dataset. By revisiting and repurposing existing data, this
dissertation project was very specifically in the “preservationist” tradition, as no new
potentially-damaging fieldwork was needed.
The Problem
Underlying decades of research into prehistoric economies in the Southwest was
the assumption, implicit or explicit, that utilitarian goods were mostly produced by
households for their own use and occasionally exchanged down-the-line within kin
networks, while more-valuable items may have circulated in prestige goods economies
(Bayman 2001, 2003; Crown 1991). Recent analyses (Abbott 2009; Abbott et al. 2007)
led to the proposal that the Sedentary period Hohokam (ca. A.D. 1050-1175) provide an
3
, example of the adoption or evolution of marketplace trading of pottery. Abbott’s
argument that marketplace exchange of occurred in the late Preclassic was an important
acknowledgement that ethnographic and historic analogues from the region may not be
sufficient for understanding economic patterns observed in the archaeological record.
Fundamentally, the problem that framed the current dissertation research was an
extension of Abbott’s criticism. First, generally accepted models seemed to inadequately
explain patterns in the distribution of Hohokam pottery, which necessitated the
development of alternate models. Second, appropriate methodological toolkits for
systematically evaluating alternate models of the Hohokam economy needed to be
identified, developed, and applied.
While Abbott’s (Abbott et al. 2007) argument for periodic marketplaces is focused
on the Middle Sacaton phase, other evidence (Abbott 2009) hints that an efficient and
reliable economic mechanism to distribute pottery from specialist producers to consumers
must have been adopted by the Hohokam, possibly as early as A.D. 500 and almost
certainly by A.D. 650. That mechanism may have been a nascent market-based system, or
any of several competing hypotheses. The current research project, described in much
greater detail below, proceeded by first documenting spatial patterning in the distribution
of Hohokam pottery in the Phoenix Basin. Then computer models provided a virtual
laboratory to generate patterns that were eventually compared against the archaeological
dataset to assess which hypothesized economic systems (if any) may have more likely been
adopted by the Hohokam. Among the models that were tested, I included both naïve
models (e.g., random, complete, nearest neighbor, and scale-free exchange networks) and
theoretically informed models drawn from economic anthropology (e.g., reciprocal
kinship-based networks and centralized redistribution) and microeconomics (market-
based trade, including marketplaces). The agent-based modeling effort contributed to a
4
