Project
Background
The Four Corners area of the U.S. Southwest contains
many remnants of a dense pre-Hispanic occupation by ancestral Pueblo farmers. A farming way of life here
dates back to before 1000 B.C. (Cordell
1997:127-151). Substantial villages with both pitstructures
and surface structures appear around A.D. 750, but the area was largely depopulated
again in the A.D. 890-930 period. After a second population expansion, the
area was depopulated in favor of locations to the southeast and south in the
late 1200s (Cameron 1995). In this
proposal we deal with an important subset of this area, comprising most of
what Varien et al. (1996:86) defined
as the McElmo/Yellowjacket District, an area sometimes called the breadbasket
of the Anasazi world. During the last 700 years
of its occupation, from A.D. 600 to 1300, the settlement and subsistence systems
of our study area underwent several well-documented transformations. There
were episodes of population aggregation and partial disaggregation
for which the general causes are under debate (Adler
1990; Cordell et al. 1994; Leonard
and Reed 1993; Stone 1992). There
is a shift from rainfall-dependent field systems towards more water control
by the A.D. 1100s (Schlanger
1988), with sediment and water-controlling features becoming gradually more
common after about A.D. 940 (Wilshusen
1997). There is an increasing restriction of large settlements to canyon heads
in the last 40-50 years of occupation (Varien
et al. 1996:106). There is, of course, the final depopulation by pre-Hispanic
farmers, accompanied by signs of strife (Lipe
1995; Kuckelman
et al. 2000).
This history cannot
be understood without reference to population growth (Kohler
and Sebastian 1996; Wilshusen
with Varien 1996), forest reduction (Kohler
and Matthews 1988), high- and low-frequency climate change affecting opportunities
for agriculture (Ahlstrom
et al. 1995; Petersen 1988), repositioning
of fertile topsoils downstream with erosion (Force
and Howell 1997), warfare, especially during the last few decades of occupation
(Wilcox and Haas 1994), and changing
extent and activity of networks of reciprocity driven by changing patterns
of agricultural productivity (Kohler
and Van West 1996). We will also suggest that changing distributions of
water sources during the last 200 years of occupation affected settlement
distributions. It has proven extremely difficult to understand how these factors
interacted to produce the behaviors instantiated in the archaeological record,
generally regarded as the best known in the world (Wenke
1990:578).
By most measurements
of cultural complexity, these societies are fairly simple. Their analysis,
however, presents serious dilemmas that confront the systematic analysis of
any similar system, contemporary or prehistoric. How can we study the simultaneous
interactions of a population with a set of resources that is changing both
in response to exogenous, climate-induced factors on several time scales,
and anthropogenic effects? How can we, at the same time, account for the effects
of the changing distributions of population itself on economic opportunities
for the actors in this system? This long and detailed record provides our
best opportunity to expand our abilities to understand human adaptation in
small-scale societies over long periods on high-dimensional landscapes, using
the techniques and tools described below.
Agent-based modeling languages such as Swarm
(Swarm Development Group
2000) make it possible to effectively study systems composed of many interacting
distributed processes. From a social sciences perspective, the primary benefits
of this research are to:
·
Refine an existing database of site
location, size, function, and chronology for this area, to the benefit of
all regional archaeologists.
·
Estimate the extent to which flow
quantities and locations of water sources changed during the late pre-Hispanic
occupation, and assess the importance of this effect for settlement systems
and for the depopulation of the area.
·
Determine whether the episodes of
aggregation and disaggregation, and the locations
of the aggregates in this area, can be accounted for by variability in exchange
practices, in the context of the other variables affecting settlement practices.
We will begin by examining the performance of a self-interested form of variance-reducing
reciprocal exchange described by Kohler
and Van West (1996). We will examine against this a variant of the strong
reciprocity model for cooperation (Bowles
and Gintis 2001), which depends on group norms
enforced (we suggest, for our case) by exclusion from exchange networks.
·
Make simultaneous tests of the interactions
of economic factors, hydrological factors, possible depletion effects, and
changing agricultural strategies on the settlement system, while at the same
time examining the effects of the evolving settlement system on the environment,
including the hydrological system.*
From a hydrological perspective,
the work we propose contributes to these goals, and to the study of paleohydrological
regimes in general, by:
·
developing interactive, computer-based
analytical and mathematical tools within a Geographic Information System (GIS),
including nested ground-water and surface-water flow models appropriate for
evaluating watershed-scale water supply, useful for anticipating the behavior
of these systems under future climatic regimes; and
·
identifying the relative roles of
human-induced and climatic processes in controlling the modern and paleoground-water and surface-water of Southwest Colorado.
Finally, this work contributes to
developing multi-agent systems and cultural algorithms (Reynolds
1994). The integration of multi-agent modeling with evolutionary learning
based upon theories of cultural change affords us the opportunity to observe
the knowledge and social structures that emerge in various situations. The
Swarm libraries that Reynolds develops to implement cultural algorithms will
be useful to the large and growing community of Swarm users. This work furthers
computational organizational research in general by confronting artificial
society models with data that can serve to either calibrate or assess their
performance. Taken together, these elements constitute a unified framework
for studying the linked social and environmental processes affecting a small-scale
agricultural system. Although each component separately has merit, we consider
their integration to be the most important product of the proposed research.