Bronze Age Urbanism Project
I am involved in a study of Bronze Age urbanism in the Southern Levant
with Dr. Steven Falconer. For this project I developed the computer-based
Monte Carlo methods of analyzing rank-size site distributions that are
described in the abstracts section of my page.
These statistical methods address critical issues of coverage and the composition
of the underlying population of sites from which the actual data have been
recovered as a sample of some unknown, but generally estimated, proportion.
Probability estimates are developed empirically, based on a long series
of random runs, in which a group of sites equal to the number actually
found are drawn from a log-nomal site universe that has the characteristics
which the researcher specifies. The number of sites in the data universe
is equal to the number of observed sites, divided by the sample proportion,
and the largest observed site is used as the largest in the data universe.
The results suggest that Levantine coastal plain urbanism is different
in scale, and in kind, than southern Mesopotamian urban expressions. And,
we show that urban expressions are dramatically different among the coastal
plain, central hills, and Jordan valley regions of the Southern Levant.
Far from being simply a case of Mesopotamian urbanism at a reduced scale,
each sub-region of the Southern Levant shows its own unique trajectory
of urban growth and decline. In addition, there are dramatic differences
in the degree of "system integration," or social-economic factors, that
held the urban and rural places together. Our results further suggest that
the rural component of the settlement system may have been organized somewhat
independently of the larger, urban centers.
Through a series of iterations at scales that decrease from the entire
region to small groups of sites, this approach may be combined with cluster
analysis of site locations to explore the growth and spread of urbanization
in different "city-states" in the Levantine Bronze Age. Some of our initial
rank-size analyses indicated that the sub-regions of the Southern Levant
may each have contained several smaller settlement systems in the Early
and Middle Bronze Ages. I used cluster analysis on site locations to separate
the sites in the three sub-regions into smaller groups, and then ran the
Monte-Carlo RankSize simulation on them again. This work is still in progress,
but the results are quite exciting so far.
One difficulty of working with these data from the Early and Middle Bronze
Ages is that there are very few texts from the region in the Middle Bronze
Age, and none in the Early Bronze Age. A nagging question has been whether
our statistical methods can reproduce site clusters that reflect real social-political
units in these remote eras, or whether they are simply statistical creations.
We have developed a test of our spatial-statistical approach that compares
Late Bronze Age settlement data with historical sources, such as the Egyptian
Amarna Texts and the Old Testament. Shlomo Bunimovitz worked out what he
felt were accurate boundaries between city-states in Late Bronze Age Canaan,
based on these and other texts. I have compiled a list of site locations
from the same period, and performed a cluster analysis on them, in the
same way we have done it for the Early and Middle Bronze Ages. Then, by
superimposing the cluster map on the boundaries ascertained by Bunimovitz,
it is possible to see that there is a remarkable correspondence between
the two. Thus, our statistical methods are probably able to recreate a
real social landscape in the earlier, ahistorical periods.
Publications Related to this Research:
|