Simultaneities of World-System Development: Cities , Empires and Climate Change Christopher Chase-Dunn Department of Sociology Johns Hopkins University Baltimore, MD. 21218 USA chriscd@ jhu.edu A research proposal to be submitted to
the National Science Foundation Sociology Program August 15, 1995
Project Summary The world-systems perspective is a structural approach for understanding the
developmental trajectories of national societies and the international political economy
as a whole. It uses a longer temporal perspective and a larger spatial scale than most
other theoretical approaches to social change.
Recently the temporal perspective has been deepened even further by scholars who are
analyzing the continuities and differences between the modern intersocietal system and
earlier systems. This proposal outlines a research project that will compare intersocietal
networks over very long periods of time in order to answer questions that have emerged from
recent extensions of the world-system perspective to earlier periods. The main problem addressed by this proposed research is: Do empires and/or cities grow and decline simultaneously in regions that are distant
from one another and, if so, what are the causes of these simultaneities? The study of the relationships between urban growth, empire size and climate is an
old topic, but the comparative world-systems perspective brings new light to it. The unit
of analysis I will use to examine cities, empires and climate change is the PMN
-- the political/military network -- a set of polities that are interacting with one
another by means of alliances and/or warfare. The focus of this research is the problem of "simultaneities."
Andre Gunder Frank (1994) has postulated, for the Eurasian world system as a whole,
the existence of 200 year phases of expansion and contraction beginning in the Bronze Age.
Recent research (Chase-Dunn and Willard 1993; Chase-Dunn and Hall 1994) has found
preliminary support for simultaneities of urban growth and changes in the territorial
size of empires in Near Eastern and Far Eastern PMNs from the middle of the first
millennium BC until the sixteenth century AD. These pilot studies indicate some support
for the hypotheses that important connections linked Eurasia in to a single interactive
system much earlier than most scholars have believed. My research will improve the data
base for examining the simultaneity hypothesis and examine three possible explanations
of this phenomenon. One possibility is that urban growth and the territorial sizes of
empires were affected in both regions by climate changes and their effects on agricultural
production. If climate change was the main cause of simultaneous growth and decline in the
distant regions, interactional factors would be disproven. Evidence of climatic change for
each regional PMN needs to be gathered and compared with the sequences of urban and empire
growth and decline to test this hypothesis. Other possible causes of simultaneous
growth-decline sequences are changes in interactional factors such as long-distance
trade patterns or changes in relations with intervening pastoral nomads. Data on trade,
migrations and warfare with steppe-nomads will be gathered in order to examine these
hypotheses. The world-systems perspective was developed to interpret and explain the developmental trajectories of modern national states and
the international political economy as a whole (Wallerstein 1974, 1979; Shannon 1989). The main concepts are "world-system"
defined as a multicultural economic division of labor, and "core/periphery hierarchy," defined as an intersocietal stratification
based on economic and political/military power. Recent research carried out by social scientists in several disciplines (e.g. ethnography,
archaeology, political science, history and sociology) has extended world-systems concepts to earlier, less-than-global intersocietal systems
( e.g. Abu-Lughod 1989; Collins 1992; Algaze 1993) Andre Gunder Frank and Barry Gills (1993) have argued that the contemporary
global system is similar to, and continuous with, a single world system that was created 5000 years ago when cities and states first
emerged in Mesopotamia. In contrast with those who study local environments and interactions (so-called "splitters"),
Frank and Gills are proponents of an approach that emphasizes large scale systemness based on very long-distance trade.
They are "lumpers."
In order to empirically resolve the debates between splitters and lumpers, Chase-Dunn and Hall (1993) have proposed a
comparative approach to world-systems that redefines key concepts (see below) and proposes a typology of world-systems that
vary from small and egalitarian regional networks of sedentary foragers to the contemporary stratified and complex global
system. It is this comparative conceptual apparatus that will be employed in the research proposed here, but hypotheses
coming from the "continuationist" school (e.g. Frank and Gills) will also be evaluated. Chase-Dunn and Hall (1993) define world-systems as
intersocietal networks in which the interactions (trade, warfare, intermarriage, etc.) are important
for the reproduction of the internal structures of the composite units and importantly affect changes
that occur in these local structures.
They spatially bound world-systems by analyzing local, regional and inter-regional interaction networks
of three kinds:
Chase-Dunn and Hall's approach is group-centric rather than system-centric. For any focal locality it asks the question: what are the spatial networks of direct and indirect interaction that compose the whole systemic entity that affects this group? The notion of fall-off is used to bound networks rather than interaction densities because it is noted that all societies interact importantly with their immediate neighbors. But for each, the consequences of interaction fall-off with distance. Small world-systems were small, not because there were major discontinuities in interaction networks, but because the fall-off of systemic consequences was far greater than in later systems in which communications and transportation technologies made long distance interaction more frequent and more important.
case studies of small-scale world-systems have helped to further refine
the conceptual apparatus for comparison of different types of systems in
order to empirically work out the important similarities and differences.1
While it would certainly be worthwhile to undertake additional case studies,
the research proposed here moves on to the next step -- the formal comparison
of several world-systems to one another. One important feature of
many systems is their multicentric structure in which non-adjacent core
regions interact with one another vis a vis long distance trade.
These are cases in which separate PMNs are included within a single larger
PGN as indicated in Figure 1. This is a phenomenon that we shall investigate
in this study of territorial empires, urban growth and climate change.
Regarding spatial inter-connections, the long-run history of the modern world-system needs to be specified as the expansion and engulfment of all other systems by the Central World-System. Engulfment of other systems generally proceeded in stages. The expanding central system generally linked first with a system external to itself by means of the exchange of prestige goods. Later the new system was incorporated into the Central PMN and, some time after that, in to the Central BGN. David Wilkinson's (1987:32) chronograph depicts the expansion of the Central PMN as it engulfs thirteen other PMNs (See Figure 2). Wilkinson (1992, 1993) has provided a careful temporal and spatial bounding of the expansion of the Central PMN and the Central PGN.
Figure 2: Chronograph of PMNs
It may be the case that most of the important processes of historical evolution go on within PMNs. At least for the purposes of studying city systems -- the networks of settlements, towns and cities that compose the great state-based civilizations -- the PMN would seem to be an ideal focal unit of analysis. The PMN is certainly superior to either single states or civilizations as a unit of analysis. Single states are virtually always in important interaction with other states and this influences their settlement patterns greatly. The cultural boundaries of "civilizations" do not usually correspond to the boundaries of interaction networks because important interactions frequently occur across big cultural divides. I shall provisionally adopt the PMN as a main unit of analysis in this research while investigating the possibility that important interactions may also occur between separate PMNs that are part of the same PGN.
Frederick Teggart's (1939) Rome and China traced contemporaneous historical events in the Roman and Han empires that demonstrated important economic connections between these distant states. Andre Gunder Frank (1993) has hypothesized the existence of cycles of expansion and contraction that began in the Bronze Age. These alleged cycles have growth and decline phases ("economic good and bad times") lasting about 200 years and they supposedly are features of the whole Afro-eurasian world system as early as 1700 BC. Frank contends that Afro-Eurasia has been a single world-system for 5000 years since the emergence of cities and states in Mesopotamia. It is likely that Frank overemphasizes the degree of integration and the spatial scale of systemic interaction in the earlier centuries. He does not distinguish between BGNs, PMNs and PGNs. We may refine his hypothesis thus: when separate PMNs become importantly linked in to a single larger PGN they will experience nearly simultaneous periods of economic, political and population growth and these simultaneities will be due to human interactional factors such as trade.
Pilot studies have already revealed fascinatingly high correlations between changes in city-size distributions and changes in the territorial size of empires for the Near Eastern and Far Eastern PMNs from the middle of the first millennium through the sixteenth century AD (Chase-Dunn and Willard 1993; Chase-Dunn and Hall 1994). Figure 3 shows the relationship between Near Eastern (Central) and Far Eastern empire sizes from 1500 BC to 1750 AD. The Pearson's r correlation coefficient is .90 for sixty time points.
It has also been found that the intermediate Indic PMN did not experience a similar sequence of growth and decline phases. The data extension and improvement proposed herein will make it possible to better evaluate these preliminary findings. If they hold up I will investigate several hypotheses about what the linkages are that account for these simultaneities. I will compare the sequences of growth with data on climate change. It is possible that climate changes explain the similar timing of growth and decline in the Near East and China. India, at a more equatorial latitude, may have experienced a very different climatic sequence. Climate change can affect urban growth and empire-formation through its affects on agricultural productivity (Nix 1985). Periods of flooding may disrupt irrigation systems, and periods of drought also may negatively affect agriculture. Recently acquired evidence indicates that the collapse of Mayan states may have been caused by an extended period of drought. Weiss et.al. (1993) contend that both the expansion and collapse of the Akkadian empire were spurred by climate changes.
If we should find significant relationships between indicators of climate change and the urban and empire growth/decline sequences we will want to examine the important issue of the direction of causality. Does climate change cause urban change or does the expansion of agriculture associated with urban growth cause climate change? It is possible that expanded agricultural activity and/or deforestation due to human exploitation of forest resources may have affects on local and regional rainfall patterns and ground water levels. Thus intense agriculture and forest exploitation related to population density (and thus urbanization) may affect climate change. There is a huge developing literature on the anthropogenic causes of climate change. It is well-known that the intensification of productive activities cause environmental degradation and that this has been a major process affecting the development of human societies from the very beginning.
is a possibility that we can also evaluate with our data on climate change,
urban growth and empire size. If urban growth episodes precede climate
change or changes in water levels then causality in the direction of human
effects on climate will be supported.
interesting explanation from the world-systems perspective is Frank's (1992)
hypothesis of the "centrality of Central Asia" as a peripheral region linking
both ends of the Eurasian continent. It is well-known that the Mongol Empire
briefly linked the Near East and China into a single polity in the thirteenth
century AD. Owen Lattimore (1940) was the first to observe the tight
core/periphery interaction between the horse pastoralists of Central Asia
and the agrarian Chinese empires. Thomas Barfield (1989) traces the long-term
linkage of the rise and fall of steppe empires with the rise and fall of
agrarian empires in China. Frank (1992) contends that processes of peripheral
migration and steppe-empire formation and their affects on the long distance
trade carried along the Silk Roads of Central Asia are the explanation
of the event simultaneities found by Teggart and also account for Frank's
hypothesized 200 year phases of growth and decline. While our research
(Chase-Dunn and Hall 1994: Figures 4 and 5) indicates only mixed support
for the timing of expansion and contraction phases as hypothesized by Frank
(1993), we do find that the sequences of growth and decline of the Near
Eastern and Far Eastern PMNs track quite closely (Figure 3 above
and others in Chase-Dunn and Hall 1994).
Perhaps it is Frank's Central Asian linkage that accounts for this. In order to find additional support for this hypothesis we need to rule out the climatic hypothesis by gathering data on climate change for the relevant regions, and to compare closely the data on long-distance trade, warfare, migrations and steppe-empire formation to sort out the causes of the simultaneities. We also need to understand why India was not affected in the same way by these processes.
I will also test the simultaneity hypothesis for other cases in which PMNs were linked to one another inside larger PGNs. The study will compare the Egyptian and Mesopotamian PMNs before they became linked in 1500 BC. I will also examine city and empire size data for the Mexican and Mayan PMNs while they were part of the Mesoamerican PGN and for the Mesoamerican and Peruvian PMNs before they were brought in to the Central PMN by the Spanish conquistadors.
addition to growth and decline sequences of the largest empires and
cities, I will also study changes in the distribution of city-sizes
within PMNs and changes in the distribution of empire sizes within PMNs.
The size distribution measures indicate the degree of spatial and power
hierarchies within city systems and interstate systems composed of empires
and states. The simultaneity hypothesis can be posed simply in terms of
the largest cities and empires within each PMN, but pilot studies have
indicated that rises and falls of hierarchy as indicated by changes in
city-size distributions also show simultaneity between the Near Eastern
and Far Eastern PMNs (Chase-Dunn and Willard 1993).
simple size and size distribution measures can be constructed from data
on the population sizes of cities and the territorial sizes of empires.
The basic strategy is to examine urban growth and changes in the territorial
size of empires in eleven PMNs over the last four thousand years. This
will involve utilizing and extending two compilations of data: Chandler's
(1987) Four Thousand Years of Urban Growth and Taagepera's (1978a;1978b;1979;1986)
studies of the territorial sizes of empires.
Tertius Chandler's (1987) excellent compendium contains population estimates for the largest cities on Earth since the beginnings of urbanization in Mesopotamia. Chandler's 1987 publication is a much improved and extended version of an earlier work (Chandler and Fox 1974). Chandler's data set is a valuable resource based on decades of research. For my purposes, however, it needs to be extended to cover all cities that have 10,000 or more people. And the time points are too far apart before 1000 AD. I propose to upgrade and expand Chandler's data set by:
I will update Chandler's city populations using recently published studies, expand the coverage to cities with 10,000 in
population and gather data for 50 year time points for the centuries before 1000 AD. For many time points Chandler presents
lists of cities ranked in terms of population size but without numerical population estimates. It will be possible to use new
studies to make numerical estimates for some of the ranked cities. It is also possible to interpolate some of these numerical
estimates based on the sizes of cities in the same region (e.g. Chase-Dunn and Willard 1993). In some regional PMNs these
will be important for estimating the city-size distributions. There are two reasons to expand the data set to include all cities that have a population of ten thousand or more.
Wilkinson (1992a) uses a method of measuring urban growth rates based on counting the number of cities in each region.
This is a useful addition to our arsenal of measures of urban growth, but there is a problem basing this measure on
Chandler's data set because the minimum size of the cities included changes over time. Expanding the coverage of
later time periods to all cities over 10,000 would make this measure more meaningful. It is also desirable to know
about smaller cities when we are looking for intervening cities between PMNs. And recall also that the time points
included by Chandler are too far apart for the earlier millennia. These will be filled in to every 50 years to match the
empire size data. This job of data enhancement and expansion will utilize substantially the same methods that
Chandler employs (Chandler 1987: 2-13). He uses several methods for estimating the population sizes of
cities -- census figures, traveler's estimates, size of the built up area, hearth counts, the size of the military, etc.
Chandler will serve as a consultant on this project and I will visit with him for a week in Berkeley to obtain
information and discuss measurement problems. I will do a search for all the literature that was not available
to Chandler for his 1984 compilation. I will also consult with him about interpolation methods. My project will
code data about the location of cities (longitude and latitude) so that our data will easily fit in to a larger world
history Geographical Information System (GIS). Empire Sizes The second data set I will work with is the territorial sizes of empires assembled by Rein Taagepera (1978a;1978b;1979;1986).
Taagepera used maps from atlases to estimate the territorial sizes of empires over the past four millennia. His estimates are for
every 50 years, but his data set has unfortunate holes because he was only concerned with the very largest empires on Earth.
Thus he excludes important regional empires that are needed when we are using PMNs as the unit of analysis and for the purposes
of constructing size distributions of empires to measure power concentration/dispersion. Taagepera's data on the territorial sizes of empires will be updated by using more recent sources than those that were
available when he compiled his data set. And coverage will be extended to empires that Taagepera excluded because they
were not among the largest on Earth at a particular time point. We also will be able to add data for Mesoamerican and
Peruvian empires to fill out our PMN cases there. In addition, I will add data on the timing of the events that led to empire
expansion and contraction. Usually these were significant wars of conquest. The dating of such events will be useful for
considering the relationship between empire size change and changes in climate. Taagepera will serve as a consultant to the project. I will visit him at the University of California-Irvine to discuss
his methods of estimating the territorial sizes of empires and to examine the files from which he produced his data set.
He will also provide suggestions about maps and atlases to be used in extending the coverage of the empire size data set. Climate In order to examine climatic variation as a cause of simultaneities it will be necessary to obtain regional data on
climate change for those regions that are found to have high simultaneity correlations with other regions. The main case
we will focus on in this regard is the Near East and the Far East. But I will also gather climate change data for all the eleven
PMNs on which I will have city population and empire size data. Comparable climate change data for the Near East and China over the relevant time periods and for a representative
sample of locations within each region may be difficult to locate. Important local variations in climate change are
well-known, so it is inferentially dangerous to rely on climatological proxy data from only a few sites. The relevant
aspects of climate change are the amount of average yearly rain fall, catastrophic flood events, and the average temperature.
These affect agricultural productivity and may have indirect effects on urban growth and empire size. One important method
for historical estimations of these variables involves the study of tree rings (dendro-chronology). Some tree ring chronologies
are being built up for China (Sheu et.al. 1994; Graumlich n.d.) but their are as yet no dendrochronological sequences for
Mesopotamia. Studies of pollen deposition patterns in lake sediments are used to infer climate change. Mesopotamian
climate change is being inferred from Persian Gulf sediments and from stratified soils (Nutzel 1976; Weiss et.al. 1993).
Glacial ice cores are also used to infer changes in precipitation. While the urbanized regions we are studying are
generally not close enough to glaciers to make this kind of evidence useful, we also want to know about weather
changes in intervening peripheral (or semiperipheral) areas. Paleoclimatologists have also used weather diaries
(verbal descriptions of weather events) to reconstruct climate sequences and to verify other proxy indicators of
climate change (e.g. Zhang and Crowley 1989; Wang and Wang 1994). Documentary evidence about the frequency
and location of devastating floods or droughts may indeed be helpful, but it will be desirable to locate less subjective
indicators (such as the Nile height measurements mentioned in Footnote #9 above) whenever possible. I will consult
with experts on climate change at the World Data Center for Paleoclimatology in Boulder, Colorado and at the
Laboratory for Tree Ring Research at the University of Arizona. Wars Warfare is a human interaction variable that is known to affect both urban growth and the territorial size of empires.
The hypothesis about processes of steppe empire formation and the migration of pastoral nomads out of Central Asia
being the key to simultaneous rise and fall of agrarian empires at both ends of the Eurasian land mass could be supported
if we find simultaneous increases and decreases in warfare between steppe nomads and agrarian states in both the Near East
and the Far East. Thomas Barfield's (1989) Perilous Frontier provides the information for the Far Eastern region.
For the Near East I will use data on warfare obtained by the LORANOW project (Cioffi-Revilla 1991;1994). Trade Ideally we would like to have comparable and reliable data on cross-Eurasian trade by both the Silk Road
and maritime routes over the whole period that we are studying. In practice trade data are quite piecemeal. Nevertheless
I will probably be able to obtain enough data to make rough estimates of changes over time in the amount of trade
flowing over the different routes linking China, India and the Near East. Price data are also useful.
Frank and Gills (1995) have recently argued the importance of the balance of payments in influencing the growth
and decline of states and regions and as an indicator of core vs. peripheral status. As demonstrated by Braudel (1979),
simultaneous and corresponding changes in prices over time in different localities are themselves an important
indicator of systemness. Yearly price data from China and the Near East will enable us to examine the simultaneity
hypothesis on a finer time scale than the city population and empire size data. This will also be helpful in evaluating
the relationship between economic activities and climate change. Measures The basic data on empire and city sizes will be used to produce five different measures. With the city data
we will construct three measures of different aspects of change in city systems. The first measure will be the estimated
population size of the largest city. The second measure will examine the shape of the city size distribution
-- the relative sizes of the five largest cities. For this we will use the Standardized Primacy Index (SPI) developed
by Pamela Walters (1985). City systems vary with regard to the steepness or flatness of the city size hierarchy.
A very flat city size distribution is one in which the largest cities are all about the same size. A steep city size
distribution corresponds to what is referred to in the settlement system literature as "urban primacy."
This means that the largest city is much larger than the other cities in the system. The SPI compares the actual
distribution of city population sizes with a hypothetical rank-size hierarchy in which the largest city is twice
the size of the second largest, the third largest is one third the size of the largest, and so forth. Deviations toward
flatness from the rank-size norm are assigned negative scores, while deviations toward steepness are assigned positive
values that increase with the steepness of the city size distribution. A third measure of urban systems is the
number of cities within a region that are larger than a certain cut-off size. This can be understood as a measure
of city density. We will improve upon David Wilkinson's(1992a) measure based on the number of cities by adding
all cities larger than 10,000 to the data set. Wilkinson's measure was based on the cities listed in Chandler's compendium.
Chandler's lists drop smaller cities as the number of larger cities increase. The problem here is that a change in the
number of cities may reflect changes in the lower cutoff rather than a real change of the number of cities within
a region. Having a constant cutoff of 10,000 eliminates this problem. Two measures will be computed using the territorial sizes of empires. The simplest is the size of the largest empire.
We will also compute a size distribution of empires using our territorial size data on the three largest polities.
We will use the SPI to compute this measure. This will provide an indicator of the relative degree of centralization
of power in regional state systems. In order to analyze the direction of causality vis a vis changes in empire size and
climate I will also construct a measure of empire size change that codes the month (or year) of key events such as
wars of conquest that resulted in changes in empire size. This will require going back to Taagepera's original files
to determine the timing of events that led to changes in empire sizes. Analysis The main method of data analysis will involve time series correlations and regressions within and across PMNs.
It will likely be necessary to detrend these variables in order to examine whether or not short term rises and falls are
correlated. The geometric rises of everything in the nineteenth century will overcome all other relationships. Another
way of getting around this problem is to exclude the years after 1800 AD. In order to evaluate the simultaneity hypotheses we need to correlate rises and falls in the same variables for
different PMNs. This is done by combining the PMNs in the same data set and calculating the temporal correlations (e.g. Figure 3).
We may also want to examine hypotheses about time lags in relationships or loosen the criterion of "simultaneity"
somewhat by widening the time periods. It will not be possible to calculate meaningful moving averages because
we have 50 year intervals and it will not be feasible to find data for smaller intervals of time, except for the climate measures. Recall that we also want to examine the question of the direction of possible causal interaction between climate
change and urban and/or empire growth and decline. In principle the issue of the direction of causality can be determined
from time series data using the procedure of Granger tests of antecedence. But Granger causality analysis is usually applied
to yearly data (e.g. Rasler and Thompson 1994: 93). Our fifty-year time points are too crude for the determination of antecedence.
If we find significant correlations between climate measures and urban or empire growth/decline phases we will need
to look at a finer time scale in order to employ Granger tests of antecedence. This can be done for the empire size variable
by using the dates of events such as wars of conquest in which empires expanded or contracted. This project will improve existing data sets on empires and cities that are important for our understanding of
world history. These data sets will be constructed so as to fit easily in to a larger world history GIS (geographical information system).
This study will produce improved explanations and evidence regarding the causes of urban and empire growth and decline,
and the interactive relations between distant regions. It will examine the hypothesis of climatological effects on social change.
And it will have important implications for some of the main conceptual and theoretical problems at the roots of the
comparative world-systems approach.