A Theoretical Introduction
to 5,000 Years of World System
History
Andre Gunder Frank
No history can be written but universal history.
—Ranke
During this period, history becomes, so to speak, an or-
ganic whole . . . all cvents culminating in a single result... .
Separate histories must be regarded as of very little use
in arriving at a realistic conception of the total picture.
For it is only by exposing side by side the threads that con-
nect each event with the whole complex, and also by point-
ing out the resemblances and differences, that it becomes
possible to achieve this.
—Polybius
New Peoples. . . transferred the centre of civilization from
the Mediterranean to the Atlantic Ocean. The distinction
between ancient and modern history, therefore, is no mere
accident, nor yet a mere matter of chronological conven-
ience. What is called modern history is, in reality, the for-
mation of a new cycle of culture, connecting several epochs
of its development with the perishing or perished civiliza-
tion of the Mediterranean states.
—Mommsen
Whoever is the lord of Malacca has his hands on the throat
of Venice.
—Tomé Pires
REVIEW, XIII, 2, SPRING 1990, 155-248 135156
Andre Gunder Frank
What has happened is a sensc of remarkable complexity
in the interplay between the Occident and East Asia. . . .
The Old World at least has long lived in a more unified
realm of discourse than we have been prepared to admit.
—Lynn White, Jr.
Few scholarly tasks are more urgent than that of learning
to see the various historical backgrounds of our common
world in relation to each other.
The point is that from a world-historical point of view, what
is important is not European history in itself, however im-
portant that be for us all, but its role in interregional his-
tory. . . . The problem of reorienting ourselves to a more
interregional viewpoint, then, is psychologically far-
reaching, and must be solved along with that of organiz-
ing the historical material.
During the last three thousand years there has been one
zone, possessing to some degree a common history, which
has been so inclusive that its study must take a prepon-
derant place in any possible world historical investiga-
tion. . . . The various lands of urbanized, literate civilization
in the eastern Hemisphere, in a continuous zone from the
Atlantic to the Pacific, have been in commercial and com-
monly in intellectual contact with each other, mediately
or immediately. Not only have the bulk of mankind lived
in this zone, but its influence has emanated into much of
the rest of the world.
[In] the following approach . . . events may be dealt with
in their relation to the total constellation of historical for-
ces of which they are a part—a method not limited to world
history, but perhaps likely to be especially appropriate in
this case. This means that we are to consider how events
reflect interdependent interregional developments. . . .
What is important is the recognition . . . that there has
been some sort of developing pattern in which all these45,000 YEARS OF WORLD SYSTEM HISTORY 157
interregional developments can be studied, as they are
affected by and in turn affect its elements as constituted
at any one time.
—Marshall G. S. Hodgson (1954, 1958)
I. CONCEPTUAL INTRODUCTION
1. Thesis
y thesis is that our present world system extends back
Jong before ap 1500 to 2500 sc and probably earlier. I believe
that we usefully can and should (re)read history this way and I intend
to support this thesis in two steps. The present “Theoretical Introduc-
tion” critically examines some other contemporary world-historical and
macrosociological approaches to build up a prima facie case for my
thesis. I hope, in a subsequent “Historical Review? to (reread and
(re)write selected empirical and analytical studies of history to support
my thesis with historical materials.
I believe it can and should be shown that there was—and still is—a
millennial systemic connection and unity of development in what has
been variously called the “world island,’ “ecumene,’ and “central civ-
ilization? which stretches from eastern Asia to western Europe and
southern Africa. The Americas and Australasia were previously affected
and later incorporated into this same one world system. I also wish
to stress the importance of the oft-neglected centrality of central and
inner Asia in the unfolding of this one world system history over some
five millennia.
2. Relevance and Purpose
This reading of history has contemporary cultural, political and
theoretical or analytical relevance. Culturally, it is more humanocentric,
that is, non- (anti-) Euro/Western-centric. Politically and analytically,
this reading intervenes in several contemporary political and theore-
tical debates:
1. Euro- and Western-centrism and its Islamo-, Sino-, Indo-
centric and other alternatives;
2. Political and ideological debates about alternative political158 Andre Gunder Frank
“systems? like capitalism, socialism, and none of the above;
3. Disputes about what kind of history to read and write, and
how to do it;
4, Ideological and theoretical debates about the primacy of cul-
tural and ideological, political and military (power), eco-
nomic and technological, social structural (including gender),
and physical and environmental factors and their relations
in the past and present;
5. Theoretical debates about the meaning of “society; “nation,”
world economy and system, mode of production; and na-
tional, civilizational, comparative, and world (systems) his-
tory, as well as social evolution and transformation, etc.;
6. Philosophical debates about materialism and idealism, and
determinism and free will.
I shall comment on some of these issues, and explicitly or implic-
itly intervene below. .
Euro- and Western-centrism (Eurocentrism for short) has been a
problem for a long time, both in the West and, through its cultural
colonialism, elsewhere. Increasingly, more objections to Eurocentrism
have recently been voiced, and so have calls for its replacement.
UNESCO has been active in this task, and recently alternatives to
Eurocentrism have had some impact within the educational establish-
ment in the United States (though mostly because the Pacific theater
and Asian immigrants are increasingly more important than Euro-
pean ones). Islamic revivalism on the right is directed against the influ-
ence of the West, while on the left the Egyptian Samir Amin (1989)
has recently published a book directed against Eurocentrism, to which
I shall return.
For a Westerner, it is extremely difficult to escape Western-centrism.
One problem is the constant Western temptation to address— even if
to give different answers to—questions previously posed by Western-
ers, which in a global context are of relatively minor importance. An-
other problem is the dependence on Western and Western-language
sources, which gives a Western bias to the selection and interpretation
of historical sources and problems. Both Western biases will be more
than evident in my attempt to write a non- or at least less Euro/Western-
centric history.5,000 YEARS OF WORLD SYSTEM HISTORY 159
Il. HISTORICAL PROCEDURES
1. Procedures
My procedure will be first theoretical or conceptual, and then his-
torical. This paper undertakes the first theoretical step to argue that it
makes sense to pursue the world system much farther back than the
year aD 1500, the date at which some (I now believe mistakenly) see
the birth of our modern world system. In Part III, I mostly dispute
as unfounded scholarship which neglects the possibility of existence
of this world system before ap 1500. To this end also, Section 3.2.2.
especially, and Section 3.3.1. to a lesser extent, examine the concept
of the “world system” and its historical continuity over time, its geo-
graphical extension over space, and some of its constituent economic
connections, political relations, and cultural diffusions.
Sections 3.3.2. and 3.3.3. critically accompany writings that, de-
spite their authors’ possible reservations about world system history, I
find most useful as a guide. Part IV offers my provisional theoretical
conclusions and conceptual proposals toward writing a historical-
materialist political economy of world system history. This conclusion
is complemented by a derivative paper with Barry K. Gills (Frank &
Gills, 1989), tentatively entitled “The Cumulation of Accumulation”
which extends and deepens our conception of the process of accum-
ulation in the world system and how to study it. Subsequently, I in-
tend to write one or more studies of the history of the world system,
tracing it as far back as the evidence and my analysis of it permit.
2. Timing
Reading and writing world system history poses some procedural
problems of timing, sequencing, and periodization. It is well to pose
them in this theoretical introduction, although I may postpone their
practical resolution to the subsequent historical review.
One problem is that of historical time and timing. “World” histo-
rians, comparative civilizationists, and social scientists who confine
themselves to comparing one place and time with another, or to study-
ing one problem or institution like power or the state, or culture or
religion, can define their historical timing to fit their chosen topics and
problems. A world system history, of the world system as a whole, how-160 Andre Gunder Frank
ever, requires single and scquential timing for the whole and all of
its parts. This problem is illustrated by a dispute I had with Perry An-
derson in an earlier book (Frank, 1978b). He rejects explicitly —and
many other historians reject de facto—what he calls “chronological mo-
nism,’ in which “events or institutions appear to bathe in a more or
less continuous and homogencous temporality” (Anderson, 1974: 10).
Therefore, in his Lineages of the Absolutist State,
there is no such thing as a uniform temporal medium: for the
times of the major Absolutism . . . were precisely, enormously
diverse . . . no single temporality covers it. . . . [On the other
hand, some] fundamental phenomena fall well within the formal
compass of the “periods” treated here, as contemporaneous . . .
[but although] their dates are the same, their times are sepa-
rate (Anderson, 1974b: 10).
In methodological opposition to this, I already wrote in the mid-1970's
that:
Anderson’s apparent attempt to make historiographic virtue out
of empirical necessity when he argues that the historical times
of events are different though their dates may be the same must
be received with the greatest of care—and alarm. For however
useful it may be [comparatively] to relate the same thing through
different times, the essential (because it is the most necessary
and the least accomplished) contribution of thc historian to his-
torical understanding is successively to relate different things
and places at the same time in the historical process. The very
attempt to examine and relate the simultaneity of different events
in the whole historical process or in the transformation of the
whole system—even if for want of empirical information or
theoretical adequacy it may be full of holes in its factual cov-
erage of space and time—is a significant step in the right di-
rection (particularly at a time in which this generation must
“rewrite history” to meet its need for historical perspective and
understanding of the single world historical process in the world
today).
‘Throughout this book the attempt is made to examine succes-
sively historical events and to analyze the processual or systemic5,000 YEARS OF WORLD SYSTEM HISTORY 161
connection between them in different parts of the world af the
same time, that is, in the same decade or year, and particularly
in the same period of crisis. Implicit, if not explicit, in this ap-
proach is the supposition of the existence of systemic connec-
tions in a single historical process by the sixteenth century, if
not earlier (Frank, 1978b: 20-22),
I may briefly illustrate this reading of history with an example from
this earlier book. I argued that events in India, and elsewhere in the
world, directly and materially contributed to initiating the American
Revolution. In general, I sought to show how the development of the
world economic crisis since 1762, through successive recessions cul-
minating in that of 1773, rallied support to the cause of independence.
More specifically, the 1774 Quebec and Intolerable Acts and “taxation
without representation” are well known. Less known or forgotten is
that they were a response to the recession and the Boston Tea Party.
The general economic crisis of the 1760's strengthened the hand of
the East India versus the West India interests in London and promoted
the East India Company's “Rape of Bengal? This in turn occasioned
the Bengal Famine of 1770-72, which nearly bankrupted the East In-
dia Company. Therefore, it sought redress in Parliament in London
and was newly granted the American market in the Tea Act of 1773.
After shipping tea to Boston with monopoly privileges that went against
the interests of local merchants, these merchants had a “Tea Party”
and dumped the tea into Boston Harbor. The Quebec and Intolerable
Acts were the British punitive response. The April 19, 1775, Lexington
and Concord “shot that was heard around the world” and the July 4,
1776, Declaration of Independence were the American responses in
turn. Thus, India contributed materially and directly to the American
Revolution through the worldwide historical process and its connec-
tions in a single world system.
Pursuing history in a single world system back many centuries and
millennia, I will have to be similarly “monistic” in timing and inter-
relating events and developments around the world. However, I may
have to lengthen my “simultaneous” timing a bit to inquire into other
historical connections. For instance, Beckwith challenges the common
interpretation of the establishment of the Carolingian dynasty in 751
as an isolated event. Beckwith observes that in the thirteen year period162 Andre Gunder Frank
between 742 and 755 there were “fundamental changes, usually sig-
naled by successful political revolts, in every Eurasian empire. Most
famous among them are the Carolingian, Abbasid, Uyggur Turkic,
and anti‘T’ang rebellions. . . . Significantly, all seem to have been in-
timately connected with central Eurasia” (1987: 192-93). Beckwith sug-
gest that the simultaneous political event in “Western Europe was not
coincidental,” and observes at the same time that “it occurred during
an economic depression in Europe? and that “the importance of the
international trade routes through central Eurasia cannot be overem-
phasized” (1987: 192-93), Earlier also, was the decline of Han China
and Rome related to their mutual trade (and if so how), and was their
decline related as well to the subsequent development of first Persia
and then India, which had benefited the most from this trade as mid-
dlemen? And later, (how) were developments in Ming China and Saf-
avid Persia common responses to the preceding Mongol invasions and
retreats?
3. Sequencing
A second procedural problem is that of finding and sequencing
the historical period(s) to be studied. Selecting any particular starting
date is likely to be either arbitrary or conventional (which may also
be arbitrary). Or, the date selected may follow some preconceived no-
tion of when what we are studying is supposed to have started (which
is likely to be both conventional and arbitrary). As we will observe
below, for instance, world system history is now “conventionally” started
around 1500, or more “precisely” in 1492 by Frank (1978b), 1494 by
Modelski (1987), and more generously about 1450 by Wallerstein (1974).
But how can we “really” know when it all began? The American his-
torian of China, John King Fairbank, poses this problem and offers
his answer in the preface of his book ade and Diplomacy on the China Coast:
[1] I think authors ought to look back and give us some record
of how their works developed. [I will try to do so in due
course]
[2] Thad two years by the tail, and both were pulling me back
to earlier times, In my observation this regression in time
is typical. . . . Asa result, would-be modernists have wound
up as medievalists. The rule seems to be, if you want to5,000 YEARS OF WORLD SYSTEM HISTORY 163
study the mid-period of a century, begin at the end of it
and let the problems lead you back. Never try to begin at
the beginning. Historical research progresses backward,
not forward.
[3] The moral seems to be: as you find your unifying concepts
[like mine above], follow them out; but don't let them carry
you off. They probably don’t belong to you in any case.
[4] [This can at best result in] a book that to its author is a
mere antechamber to a whole unwritten library, bursting
with problems awaiting exploration . . . (Fairbank, 1969:
vii-xii).
Fairbank’s second rule, to proceed backward, avoids arbitrary and
conventionally mistaken beginnings. Thus, this “backward” procedure
can help us find answers to the problem of timing the beginning and
changing perspective of our problem. However, this procedure also
poses some narrative and analytic problems of its own. Of course, Fair-
bank does not say we must also narrate backward. We can work back-
ward in history, first as far as the evidence and our analysis thereof
permits, and then narrate it forward. Fairbank does so himself.
However, this procedure may still pose some conceptual and an-
alytic problems. One is that I may be accused of stretching my present
conceptions back into history—not as far as they will go—but farther
than the evidence fits. I may be trying to squeeze it into a procrustean
historical bed. True, but not necessarily any more so than using any
other historical procedure or conceptualization, including those that
choose starting dates arbitrarily defined by their preconceptions. Is
Ranke’s history “wie es eigentlich gewesen ist” ever possible?
The other analytic problem is twofold. First, the real historical proc-
ess did “eigentlich” move forward and not backward. So there are good
grounds not only to tell, but also to analyze, history with the stream
rather than against it—once we have pursued it far enough back to
find its origin!
However, if we want to analyze the process of historical cumula-
tion, or even evolution or development, in a single system, then we
must also follow and analyze this system-wide process progressively
and cumulatively from its inception. Fairbank’s first, third, and fourth164 Andre Gunder Frank
rules are reproduced here for what they are worth. I believe they are
useful as guides and cautions in doing and communicating this his-
torical analysis.
4. Periodization
In my proposed historical review, I intend to try a compromise
resolution of these procedural problems, which is based on periodi-
zation. In embarking on this historical venture, I take it as provision-
ally established that there has been a single modern world-system since
1500, despite the critiques of the work by Immanuel Wallerstein (1974,
1980, 1989), and the lack of them in my own above-quoted World Ac-
cumulation 1492-1789. Therefore, the differences between these books
and the present theoretical introduction and proposed subsequent his-
torical review are the following: I now also stress and examine “sys-
temic connections in a single historical process” extending back much
earlier than 1500. I now examine these systemic connections in a sin-
gle historical process over a much wider geographical and social range,
including at least the entire Afro-Eurasian ecumene, of which Europe
and its world is only a part. Thus, the historical and socio-geographical
scope of this process is no longer seen as beginning and centered in
Europe, which, on the contrary, joined it rather late. I will also ques-
tion the supposed historical uniqueness and perhaps the social-
theoretical relevance of the modern capitalist mode of production.
Therefore, I now also propose to work as far afield geographically and
as far backward in time, and to consider as diverse modes of produc-
tion, as the evidence, its systemic interpretation, and my ability per-
mit. Of course, I still regard it as essential “to analyze the processual
or systemic connection between [historical events] in different parts
of the world at the same time” (Frank, 1978b: 22).
To do so, I may, however, wish to introduce some periodization
or “stages” into my reading and writing of history. In my subsequent
“Historical Review? 1 intend to (re)read the history of major periods
before ap 1500. For the reasons explained above, I will start in ap 1500
and work backward from period to period. For convenience, I will di-
vide history before ap 1500 into periods of 1,000 years each, and further
subdivide some into periods of 500 years. I will examine each of these
periods in reverse chronology, starting from ap 1500 and by implica-
tion or reference from today. However, within each of these periods,5,000 YEARS OF WORLD SYSTEM HISTORY 165
I will also read and write historical development forward. These per-
iods are: the modern period since a 1500; the medieval period from
AD 1500 to ap 500; the classical period from Ap 500 to 500 Bo; the
period of late antiquity or ancient history from 500 ac to 1500 nc;
and the periods of earlier antiquity before 1500 nc. Of course, my in-
tent is not to create arbitrarily convenient divisions of historical time.
Instead, I will argue that these convenient divisions roughly corres-
pond to historically significant major turning points, which also merit
special analysis in themselves. Where and when these dates do not co-
incide, I will make the necessary amends, which correspond to the his-
torical evidence.
5. Classification
In the present “Theoretical Introduction? I will (re)examine other
approaches and (re)read others’ renditions of history. On the concep-
tual level, I will examine world-historical, macrosociological, and
political-economic approaches, beginning with those most distant and
moving successively closest to my own. On the temporal level, how-
ever, this theoretical discussion will examine approaches, beginning
mostly with those that deal with historical periods closer to our time
and proceeding more or less successively to the times most distant from
our own. The purpose of this introductory exercise is not simply to
review some of the state of the art. The purpose is more to review
omissions of, or objections to, world system history by others, and to
argue why these omissions and objections are not warranted. In other
words, I review the work of others to build a case for my own. By ex-
amining, and where warranted, rejecting the “anti-world-system-before-
1500” arguments of others, I intend to build up a prima facie case for
my thesis that our world system began thousands of years earlier.
World historians in the United States and elsewhere have sought
alternatives to Eurocentric history. However, most of their proposed
alternatives have centered largely on the additional or comparative study
of other world regions and earlier civilizations, particularly in Asia.
Only a few historians, notably William McNeill and L. S. Stavrianos,
have written and pleaded for a more unified world history. Even they,
however, have laid less stress than I would like on analyzing interde-
pendence at the same time in a single world system.
I have examined numerous world-historical works with suggestive166 Andre Gunder Frank
titles like A World History (McNeill, 1967), New Perspectives in World History
(Engle, 1964), Perspectives on the European Past: Conversations with Historians
(Cantor, 1971), History for a United World (Dance, 1971), The World to 1500:
A Global History (Stavrianos, 1970), and Cross-Cultural Trade in World His-
tory (Curtin, 1984). I have also examined works on historiography and
historical method and the Encyclopedia of World History (Langer, 1972),
the Penguin Alas of World History (1974), and the Penguin Atlas of Ancient
History (McEvedy, 1967). Additionally, we may draw on many works
of historical and comparative macrosociology like Big Structures, Large
Processes, Huge Comparisons (Tilly, 1984), The Sources of Social Power: A His-
tory of Power from the Beginning to 4.D. 1760 (Mann, 1986), The Political
Systems of Empires (Eisenstadt, 1963), and Comparative History of Civiliza-
tions in Asia (Farmer et al., 1977).
6. Trinities of History
Almost none of these approach “world history” as recommended
here by me, or in the classical opening epigraphs I cited in my sup-
port. Curiously, several of the surveys on how history is, or can be,
written, review a trinity of “three” ways to do it. More curiously, their
trinities are very different from each other. For instance, Canter (1971)
distinguishes “humanist historiography” (including narrative and bi-
ographical history), “romantic historiography” of national histories, and
“comparative history? which is again subdivided into three kinds, im-
pressionistic, positivist, and model-building. Edith West recognizes in
Engle “three main organizational schemes for organizing history con-
tent; history can be studied by geographical regions, chronologically,
or topically (1964: 595), Most curiously, none of these sets of three
possible and recommended ways to do history (or not to do history)
includes one world history, and least of all one like the one I recom-
mend above and try to practice below, Very few historians or macro-
sociologists even consider the possibility, let alone the desirability, of
world systemic history, even though some of them also pursue non-
or anti-Eurocentric ends.
Therefore, I hope I may be forgiven if I also invoke my own trin-
ities of historians and historical macrosociologists with regard to how
they do, and would have others, write world systemic history.
1. The largest majority swear by the historical gods of the above-
named or other trinities, and forsake world history altogether for re-5,000 YEARS OF WORLD SYSTEM HISTORY 167
gional, sectoral, personal, or whatever other “history?
2. Many historians and macrosociologists claim to write the his-
tory of the world, but without ever attempting to do world history.
‘They distinguish various civilizations or other systems, and sometimes
study one problem or another, like ideology, power, economy, tech-
nology, etc., through many if not all of them. Toynbee (1946), Quigley
(1961), and more recently Mann (1986) are among them.
3. Some students of world history or historical macrosociology do
consider something like my kind of world history and for some of the
same reasons. Following precedent, I shall subdivide this third cate-
gory into three subcategories as well.
3.1. Some consider world system history, but reject it outright. Two
examples are Anderson (1974), who was quoted above, and Tilly (1984),
to be examined below.
3.2. Others indicate some real or potential agreement in principle,
and sometimes in part of their analytical practice, with world system
history. Then however, they devote most of their own historical and
macrosociological practice to comparative, model-building, or region-
al history instead. Examples are Farmer et al. (1977), Chase-Dunn
(1986), and Ekholm and Friedman (1982).
Some of the authors in this second category, and also others who
might fit there, also help to raise and partly to answer two important
theoretical/conceptual and empirical/historical questions about world
system history: (a) what do we mean by world system and how do we
define, recognize, and bound it? and (b) does this system have a rec-
ognizable, albeit perhaps changing, structure, in particular a hege-
monical (?) core-periphery structure? Among those who think so, a
minority prays at the altar of the third divine form in my trinity.
3.3. Thus, a small minority finally does some—and in a few cases
very much—world system history. Some, including Marxists, Waller-
stein (1974), and formerly myself (Frank, 1978b), limit their history
and their conceptual definition of the world system to the period after
1500. This date already represents an advance, chronologically 300
years backward, and conceptually wider. Most Marxists, and lately
also Eric Wolf (1982), focus on industrial capital(ism) and therefore
define and date the capitalist world system only from 1800. Others
extend the world system back further, to the thirteenth century, like
Abu-Lughod (1989), and to the Islamic, Indian, and Chinese world168 Andre Gunder Frank
and system of Hodgson (1974), Lombard (1975), Chaudhuri (1985),
and Gernet (1985). Finally, the smallest minority, especially McNeill
(1964), Stavrianos (1970), and to some extent Curtin (1984) and Wil-
kinson (1987), extend their systemic purview still farther backward in
history and outward in the world. However, even they shy away from
pursuing it to its “logical” conclusion.
7, Historical (Preview and Central Asian (W)hole
Below, I will mine and refine some of these above-mentioned and
other approaches and studies as inputs into my own. I will also draw
on students of interregional trade, migration, conquest, and techno-
logical and cultural or religious diffusion. All can help us extend these
systemic connections farther back in time and space. I will reexamine
Wallerstein’s transition to capitalism around 1500. I will examine Abu-
Lughod’s Thirteenth-Century System (1989), and will look at the me-
dieval Islamic, Indian, and Chinese world through the perspectives
of Hodgson (1974), Lombard (1975), Chaudhuri (1985), and Genet
(1985). Scholars like Hirth (1885), Hudson (1931), Needham (1954-84),
and Yu (1967), document the relations between imperial Rome and
Han China via the Parthian empire and India, and to their outlyers
in Africa, southeast Asia, and Japan. We should ask how systematic
and systemic these relations were. Similarly, we should inquire into
Hellenic-Indian and still earlier religious and cultural diffusion and
other connections. Most of all, I will rely on McNeill (1964), Stavri-
anos (1970), and Wilkinson (1987), as my most systematic and systemic
predecessors and guides in the journey back through world-historical
time.
Even these world system historians, however, give central Asia less
emphasis than I would like. I argue that we must devote much more
attention and systematic analysis of systemic connections in and through
central Asia. This area, its peoples and their often decisive importance
for world history, is an enormous hole in most historical work, espe-
cially that on ‘civilization(s)? and also on our historical world system.
(1 am tempted to call this omission a central Asian black hole, except
that it seems to repulse and expel rather than attract historians and
others.) During several millennia, repeated waves of migrating, con-
quering, but also newly socializing, “hordes” of “barbarians” emerged
in all directions out of central Asia. They fundamentally changed the5,000 YEARS OF WORLD SYSTEM HISTORY 169
lives, civilization, and relations, of the peoples of cast, central, and
west Asia. They also affected the history of Europe, Africa, and the
Americas. Yet these peoples and their movements are largely ignored;
considered uncivilized. At most, some consideration is given to their
impact elsewhere, but without considering the dynamic and cause of
this “Pulse of Asia” (Huntington) itself. Least of all has this central
Asian dimension been incorporated into a reading of the history of
the world system itself. Lattimore (1962), Krader (1963), Stein (1974),
Rossabi (1982), Beckwith (1987), Barhold (1962), and other Russian
and Soviet, as well as, Chinese authors can help us.
Let a hundred flowers bloom for each approach of the trinity. I
certainly welcome and draw on all kinds of historical, sociological, an-
thropological, ideational, religious, psychological, political-economic,
technological, biographical, national, sectoral, comparative, and other
studies. I also welcome contemporary worldwide, and not so wide-
ranging, histories of cycles of civilizations, war, hegemony, colonialism,
economic organization, ideologies, social movements, etc. All of these
can advance our production and consumption of knowledge directly,
and also indirectly, as raw material inputs for other flowery approaches.
This study welcomes the other studies and draws on them and their
various historical realities to connect them into a single historical whole
as far as possible.
However, I also request similar acceptance and place, if not pride
of place, for the systematic study of world system history which I am
trying to advance here. Therefore, I ask any possible true believers
in their own methodological gods or trinities, at least to accept the
equal right and merit of the one proposed here to exist as well.
Ill. TRINITY OF WORLD HISTORICAL APPROACHES:
ALTERNATIVE AND COMPLEMENTARY
1. Specialized Histories
I will draw on non-world-historical and social-scientific studies when
and where they are useful as raw material inputs in this study, but
without further discussing them per se.170 Andre Gunder Frank
2. Histories of the World, but not World Histories
It may be useful to cite and comment on some selected histories
of the world, which are not world histories, in order to compare and
contrast them with my own. In so doing, indeed, I also wish to argue
that these authors have few or no good reasons to reject my road to
our often common goal. We may start with the best known of these
historians, Arnold Toynbee.
Arnold. Toynbee (1946: 34-40) finds (nineteen) or twenty-one civili-
zations, five still living and sixteen dead (though “most of them
[were/are] related as parent or offspring to one or more of the others”
(1946: 34). He rejects the following alternative “misconceptions” or “il-
lusions,’ which he seems to think exhaust all alternatives, which they
do not: “The egocentric illusion [of] the misconception of the unity
of history—involving the assumption that there is only one river of
civilization, our own” (Toynbee, 1946: 37). We should indeed reject
this Euro/Western egocentric illusion, but it is Toynbee’s misconcep-
tion to assume that there cannot have been, or be, a single unifying
river unless it was or is that of “our” (Western, or for that matter, any
other single civilizational) river. I suggest there was much of a com-
mon river and unity of history.
‘Toynbee also rightly rejects “the illusion of ‘the unchanging East?”
“The East” has no historical existence. Indeed, it was a Euro/Western-
centric invention. Moreover, of course, the many peoples and regions
of “the East” have been very different and everchanging. This fact, and
reading of history, need not and should not, however, exclude these
peoples and regions from participation in a common stream of history
or historical systemic unity.
Thirdly, Toynbee rightly rejects “the illusion of progress as some-
thing which proceeds in a straight line” (1946: 38). Leaving the criteri-
on of progress or not aside for the moment, it can certainly have had
its historical ups and downs in any and all parts of the system and
maybe in the whole system itself, while still maintaining and perhaps
(as in natural selection) even furthering systemic unity and continuity.
Finally, Toynbee rejects the “very different concept of the unity of
history” (1946: 37), which regards it as the diffusion of Egyptaic civ-
ilization over thousands of years. The rejection of this fourth concept
of unity, like that of the Eurocentric first one, also does not require5,000 YEARS OF WORLD SYSTEM HISTORY 171
the rejection of the unity of history or of a single historical world sys-
tem itself.
Therefore, Toynbee’s rejection of both historical unity and histor-
ical continuity is unnecessary on all five counts, his own four rejections
and his own alternative reading.
Carroll Quigley (1961) devotes more attention than Toynbee to the
interrelations and mutual influences among civilizations and the
rise and declines of each through their seven stages of mixture,
gestation, expansion, conflict, universal empire, decay, and invasion.
Nonetheless, he still recognizes sixteen separate civilizations, which
resemble, but also differ from, those of Toynbee. Thus, Quigley
also writes a history of the world without attempting to write world
history.
Michael Mann (1986: 1-2) sums up his approach in two statements.
Both could offer justification and basis for my world system historical
approach. However, in Mann's hands they do the opposite:
Societies are not unitary. They are not social systems (closed
or open); they are not totalities. We can never find a single
bounded society in geographical or social space. Because there
is no system, no totality, there cannot be “sub-systems,” “dimen-
sions” or “levels” of such a totality. Because there is no whole,
social relations cannot be reduced “ultimately” “in the last in-
stance” to some systemic property of it—like the “mode of ma-
terial production? or the “cultural” or “normative system? or
the “form of military organization.” Because there is no bounded
totality, it is not helpful to divide social change or conflict into
“endogenous” and “exogenous” varieties. Because there is no so-
cial system, there is no “evolutionary” process within it... . There
is no one master concept or basic unit of “society”. . . . would
abolish the concept of “society” altogether.
The second statement flows from the first. Conceiving of so-
cieties as multiple overlapping and intersecting power networks
gives us the best available entry into the issue of what is ulti-
mately “primary” or “determining” in societies. . . . [There are]
four sources of social power: ideological, economic, military,
and political (IEMP) relationships (1986: 1-2).172 Andre Gunder Frank
I can only agree with Mann’s proposal to abolish the concept of
society and with his rejection of the search for a single, ultimately de-
terminant, property thereof. For most of Mann's rejection of the prem-
ises of orthodox history and social science, right and left, also eliminates
many underbrush obstacles on the way to the world system history
I propose. However, I have some reservations about his prima facie
rejection of all totality and systemic property, as well as to his preoc-
cupation with power alone. In particular, I cannot be satisfied with
his inquiry only into “the sources of social power” at different times
and places, without a systematic attempt to investigate possible con-
nections between here and there, and to trace possible continuities be-
tween then and now. Moreover, as per Section 3.2.2. below as well
as Gills and Frank (1989), I suggest that Mann’s focus on power itself
devotes insufficient attention to the use, if not the motive, of power
for ulterior economic ends.
3. World System History Considered and
(@) Rejected, (2) Neglected and Conceptualized, (3) Pursued
3.1. Wort System History ReecTe>
An example of the consideration and outright rejection of world
system history is Perry Anderson (1974), as cited above.
Charles Tilly (1984: 19, 62, and passim) inveighs against some of the
same premises as Mann, which Tilly calls the “pernicious postulates
of twentieth-century social thought” The first four, and most impor-
tant, of these are (1) that “society” is a thing apart and that the world
is divided into “societies”; (2) that social behavior results from indi-
vidual mental events; (3) that social change is a coherent general phe-
nomenon, explicable en bloc; and (4) that social change is a succession
of standard stages, each more advanced than the one before. I, and
I presume Mann, agree. Elimination of these postulates can also pave
the way toward a world system history.
Moreover, Tilly asks “how can we improve our understanding of
the large-scale structures and processes that were transforming the world
. 2” Tilly answers by arguing that “the most pressing theoretical prob-
lems are to connect local events to international structures of power
and to improve existing models of these international structures” He
considers doing so at the world-historical, the world-systemic, the macro-5,000 YEARS OF WORLD SYSTEM HISTORY 173,
historical, and the microhistorical levels. “If the world forms but a sin-
gle coherent network, the first two levels collapse into one. . How
many levels exist and what units define them are partly empirical ques-
tions” But “if any connection counts, we will most likely discover that
with trivial exceptions the world has always formed a single system?
Tilly rightly rejects counting any connection; but he jumps to the un-
founded conclusion that therefore “only in the last few hundred years,
by the criterion of rapid, visible, and significant influences, could some-
one plausibly argue for all the world as a single system. . . . [This] im-
plies that human history has seen many world systems, often
simultaneously dominating different parts of the globe” Therefore, Tilly
argues that we must study many “big structures, large processes, huge
comparisons” (1984: 19, 62, and passim). Yet Tilly’s own objectives and
alternative criteria to the pernicious postulates also permit alternative
plausible arguments, which have been made by some others and that
are closer to my own below. To begin with, there could have been a
multi-centered and yet single system. (Tilly does not accept this ar-
gument. On the contrary, in private correspondence of July 30, 1989,
he suggests that we would have to adopt precise numerical criteria of
degrees of influence to measure significance, which in turn I reject
as deleterious.)
3.2.1. Connections Recognized but Neglected
Edward Parmer (1985) recognizes world-historical connections and
interactions, which may be systemic; but he rejects their study in favor
of historical comparisons instead. Farmer represents the longstanding
teaching program on world history at the University of Minnesota.
In an article in The History Teacher, entitled “Civilization as a Unit of
World History: Eurasia and Europe's Place in It? he writes:
My starting point is the proposition that European history needs
to be placed in some kind of global context if we are to be able
to think clearly about the history of the non-Western or non-
European world . . . [and meet] the need to counter Europo-
centrism. . . . The way to put European history in its place is
to consider it from the perspective of Eurasian history as a
whole. .... [For] a balanced view of the whole human com-
munity, the answer I propose, at least for the period before 1500,174
Andre Gunder Frank
is to employ a comparative perspective. . . . World history [before
1500], according to this definition, includes the history of the
Eurasian land mass and [comparing] its four major civilizations
developing in parallel. . . . In addition to comparing civilizations
we can also talk about their inéeraction . . . [through] religions
. .. trade and warfare (345-54; emphases in original).
Elsewhere, Farmer et al. (1977) recognize interaction and reject
it even more explicitly.
Early modern empires were founded in all cases by military
conquest, . . . South Asia, the Indian subcontinent, has through-
out its history interacted with West Asia and Central Asia along
overland communication routes. India has also played a role
as an intermediary in maritime trade between Southeast Asia
and areas to the west, and South Asian values have made a
profound impact [elsewhere].
[Nonetheless] first of all, we are not primarily concerned with
seeking the origins of civilized society. . . . Second, interaction
among civilizations is not the focus. Contact between one part
of Eurasia and another, the influence of one culture on another,
is a constant and substantial portion of our picture, but it is
not the interaction itself that is of primary interest, but the effects
of interaction on the civilizations involved. .. .
[Instead, ] the comparative perspective is useful in studying Asian
civilizations for two principal reasons. First of all, by treating
the history of civilizations comparatively, one underscores the
differences among them and hence their separateness in a way
that dispels the notion of Asian unity. Second, comparison stim-
ulates awareness of contrasts.
Because there is no single Asian civilization, it follows there
can be no uniform history of Asia . . . [but] five culture zones,
three of which [West Asia, South Asia, and East Asia] are clearly
identified with civilizations . . . which historically interacted
. . . with Southeast Asia and Central Asia. . . . [The latter mer-
its less study because it] was generally inhospitable to the de-
velopment of civilization, although in some parts of this region5,000 YEARS OF WORLD SYSTEM HISTORY 175
oasis cities flourished as local variants or blends of major civ-
ilizations (1977: Preface and Introduction, passim).
Ofcourse, I agree about countering Eurocentrism and putting Eu-
ropean history in its proper place. But why do so only through making
comparisons, rather than also by studying interactions (which if they
are systematic enough may be systemic)? For a balanced view of the
whole human community, and of Europe within it, we also need a
history of the whole historical system and the interactions and com-
parisons within it. Moreover, the comparison of contrasts within a sys-
tem of interaction is more revealing than studying differences without
connections, let alone unity. Yet, although Farmer notes some inter-
actions, he insists on teaching only comparisons. For instance, he writes:
‘The early modern empires appeared in the wake of the Mongol
world empire. In some respects the early modern empires had
the character of rallies by civilizations seeking reintegration af-
ter the Central Asian intrusion. In some cases they were con-
quest empires built by Central Asians following the Mongol
model. In all cases the influence of the Mongol experience can
be clearly discerned. Among the features of these regimes are
the following elements: conquest of the contiguous land area;
a strong ruler with highly centralized authority; a tax system
designed to exploit an agricultural resource base; reliance by
central authority on a military machine built around horse-
mounted cavalry; ideological ties to the indigenous religious
system; and efforts to stabilize the social order by legislating
class statuses.
Examples of early modern empires are: in West Asia the Ot-
tomans (1453-1918) based in Anatolia, and the Safavids
(1501-1722) in Iran; in South Asia the Moghuls (1526-1739);
in East Asia the Ming (1369-1644) and Qing (1644-1911) Dy-
nasties in China, and the Tokugawa state (1600-1868) in Japan;
and in Eastern Europe and Central Asia the Russian Empire
(1480-1917) (Farmer, 1985: 357-58).
Of course, the comparative method is important for displaying com-
mon features, such as those Farmer discerns above, However, to under-
stand why these common features appear at particular times, we need176 Andre Gunder Frank
also to study the common problems, experience, interactions, and back-
ground. In Asian history in particular, moreover, we need to study
the possible centrality of central Asia, which much comparative study
lets fall into an empty hole between the several “civilizational” stools.
The comparison between the above-mentioned Safavid empire in
Iran and the Ming empire in China was primary in a Farmer seminar
I attended at the University of Minnesota (when I was teaching world
history there last spring). All my attempts to make connections, as
well as comparisons, between the analogous reactions in different parts
of the erstwhile Mongol empire of Genghis Khan to their immediately
prior common experience were rejected by Farmer, his colleagues, and
their students. They all agreed that mine was not the right way to study
world history. It is hard for me to understand the resistance to my mod-
est proposal of arguing for connections in addition to, but not instead
of, their comparisons. My amazement has only increased since finding
that Polybius had insisted on the same thing a bit earlier, as cited in
my opening epigraph: study “the threads that connect each event with
the whole complex, and also by pointing out the resemblances and
differences”!
Christopher Chase-Dunn (1986, 1989) devotes considerable explicit at-
tention to extending world system theory historically backward before
1500 and also to prehistory. I shall return below to his theoretical prop-
osition of criteria for the definition and bounding of a world system.
Chase-Dunn also devotes much analytical and empirical attention to
interconnections, especially commercial ones. On many occasions, he
points to wide-ranging interregional interconnections in his work on
world systems, core-periphery hierarchies, semiperipheries, and modes
of production. Yet, rather than systematically pursuing such systemic
interconnections in a single evolving world system, Chase-Dunn still
prefers to emphasize comparisons among different world systems, hier-
archical organizations, and modes of production.
Chase-Dunn, however, advances my cause not only by his admis-
sion of pre-1500 and prehistorical interconnections and world system/s,
but also by explicitly rejecting modes of production as the primary,
predominant, and determinant standard or criterion of system defini-
tion and bounding. I welcome this challenge to the mode of produc-
tion, which is dominant de jure in Marxist and de facto in many
non-Marxist studies. I have long claimed that this preoccupation with5,000 YEARS OF WORLD SYSTEM HISTORY 177
modes of production in general, and in particular with the capitalist
(plus historically the feudal and ideologically the socialist!) modes of
production, is an obstacle to systematic thinking about systemic inter-
connections. Exclusive or predominant attention to the mode of pro-
duction necessarily blinds us to the wider systemic connections and
historical antecedents and development of the modern world system.
‘The same preoccupation with modes of production probably also gen-
erates unrealistic hopes about the socialist future of our world system.
Indeed, this may be the functional, if not causative, political ideolog-
ical raison detre of insisting on mode of production as the key (to so-
cialism?).
I also vote for abandoning primary reliance on the mode of pro-
duction as the determinant criterion of historical development. I vote
for replacing the focus on mode of production with a focus on the modes
of accumulation in the world system. Of course, this also means that
we must abandon the definition of one in terms of the other. On the
other hand, my wife Marta Fuentes argues that we should not entirely
abandon the study of modes of production, insofar as they may help
shape the gender structure of society. This proposition sounds intui-
tively persuasive and merits further study. However, this means using
the idea of modes of production for something very different, and prob-
ably more useful, from the debate between Marxist collectivists and
bourgeois free marketers. Moreover, perhaps feminists can consider
“mode of accumulation” to include patriarchal reproduction.
Chase-Dunn shares my reservations about the thesis that there
already is a socialist mode of production in a separate world system.
We both see the existing socialist countries as part and parcel of the
same world capitalist system (Chase-Dunn, 1982; Frank, 1980). He
also shares some of my skepticism about the future, although he looks
forward to a single world state as an arbiter to avoid future world wars.
Despite Chase-Dunn's help, I nonetheless find him to be still excessively
reticent to pursue the analysis of the single world system backward
through history as far as it will take us.
Exkholm and Friedman (1982) do take our world system farther afield
and farther back in history. Moreover, they also reject the blinders that
the production restriction imposes on seeing, reading, and writing about
the systemic continuity of the world system. Yet their long historical
and world systemic perspectives in theory are contradicted by their178 Andre Gunder Frank
continued insistence on referring to world systems, which they com-
pare over time and space between the here and now and the then and
there of Rome, China, Egypt, Sumer, etc. Their continued insistence
on only comparative analyses of, for instance, the role of the semi-
periphery in the system, and their virtual exclusion of tracing the con-
tinuity of the whole system is all the more strange, since Kajsa Ekholm
has herself written that
an international production system with all its concomitant di-
vision of labor has been a constant in the world economy, more
general than modern capitalism. . . . The decisive importance
of trade in economic growth is related to the fact that expan-
sion of production has always been primarily based on export
production, that growing accumulation has depended on increas-
ing export, with or without local production as its base. . . .
The importance of external trade and markets for economic
development is easy to document by simply enumerating the
great number of wars that have been fought in order to main-
tain and expand control over such “resources.” In fact, loss of
such control has invariably resulted in economic regression.
Europe's expansion around Africa into the Indian Ocean led
toa clear decline in the Middle East which had previously been
strategically located in terms of world markets. . . .
It is this reductionist error [of Marxist insistence on local “inter-
nal” production and class struggle] that is the basis for the mis-
conception that capitalism developed in Europe via an internal
process. . . . Thus the development of capitalism in Europe is
not the result of an evolution from feudalism as a system, but
the result of a shift in a single system within which the shift
is itself an aspect of the local transformation in Europe. In effect,
the capitalist evolution of Western Europe represents a contin-
uation and further development of processes already in existence
in the Middle East. At the same time, the Middle East expe-
riences economic decline, political fragmentation and a tendency
to a kind of feudalism reminiscent of Europe in the earlier pe-
riod. . . . In fact, every peripheral sector is a potential center in the next
period, and centers are usually transformed peripheries from a previous pe-
riod (1980: 157-61; emphases in the original).5,000 YEARS OF WORLD SYSTEM HISTORY 179
This is the direction in which I propose to go, following not only
Ekholm but also Polybius before her! However, if we begin our his-
torical journey backward through time from 1500, we will soon find
more than the Mediterranean and Middle East (which is a Eurocen-
tric term for west Asia), In the immediately preceding centuries, Gen-
ghis Kahn and his successors up to Tamerlane were active in the whole
Asian expanse from China to Europe. Moreover, in the preceding cen-
turies also, China flourished more than any other part of the world
and was its center, if any, according to McNeill (1983). I shall return
to this subject in the historical review below. Here already however,
it raises questions about the geographical and social extent of the world
system, about its hegemonic centers if any, about their rise and de-
cline, and about possible shifts of hegemonic centers from one place
to another.
3.2.2. World System Connections Conceptualized
The didactic advantages of retaining some of the development (pro-
gression?) of my conceptualization of world system history through
three drafts of the present paper I hope will outweigh the stylistic costs
of taking the reader through some of this conceptualization more than
once. Later drafts were influenced by critiques of, and conversations
about, previous drafts with people named in my acknowledgments and
particularly with Barry Gills, as represented in Frank and Gills (1989).
World System Criteria and Boundaries. Before discussing the conceptual
definitions cited below for this purpose, in my first draft I hazarded
the following preliminary proposition:
‘Tentative brief definitional criteria of what (1) system, (2) con-
tinuity, and (3) systemic continuity might be: (1) the criterion
of the range and participation in a/the world system is that it
made/makes (local) historical (including present) events differ-
ent from what they would have been in isolation from/outside
that system; looking at it the other way around, historical events
and experience could not have been the same in isolation from.
the system, or if the system had not existed; (2) the criterion
of continuity is that the past has significantly affected the histor-
ical present and future, and the historical present would have
been different if the past had been different; (3) systemic continuity
means that these criteria operate[d] (are [were] met) over the180 Andre Gunder Frank
world system as a whole, so that the (apparent) discontinuity
(birth/death life cycles) in/of a (regional or social) part does not
negate the historical continuity at the (whole) systemic level;
indeed, social transformations of the parts appear to have been
an essential constituent element of the systemic existence and
historical continuity of the whole.
I wish to state right here, however, that I am conscious that
this “whole” is also relative. In particular, even the pieces of
its history have been written almost exclusively by men about
men, who constitute(d) at most half of the whole (population).
‘The “hidden history” of women’s participation —and therefore
of the entire relations between men and women—in this whole
history (and not just over 300 years as in Sheila Rowbotham’s
book of that title) largely remains to be written. And I have
no hope of even beginning to work on shis task here, other than
to express my awareness that it is at least half of the task.
Few women, let alone men, have yet devoted themselves to this
half of the task; and those who have, like Gimbutas (1980, 1982) and
Eisler (1987), only pursue part of the history of the world system. So
let us return to the other half. (Two readers—both women—have ob-
jected to this paragraph. One said it is a selfserving attempt to absolve
myself from responsibility for undertaking—do it, or shut up about—
this whole task. The other, my wife Marta Fuentes, said that women
probably really did not have much role in “my” kind of history. I ap-
preciate and respect both comments, but I also disagree with both.
Asking the right question may not be as much as half the right answer,
but it can Aelp—even me—find it; and I do think that both women per
se and the patriarchal structure of society, and process of reproduc-
tion, did play a major role in all of our history as well as in my view of it.)
Tilly (1984) poses the question of how to define and bound the sys-
tem after rightly rejecting “society” as an answer. On the one hand,
he says if any connection counts, the answer and the “system” is trivial.
On the other hand, if we use criteria taken from nineteenth-century,
let alone contemporary, reality, we could find no system before that.
Somewhere between those extremes lie all useful accounts of
human connectedness. A sensible rule of thumb for connected-
ness might be that the actions of powerholders in one region5,000 YEARS OF WORLD SYSTEM HISTORY 181
of a network rapidly (say within a year) and visibly (say in
changes actually reported by nearby observers) affect the wel-
fare of at least a significant minority (say a tenth) of the pop-
ulation in another region of the network. Such a criterion
indubitably makes our own world a single system (Tilly, 1984:
62).
But Tilly begs several questions with his “rule of thumb? and maybe
we need more dexterity. First of all, he transfers part of the problem
to defining and bounding his different regions. Why then should sig-
nificant effects between here and there be limited to one year? How
can we demand local observer reports? Many people, even now, do
not know what hit them here from there. Even if past people did re-
port what hit them here, their reports may long since have been de-
stroyed or lost. Yet indirect and inferential evidence may still be available.
Why should we discount that? The effects themselves may also be in-
direct. For instance, direct effects may affect very few, but these in turn
can affect the welfare of many. Also, why include only the actions of
powerholders and not of others in the system? In short, we need more—
or is it less?—than Tilly’s “rule of thumb” to seek and find the world
system.
Chase-Dunn also tries his hand:
I shall contend that any regularized and repetitive exchange
of material goods which substantially affects the welfare of a
population or the maintenance of a power structure constitutes
part of a single world-system network (1986: 102).
This sounds better to me, because it is more flexible. However,
it still begs questions of what substantial, welfare, power structure, and
population are. Moreover, what is the time-frame of regularized and
repetitive? Also, why only exchange of material goods? Furthermore,
Chase-Dunn continually refers to “societies” as though he and we knew
or could know what they are.
Jane Schneider (1977) addresses another problem of bounding under
the title “Was There a Pre-Capitalist World System?” One reason she
answers affirmatively is that she challenges and rejects the bounding
criterion of trade in essentials and luxuries. Wallerstein and others ex-
clude trade in luxuries, and therefore the regions from which luxuries182 Andre Guader Frank
are imported, from the division of labor and the world system. Source
regions of luxury trade are thereby relegated neither to the core nor
to the periphery of the system, but to the outside of the system.
Schneider “suggest[s] that this dichotomy is a false one which obscures
the systemic properties of luxury trade” (1977: 21) and the systemic
connection of regions, peoples, and their producers who are falsely
excluded from the world system. Moreover, Schneider also relates trade
in luxuries to trade in precious metals and both to the predatory use
of force.
Since antiquity, expanding empires have acquired precious met-
als and other highly valued luxuries, not only by plunder and
piracy, but by exporting finished goods, above all cloth. . . . It
is possible to hypothesize a pre-capitalist world system, in which
core-areas accumulated precious metals while exporting man-
ufactures, whereas peripheral areas gave up these metal (and
often slaves) against an inflow of finished goods. .. . Because
the tribute system was maintained through the distribution of
luxuries, these goods were also “essential”. . . . It is necessary
to view luxury trade not merely as a stimulus to production,
or an adjunct to stratification, but also as a series of long dis-
tance exchanges of relevance to the capture of energy (1977:
24-25).
Chase-Dunn (1986) agrees in principle, as do I, but I would go
even further. The “highly valued luxuries” were part and parcel of the
exercise of power and the store of value. The “adjunct to stratification”
also was an “adjunct” to exploitation, and the “capture of energy” in-
cluded economic surplus. Empirical confirmation is to be found, for
instance, in Lombard’s analysis of medieval Afro-Eurasian-wide trade
in precious metals and luxuries in exchange and division of labor (or
production) for (other) “necessities” However, it may be argued that
luxury trade was also vital in classical and ancient times. The forth-
coming historical review will analyze the important role of the silk (and
other luxury) trade between Ming China and Imperial Rome via inter-
mediaries in Parthian Persia and India. Moreover, this trade may have
contributed to the decline of the former and the development of the
latter. Luxury trade was also important for the entire Fertile Crescent
region and its outlyers in the earlier Axial Age.5,000 YEARS OF WORLD SYSTEM HISTORY 183
Indeed, it may be argued that for this earlier history, even Schneider
does not value luxury trade enough! Luxury (not to mention precious
metals) trade may have been even more significant than she suggests
(and more important than staples trade?) for the following reasons:
Not only was luxury consumption part of the ideological and social
cement holding up the hierarchical structure of society and its pro-
duction of social surplus, but long-distance luxury trade also tied the
luxury consuming elites to each other—and therefore to the hierarch-
ical structure of surplus production and of exploitation here and there.
Thus, luxury trade may be said to have been an integral, instead of
only a marginal or inessential, part of economic, social, political, and
ideological organization everywhere. Therefore, “luxury” trade is also
a very significant element of the world system as a whole.
Another approach to defining, if not bounding, the world system
is to appeal to general systems theory in general, and/or to analyses
by Prigogine and others (Prigogine & Sanglien, 1988), in particular
of the emergence of order out of chaos.
Wilkinson (1987: 33), to whom I shall return below, is also intent
on defining connections and unity, albeit of what he calls “Central Civ-
ilization” rather than of a world system. For present purposes, how-
ever, he stresses that “fighting is bonding” and “conflict, hostility, and
even warfare, when durable (habitual, protracted, or inescapable), are
forms of association” (Wilkinson, 1987: 33). I agree.
However, Wilkinson implies that not only material exchange, or
even an international (sic!) division of labor, but also various political
relations are systemic. This places Wilkinson in good company. Thirty
years ago, I wrote an essay on “Dyad, Triad, Cohesion, and Social
Science Theory: Or Toward Conflict Resolution among Theories of
Social Conflict” (Frank, 1959). I examined eight theories or approaches,
to wit: (1) Simmel’s sociology of the dyad and triad; (2) Gluckman’s
thesis about cross-linkage; (3) the balance of power approach; (4)
Galbraith’s propositions about countervailing power; (5) the theory of
games; (6) the economist’s market analysis; (7) oligopoly theory; and
(8) some aspects of Marxian theory. All of these, whether dealing with
conflict or cohesion, and no doubt others, refer to what I called “struc-
tured interaction . . . [which] is cohesive for the system which the units
constitute. It’s just that the system is transformed” by some of this inter-
action (1959: emphasis in original).tee Andre Gunder Frank
Moreover, we may recall Mann's four sources of social power—
ideological, economic, military, and political (IEMP)—all of which are
relational. But so are the repeated, if not continual, migrations, in-
vasions, conquests, and assimilations of populations around the world
over the millennia. All of these have a place in the interactions that
make up our world system, in the present and in the past. My initial
proposition about system extension and continuity still seems relevant
for my inquiry. How far afield and how far back shall I go? That is
the question.
Other than the ubiquitous but invisible society, the most popular
large social systems are and were (cultural?) civilizations, political em-
pires or other state systems, and world economies or world economic
systems. Often, the three overlap in fact or are so viewed by their stu-
dents. Before 1500, civilizations were not thought to be (or even be-
come) worldwide in scope, although the cultural influences of the major
religions spread extensively. Systemic contacts among the various em-
pires may, however, have covered much of the world, as did population
migrations and some military conquests. What about economic rela-
tions? And how interactive, regularized, repetitive, and welfare-affecting
need they have been to be termed systemic? Or, how related were they
to ideological, political, military, migratory, and other relations?
Thus on second thought, I would use the criteria reviewed above
only to extend, but not to replace, my first-draft criteria for world sys-
tem bounding. My emphasis would still be on “significant” difference
or change, where one structure and/or process there affects another
here, and some continuity between then and now. Of course, these
changes and continuity can be transmitted system-wide through all
sorts of economic, political, social, cultural, and physical or migratory
relations.
3.3. Wortp System Tuzortes anp Hisrortes
We finally arrive at several points of transition. One transition is
from the above primarily theoretical/conceptual considerations to (their
application to) the study of history before 1500. Another is the sup-
posed “transition to capitalism” around 1500. A third transition (back-
ward chronologically but I hope forward conceptually) is from a world
system centered in Europe, to a world system which is centered else-
where or not at all. The connection between the two last-named tran-5,000 YEARS OF WORLD SYSTEM HISTORY 185
sitions, of course, is that capitalism and its capitalist world system
supposedly arose through an indigenous transition from feudalism in
Europe between 1500 and 1800, according to all Weberian, Marxist,
world system, and other orthodoxies. The connection between these
two transitions is also related to the supposed primacy and uniqueness
of capitalist production, considered distinct from previous feudal or
other tributary, and future socialist, modes of production. I, of course,
now challenge the verity and utility of all these connected theses. As
to transitions, I may usefully recall that while we were waiting for,
and debating about, the transition to socialism in Chile between 1970
and 1973, it became increasingly clear that a transition is a transition
between a transition and a transition.
3.3.1. Neo-Marxists and World-Systems Schools
Marx and Marxists read history as a succession, and sometimes com-
bination, of “modes of production.” The now discredited Stalinist read-
ing of history makes it a necessary and progressive succession of modes
of production from the “Asiatic” and slave modes to feudal, capitalist,
socialist, and communist modes of production. Some(times) Marxists
read this succession into world history, some(times) into regional or
national history, but with little historical accuracy. Less dogmatic and
more catholic Marxists find varieties and combinations of these and
other modes of production and the transitions among them. However,
the central focus remains the production in world or smaller societies
and not the world systemic unity or historical continuity. On the con-
trary, this fixation with supposed modes of production prevents these
Marxists from looking for, or even seeing, systemic unity and continuity.
Thus, if capitalism is industrial, and industry was not prevalent
(which is itself a doubtful proposition) before 1800, neither capitalism
(that is the capitalist production), nor systemic continuity could have
existed before 1800. Thus, in his self-designated Marxist book, Europe
and the People without History, Exic Wolf (1982) takes a giant theoretical
step backward by dating the beginning of this world capitalist system
in 1800 for the above reason. This represents a step backward for Wolf,
because he had previously written. several important books and arti-
cles (which were very influential in my own thinking) showing that
New World peoples had been incorporated into the self-same system
since 1500, and how.186 Andre Gunder Frank
Some Marxists, like Samir Amin, have of course pushed this date
back to 1500. However, until recently, Amin has insisted that there
cannot have been any systemic continuity into times before 1500, because
only then did the production begin to change from tributary to cap-
italist. Amin’s LEurocentrisme (1988), however, represents a welcome re-
consideration of this Marxist position. The reason for my welcome
is that it removes some obstacles to looking for, and seeing, systemic
continuity before 1500, and it consciously offers alternatives to Euro-
centrist perspectives.
Samir Amin (1988) inveighs against Eurocentric Western misreading
of history and Western misappropriation of what it finds useful to sup-
port its Eurocentric ideology. Amin challenges both the Western di-
vision between East and West, and the division of history into epochs
by the latter. He recalls that Christianity, Hellenism, and Islam were
all born in what Westerners now call the East. However, the West has
(mis)appropriated the first two traditions; so much so, that in the pop-
ular image the Holy Family is blond! (Amin, 1988: 68). Amin agrees
with Martin Bernal who makes a convincing case for Black Athena: The
Afroasiatic Roots of Classical Civilization (1987). Greece followed the East
more than it led the West, notwithstanding Western claims to Athens
(and the Christian part of Jerusalem!) as the fountainheads of its own
civilization.
Similarly, Amin challenges the Eurocentric periodization of history,
which divides antiquity from the medieval age at the end of the Ro-
man empire (in the West!). However, this temporal division is not sig-
nificant for the East (the later Byzantine part of the Roman empire
continued!), Amin would rather place a dividing line at the time of
Alexander the Great, when he unified the Hellenist East (1988: 44).
Islam again unified and expanded this region and provided “the
objective basis for the progress of the productive forces and thus of
the success of the state based on the tributary production” (Amin, 1988:
40). This and other tributary societies are not immobile, and they made
remarkable progress in their development of productive forces (Amin,
1988: 108). Furthermore, Islam and other religions and peoples already
had universalist aspirations (Amin, 1988: 71). Moreover,
doubtless, embryonic forms of capitalism (private enterprise,
market exchange, free wage labor) existed in the Mediterranean5,000 YEARS OF WORLD SYSTEM HISTORY 187
region for a very long time, particularly in its Arab-Islamic and
Italian parts. In a certain way, this Mediterranean system . . .
constituted the prehistory of the system of the capitalist world-
economy (Amin, 1988: 52).
Finally, and also in support of my argument here, Amin challenges
an axiom which is almost universally accepted without question from
right to left: that factors internal, rather than also or often predom-
inantly “external” to societies, are decisive in their transformation. In
this regard, Curtin specifies even more in the opening words of his
Cross-Cultural Trade in World History:
‘Trade and exchange across cultural lines have played a crucial
role in human history, being perhaps the most important ex-
ternal stimuli to change, leaving aside the unmeasurable and
less-benign influence of military conquest. External stimulation,
in men, has heen the most important single source of change
and development in art, science, and technology. Perhaps this
goes without saying .. . (Curtin, 1984: 1).
This consideration brings us to the “school"(s) of writers who inten-
tionally and explicitly posit a large world system.
World-systems writes, like Braudel (1982-84), Wallerstein (1974; 1978;
1987), Frank (1967; 1978b), and more recently Modelski (1988), Gold-
stein (1987) and others, have written from different but related per-
spectives of the existence of a (capitalist) world-system since about 1500,
or perhaps 1450 for Wallerstein (1974), 1492 for Frank (1978b), or 1494
for Modelski (1987). For all, however, their reading has restricted the
systemic unity of world history to only about 500 years and to only
those parts of the world that have since then been incorporated in the
“Modern World-System” centered first on western Europe and then
on America. For Modelski, and more recently for Paul Kennedy in
his bestselling The Rise and Fall of the Great Powers (1987), the criterion
of systemic pertinence is the existence of and challege to hegemonic
world power, which omits most of mankind that neither has been or
sought to be hegemonic.
For Wallerstein and Frank, the criterion has been capitalist pro-
duction, pushed back to 1500 but no further. Before learning of
Fairbank’s second rule (though when I did so, I belatedly added it to188 Andre Gunder Frank
my preface), I thought I could identify and start at the “beginning”
and entitled my book on the history of the capitalist world-system, World
Accumulation 1492-1789. Wallerstein admits of other world-
systems/economies before and elsewhere, but with different modes of
production. Therefore, his reading of this “modern world-system” does
not and cannot include any regions, historical times, or world “systems?
which fall outside his, I now believe, excessively restrictive definition.
Common to all of these readings are the twin ideas that the mod-
ern world-system developed through the transition from a feudal to
the capitalist mode of production, and that this ocurred in southern
and western Europe. Anderson and Brenner, among others, deny that
this transition took place so early and so generally in Europe. There-
fore, they deny even this, from my present point of view, excessively
limited temporal and geographical scope of the world system and cap-
italist economy. In The Modern World-System, Volume III, Wallerstein
(1989: 129-38) devotes considerable analytic energy to trying to elab-
orate a criterion of inclusion and exclusion in the systemic division
of labor. He continues to insist on the distinction between “essentials”
and “luxuries” as a criterion of inclusion/exclusion. He concedes that
there is little point debating the essentiality or not of luxury imports;
if someone is willing to pay for them, who are we to say that they are
not essential? However, luxury exports, among which Wallerstein spe-
cifically includes precious metal bullion, are still a criterion of exclu-
sion for him. Therefore, and for other reasons as well, Wallerstein’s
revised criteria of a “production process and division of labor that are
integral and responsive to ever-changing market conditions of a world-
economy” still do not meet the standards of the criterion set out in
the discussion of system definition and bounding above. A number
of critics have pointed to the contradictions in Wallerstein’s position,
which obliges him to accept a division of labor and trade earlier and
elsewhere in other “world-systems,” but to exclude them from his read-
ing of the “modern world-system” The reason is that his world system
criteria are not only an extensive division of labor, but also a partic-
ular production. This is another reason to abandon the fixation with
production.
Brenner on Wallerstein. Brenner identifies the contradiction in Waller-
stein’s argument:5,000 YEARS OF WORLD SYSTEM HISTORY 189
Thus, since Wallerstein admits, even if only implicitly, that the
medieval world-system seems to fit into the model of the world-
economy, he maneuvers himself into a dilemma, which reflects
the fundamental weaknesses of his conceptual framework. On
the other hand, the medieval economy was apparently a world-
economy. However, it did not function in a capitalist way: It
was not an economic system, which systematically produced
according to the profit motive and that correspondingly accu-
mulated innovatively. That is why this system has to be clearly
distinguished from the capitalist economy, which followed in
the sixteenth and seventeenth centuries, at least in some regions.
However, since the medieval world-economy nonetheless coin-
cides in its basic features with the “modern” world-economy,
Wallerstein is obliged to explain why the medieval economic
system functioned differently from the subsequent world-
economy. Since, however, the world-system theorists have not
worked out the prerequisites of a capitalist economy, they also
cannot explain why one “world-economy” operates with capi-
talist methods, but the other one does not (Brenner, 1983: 90;
retranslated from the German).
Brenner argues that preexisting class and property relations gen-
erated the obligation to compete, the profit motive, and thereby tech-
nological innovation and capital accumulation, which are the marks
of modern capitalism. Brenner then points his finger at Wallerstein’s
inability to account for the same in his capitalist world-economy and
modern world-system.
Wallerstein could not have been clearer in saying (admitting)
that the older as well as the newer world-economy had no in-
herent power to develop their productive forces. Therefore, the
modern world-economy is only the last of a whole series of pre-
decessors, none of which ever had the ability to alter the forces
of production in any major way. [The modern world-economy]
therefore, might equally well have developed like the previous
ones; since its organization of production was not significantly
different —except for the inexplicable appearance of the “meth-
ods of capitalism and of scientific technology.” In Wallerstein’s190 Andre Gunder Prank
system, progress and innovation —so far as they appear at all—
operate as a Deus ex Machina (Brenner, 1983: 95).
So maybe, as Amin now also begins to suspect, the “modern” world-
system had a medieval precedent — or even continuity—after all! The
way out of Wallerstein’s dilemma, and the way to bring the machine
that produces the technological god down to earth, is threefold, One
is to cut (out) Brenner’s and others’ Gordian production knot. That
will also release us to tread the other ways. The second is to abandon
undue distinctions between trade in necessities and luxuries. The third
is to pursue the entire division of labor—and political and other
relations — much farther back. Then we may also find ample trade and
production for profit, its reinvestment, capital accumulation, and tech-
nological progress before 1500. Indeed, we may ask if none of these
existed before 1500, how could the modern world-system have taken
off after 1500? Moreover, we may ask, if nothing of interest had been
going on and nodhing accumulating in eastern and svuthern Europe
before 1500, why would Europeans have sought to expand there?
3.3.2. The Thirteenth-Century World System
Janet Abu-Lughod (1989) begins taking this tack. She makes a major
contribution to the writing of world history by pushing the starting
date for the world system back to 1250. In doing so, she has finally
cut the Gordian knot of the supposed break in world history in 1500,
as per Wallerstein (1974) and others. This is a major step forward in
my opinion. She denies that the present world system emerged in Eu-
rope through the transition from any previous production. She argues
instead that whatever production existed in the sixteenth century also
existed in the thirteenth century in Europe—and in the “Middle East?
India, and China.
Abu-Lughod argues that eight interlinking city-centered regions
were united in a single thirteenth-century world system and division
of labor. According to her reading, however, this world system econ-
omy experienced its apogee between 1250 and 1350 and declined to
(virtual) extinction thereafter, before being reborn in southern and
western Europe in the sixteenth century, In her words, “of crucial im-
portance is the fact that the ‘Fall of the East’ preceded the ‘Rise of the
West” (Abu-Lughod, 1989: 338).5,000 YEARS OF WORLD SYSTEM HISTORY 191
Abu-Lughod’s thirteenth-century Afro-Eurasian world system has
eight interlinked regions in three related and interlocked subsystems.
‘These were: (1) the European subsystem, with the Champaign Fairs,
industrial Flanders, and commercial Genoa and Venice regions; (2)
the mideastern Heartland and its east-west routes across Mongol Asia,
Baghdad and the Persian Gulf, and Cairo and the Red Sea; and (3)
the Indian Ocean —east Asian subsystem of India, southeast Asia, and
China. The fortunes, and mid-fourteenth-century crisis and Black Death
epidemic misfortunes, were more or less common to all of them.
In Abu-Lughod’s world system, the European regions were tied
to each other, and through the Crusades and later the Mediterranean
port cities, were tied into the Middle East. There the Venetians vied
with the Byzantines and Ottomans and allied with Egypt under Mam-
eluk rule—but not before the decline of Baghdad. This city prospered,
perhaps like no other, in the ninth-tenth centuries and benefited from
the most geographically and commercially favored route to the Orient
via the Persian Gulf. But Baghdad’s fortunes declined in the eleventh
and twelfth centuries, long before its (and Iraq’s and Persia’s) conquest
by the Mongols. Partly for that reason—and partly due to problems
with overland trade to the Orient created by the Mongols—at the end
of the thirteenth century Cairo rose to regional dominance and per-
haps became the crossroads center of the world under Mameluk rule,
which it shared with Syria. The topologically and economically less
favorable rival Red Sea route to the Orient then displaced the Persian
Gulf route and promoted Cairo’s fortunes.
All these regions suffered terribly after 1348 from the Black Death,
which had been imported from the East. Nonetheless, it may be ar-
gued that some of the ravages of the Black Death should be attributed
to the already economically, socially, and politically weakened (there-
fore unable to organize and resist the plague) Mideastern (including
Egyptian) and European societies as a result of the economic crisis,
which had started at least a half-century earlier. Abu-Lughod cites am-
ple evidence that transport and other infrastructural investment and
expansion in Venice and Genoa had declined and then halted two dec-
ades before the arrival of the plague.
The Indian coasts, and especially southeast Asia and China, had
been experiencing a centuries-long economic boom since the eleventh
century, which was manifested in fast growing intra- and interregional192 Andre Gunder Frank
trade. Arabs, Persians, and of course Indians, on the west sidc of the
Indian subcontinent, and Indians, Malays, “Indonesians, and Chinese,
on the other side, were especially active in the interregional trade. This
was the time of Chinese fleets calling on southeast Asia, India, the
Middle East, and Africa with much larger and more numerous ships
than would sail the world’s oceans for many centuries. Chinese arts
and sciences—and public works, technology, and industrial products—
were preeminent in the world. In his The Pursuit of Power, and now even
more so, McNeill (1983, 1990) insists that China was by far the most
developed and important (and therefore by Western historians most
significantly neglected) region in the world at that time.
The reasons for the sudden Chinese retreat from overseas com-
merce in the early Ming Dynasty at the beginning of the fifteenth cen-
tury has been the source of much speculation. Abu-Lughod suggests
that a (not usually considered) factor may have been the effects on
China of the fourteenth-century economic crisis in the worldwide com-
mercial system, as well as the consequences of the Mongol invasions.
Dealing with the latter was not only costly to the Chinese, it may have
swung the domestic political balance of power from supporting or tol-
erating overseas expansion to defense against overland invasions.
The Mongol conquests to the east (into China), to the south (into
the Indian subcontinent), to the southwest (of Persia and Iraq), and
to the west (through Russia into Europe) changed the face of much
of the earth. Of course, they changed the social, commercial, political,
racial, and cultural face of inner Asia and downgraded the role of Sam-
arkand and other commercial, political, and cultural centers. But the
domino-like repercussions went far beyond the reach of the Mongol
armies. At the end of this period, and in direct or indirect response
to the changes wrought by the Mongol invasions, the Ming Dynasty
rose in China, Akbar’s empire rose in India, the Safavid Empire rose
in Persia, and Europeans began a worldwide imperial and now also
trans-Atlantic venture in the West.
Was that a time of first waxing and then waning Chinese and Mon-
gol hegemony during and after the lifetime of Genghis Khan? Did the
hegemonic center shift from East to West? Posing and answering these
questions permit us to see that the post-1500 world system did not rise
in Europe out of the transition from feudalism a la Wallerstein (1974),
but that it represents a shift of the hegemonic core to Europe in the5,000 YEARS OF WORLD SYSTEM HISTORY 193
gencral direction from East to West, although we may also identify
more complicated prior West-East shifts within Asia. (This core shift
westward, of course, has also continued since then across the Atlantic
and continues still today across the Pacific. Modelski [1987] observes
this westward shift around the globe in the direction of its earlier locus
toward Japan and perhaps onward toward China.)
However, there are also problems with Abu-Lughod’s procedure.
I accept the statement in her first chapter that in “this book [she] is
less interested in identifying origins [of the system] and more [inter-
ested] in examining a crucial moment in [its] history” (Abu-Lughod,
1989: 9). Nonetheless, we cannot properly understand this, or any other
crucial moment, outside of the entire historical context which gener-
ates, shapes, and gives it meaning.
My own experience is relevant in this regard. An editor asked me
to write a comment on an early version of Abu-Lughod’s work (Frank,
1987-88). Thus, he provided me with the incentive and opportunity
to think farther back than the year 1492 at which I had begun my own
book. But once I began this route, with or without Fairbank’s rule, I
had to continue. I wrote about the period of crisis between 1300 and
1450. But if there had been a systemic crisis, there must already have
been a system, whose previous expansion during the time of the Cru-
sades led to this crisis. So I then thought about going back toward ap
1000. Only later did I attempt to follow Fairbank’s rule two farther back.
From this perspective, Abu-Lughod’s reading also suffers from
contradictions. On the one hand, she notes, correctly I believe, that
“some two thousand years ago, an earlier incipient world system ex-
isted which involved almost all the regions (except northern Europe)
participating in the thirteenth-century system” (Abu-Lughod, 1989:
343). Yet “that system also ‘failed’ after the fall of Rome and the loss
of (Chinese] Han unity, only to be restructured eventually through
the ‘rise’ of the Islamic world” (Abu-Lughod, 1989: 343). Elsewhere,
she refers in passing to the subsequent and intervening flowering of
a (part) Islamic-centered world from ap 800 to ap 1100. Yet, in her
reading, in each case the system “failed” Nonetheless, she argues that
systemic changes should rather be viewed as shifts in the direc-
tion and configuration of central trends (or vectors) . . . [and]
if we assume that restructuring, rather than substitution, is what194 Andre Gunder Frank
happens when world systems succced one another, albcit after
periods of disorganization, then failure cannot refer to the parts
themselves but only to the declining efficacy of the ways in which
they were formerly connected. In saying the thirteenth-century
world system failed, we mean that the system itself devolved. . . .
From earliest times, the geographically central “core regions”
. - . were Central Asia and the Indian Ocean, to which the Med-
iterranean was eventually appended. These cores persisted
through the classical and thirteeth-century world systems. A
decisive reorganization of this pattern did not occur until the
sixteenth century (Abu-Lughod, 1989: 343-45).
Why then does Abu-Lughod shy away from consistently extending
her own argument to reading—if not writing—a historical continuity
with a shifting core (or during periods of temporary disorganization
no single core) in a single world system over at least the two thousand
years she mentions? It seems plausible, if not obvious, to argue that
between the fourteenth-century decline of the East, and the fifteenth-
sixteenth-century rise of the West, there occurred a “declining efficacy”
and “disorganization” of “the ways in which they were formerly con-
nected” (Abu-Lughod, 1989: 343). In that case, consequently, there
would have been a shift of the center of gravity in the system from
East to West—but not a complete failure of the system as a whole.
On the contrary, this temporary disorganization, and renewed reor-
ganization, could, and I belicve should, be read as the continuation
and evolution of the system as a whole. But if we make this argument
for the thirteenth-sixteenth-century period, then why not make it for
the earlier periods, including the rise and decline of Islamic, before
that of Roman and Chinese, connections and “restructuring rather than
substitution” within a single and continuous world system which de-
volved rather than failed? I believe we should. Therefore, we should
write a world system history which goes back farther than 2,000 years
and also includes more peoples and regions of the world than those
mentioned above.
Similarly, Jacques Gernet also remarks that
what we have acquired the habit of regarding— according to
the history of the world that is in fact no more than the history
of the West—as the beginning of modern times was only the5,000 YEARS OF WORLD SYSTEM HISTORY 195
repercussion of the upsurge of the urban, mercantile civiliza-
tions whose realm extended, before the Mongol invasion, from
the Mediterranean to the Sea of China. The West gathered up
part of this legacy and received from it the leaven which was
to make possible its own development. The transmission was
favored by the Crusades of the twelfth and thirteenth centuries
and the expansion of the Mongol empire in the thirteenth and
fourteenth centuries. . . . There is nothing surprising about this
‘Western backwardness: the Italian cities . . . were at the ter-
minus of the great commercial routes of Asia. . . . The upsurge
of the West, which was only to emerge from its relative isola-
tion thanks to its maritime expansion, occurred at a time when
the two great civilizations of Asia [China and Islam] were threat-
ened (1982: 347-48).
3.3.3. Islam, India, and China in the Medieval World System
Marshall Hodgson (1974) wrote his three-volume Venture of Islam tour
de force some years after we had shared and apartment together in
Chicago in 1953. But I remember him telling me then that he already
regarded the Islamic regions as the center of a world-historical drama.
Reuben Smith, the editor of this posthumously published work on Is-
lam, mentions in the preface that Hodgson was simultaneously work-
ing on a world history as well. The manuscript, unfortunately, remained
in an unpublishable state at the author's death. But although the pub-
lished work is deliberately confined to Islam, Hodgson's world-historical
perspective shows through, and I believe, enriches the interpretation
of his historical material. In his chapter on “The World Before Islam?
and under the subtitle “Cosmopolitan and Mercantile Tendencies in
Early Agrarianate Society” Hodgson writes, for instance, that
change was steady and far-reaching; and from very early such
changes were interdependent across a large part of the Afro-
Eurasian land mass. Local societies were less and less indepen-
dent in their cultural development. ... Through the Afro-
Eurasian land mass, all peoples involved came to be historically
interrelated to some degree, tied into the trade network and
subject to at least the indirect impact of the historical devel-
opments that arose in the cited regions in the older agriculturalAndre Gunder Frank
areas. . . . At the expense of some schematizing, Afro-Eurasian
citied history thenceforth can be presented as the interdepen-
dent and more or less parallel development of four major com-
plexes of civilized traditions [Chinese, Indic, European, and
Nile-to-Oxus]. . . . As these regions were in contact, there was
much mutual influence and even sharing of common heritages,
for instance in commerce and art, in religion and science. Still
more important, perhaps, than simple borrowing among the
regions was their common historical context. . . . In sum, the
whole Afro-Eurasian Oikoumene was the stage on which was
played all civilized history, including that of Islamic civilization
. .. (Hodgson, 1974: I, 109-14).
Hodgson proceeds to trace the birth, rise, expansion, and subse-
quent leveling off of Islam within this Afro-Eurasian Oikoumene, Of
course, I cannot do the same here. I can mention only a few high points
selected for their relevance to the construction of a world-historical
system. Hodgson repeatedly emphasizes the unity of the Nile-to-Oxus
(flowing into the Aral Sea) region, and its Fertile Crescent and Iranian
highlands core, where Cuneiform and Aramaic languages predomi-
nated. From very early, this region was climatically and geographically
favored, relative to more arid regions to the southwest and northeast.
The region was also a focal point or crossroads for the commerce of
goods and ideas, not to mention the migration and invasion of peoples
from some of these other regions. The region took advantage of these,
and, at least in part, became the site of earlier great civilizations and
major religions. Islam would rise to become another.
Maurice Lombard (1975) sees the Islamic area and period similarly,
but even more so as a single (world) economic system. He writes in
the
“Introduction” and “Conclusion” to his The Golden Age of Islam:
The centre of the Muslim World was situated in the Isthmus
region, bounded by the Persian Gulf, the Red Sea, the Med-
iterranean, the Black Sea, and the Caspian Sea. It was, there-
fore, set at the intersection of two major economic units: the
Indian Ocean and the Mediterranean area. These two territor-
ies, united in Hellenistic times, but later split into two rival
worlds, the Roman-Byzantine and the Parthico-Sassanian, were
now reunited by the Muslim conquest, so as to form a new,5,000 YEARS OF WORLD SYSTEM HISTORY 197
vast territory which was economically onc. This unity rested
on large-scale trading relations along caravan and maritime
routes, on one currency, the Muslim dinar, and one interna-
tional commercial language, Arabic . . . (Lombard, 1975: 9-10).
The network of capital cities constituted the economic, social,
and cultural framework of the Muslim World. From the eighth
to the eleventh centuries the main points along this central axis,
namely Baghdad, Damascus, Cairo, Kairouan, Fez, and Pa-
lermo, all of them important staging-posts on the route from
Samarkand to Cordoba, bore witness to the amazing unity of
a syncretic civilization with its vast movements of men, mer-
chandise, and ideas, a civilization superimposed on the older
regional, rural, or nomadic background. . . . [This] created a
vast common market stretching from central Asia to the Indian
Ocean, from the Sudan to the barbarian West and the region
of the Russian rivers. This large unit which overlapped the three
former areas, the Sassanid empire, Byzantine Syria, and Egypt,
and the western, barbarized Mediterranean, enjoyed an influx
of gold, a plentiful supply of slaves (Turks, Africans, Slavs), and
a network of major trade routes which stretched from China
to Spain and from black Africa to central Asia and encouraged
tremendous urban expansion (Lombard, 1975: 120).
Moreover, this “common market” was fueled by growing urban de-
mand and rested on a truly international division of labor with a unified
monetary system. Lombard details the generally eastward flow of gold
and silver, in which the West lost the gold and silver to Byzantium,
which was also dependent on Africa for them, and which lost it in turn
to the Sassanid east, from where it flowed further east. Lombard an-
alyzes the trade in cereals — including wheat, barley, sorghum, and rice
(1,000 camel loads of cereals daily from Beja to Kairouan and Tunis)—
especially into Baghdad, whose population may have reached two mil-
lion; other foodstuffs like sugar, dates, olives, grapes/wine (despite Is-
lam); stock rearing, especially of sheep, horses, and camels; timber
and other forest products, like wood/charcoal for smelting metal ores,
and the different metals themselves, all of which were in especially short
supply in the Fertile Crescent; wool, flax, cotton (from India), silk tex-
tiles and dyes for tents and carpeting, as well as clothing; stone, clay,198 Andre Gunder Frank
and glass products; salt and other sca products; writing materials; and
medicines. Armenian, Italian, and other Christians, and especially Rad-
hanite Jews played essential roles as middlemen.
Last but not least, there was a major trade in slaves. “Following
the example of the Ancient World and the Byzantine Empire, the Mus-
lim World was a civilization based on slavery. . . . But no slaves were
to be had inside the Muslim world. . . . The slave-traffic was, then,
of the utmost commercial importance” (Lombard, 1975: 194-96). These
slaves were imported from African Ethiopia, Nubia, and Sudan; cen-
tral Asian Turkestan; Slavic eastern Europe; but also from Frankish
western Europe, wherein Jewish merchants managed the western trans-
Mediterranean route and whose hub of the trans-Alpine and east Med-
iterranean slave trade was Venice.
The second half of the eleventh century did bring crises, dis-
turbances, invasions, and with them urban decline and the dis-
ruption of trade. Whole districts of Baghdad and Cairo stood
in ruins; people left. . . .
After the eleventh century the center of gravity of the Ancient
World swung from one place to another. From now on, the nerve
centres and centres of influence of an expanding economy were
no longer in the East, in the cities of the Muslim World. They
moved westward and became established in the mercantile ci-
ties of Italy and Flanders and, half-way along the great trade
route linking them with each other, in the trade fairs of Cham-
pagne .. . (Lombard, 1975: 236-37).
At this point, I would express only three reservations to this intro-
ductory account. One is that, as Lombard himself shows later in his
book, and as, following him, we will examine in the proposed histor-
ical review, the Muslim world and its economic system extended even
farther. Moreover, the changing Muslim world did not exist in isola-
tion. In his own examination of The Venture of Islam, Hodgson empha-
sized its relations with China.
The period was one of great prosperity. It is not clear how far
this was the case throughout the Afro-Eurasian Oikomene, but
at least in China at that time what may be called a “commercial
revolution” was taking place. Under the strong government of5,000 YEARS OF WORLD SYSTEM HISTORY 199
the Tang dynasty ... commerce became much more exten-
sive and more highly organized. . . . Chinese economic activ-
ity was directly reflected in the trade in the Southern Seas (the
Indian Ocean and eastward), where Chinese ports became an
important terminus for Muslim vessels. More generally, at this
time began a long period of limited but unmistakable Chinese
cultural ascendancy in the Oikomene as a whole—replacing
the recent ascendancy of Indic impulses and the still earlier as-
cendancy of Hellenism. . . . It can be surmised that the com-
mercial life of the lands of Muslim rule was given a positive
impetus by the great activity in China, especially considering
its important connections with China both via the Southern
Seas and overland through central Eurasia. In any case, com-
merce also enjoyed the great benefits of extended peace (Hodg-
son, 1974: I, 234-35).
The second reservation to Lombard’s account is the early dating
of the shift in the center of gravity and the omission in the meantime
in Lombard’s account of China and the Mongols. The third reserva-
tion is that the two earlier areas and periods, which Lombard distin-
guishes, can also be seen as belonging to a single system and historical
process if we look at them from the wider world systemic and longer
historical perspective I propose here.
K. N. Chaudhuris (1985) subject is Trade and Civilization in the Indian
Ocean: An Economic History from the Rise of Islam to 1750. Nonetheless, he
also emphasizes the importance of T’ang China and the relation of
the overland trade to his subject.
‘The expansion and the new activities which became faintly ev-
ident in the rhythm of both caravan and trans-oceanic trade
from the seventh century onwards in northern and southern
China received a great deal of impetus from the domestic as-
pirations and developments of the T’ang and Sung empires.
However, in the West it was joined by the second and most pow-
erful of the historical forces of the time, the rise of Islam and
its expansion across the fertile lands of the Near East and South
Asia... (Chaudhuri, 1985: 34).
‘The Arab and Chinese geographical texts dating from the early
centuries clearly speak of the hazards of the sea. .. . Bounded200
the
paii
Andre Gunder Frank
by the Pacific at one end and extending all the way to the Med-
iterranean at the other, it was this latter circuit of trade [China
Sea, Bay of Bengal, and Indian Ocean] that constituted the
foundation of pre-Columbian world economy in both the east
and the west. In Asia commercial traffic was in the hands of
highly skilled professional merchants, who operated as private
individuals with little substantive state support. . . . The Asian
ruling classes sometimes took part in shorter commercial voy-
ages when opportunities for profit were clearly visible (Chaud-
huri, 1985: 15-16).
The medieval trade of Asia was really founded on the economic
and social acceptance of the four great products of eastern
civilisation—silk, porcelain, sandalwood, and black pepper—
which were exchanged for incense (Arabian gum resins), thor-
oughbred horses, ivory, cotton textiles, and metal goods. . . .
The scalanes of dhe Indian Ovean were supplemented all
through history by the northern caravan routes with sturdy two-
humped camels, horses, and mules as transport (Chaudhuri,
1985: 39).
Although the social position of the trading communities was
never very high in any Asian country, the ruling elites were
totally dependent on them for the fulfillment of two essential
functions . . . to create a market system or to centralize the tax-
ation system. . . . Furthermore, rare and luxury commodities
. . could be obtained only through trade. Historians on the
whole have avoided analysing the fundamental factors which
propel long-distance trade (Chaudhuri, 1985: 16).
Jacques Genet (1982) and Wolfram Eberhard (1977) surely rank among
most authoritative Western historians of China. Yet both are at
ins to emphasize the relations between China, her neighbors, and
other parts of the world, and their importance for the development
and our understanding of China itself—and of the rest of the world.
Th
us, Eberhard explains in his “Preface” that he has “also been con-
cerned not to leave out of account China's relations with her neigh-
bors” near and far (1977: xviii). For his part, Gernet already stresses
in his “Introduction” that5,000 YEARS OF WORLD SYSTEM HISTORY
an important place must be given to the influence of distant
civilizations. The Far East in general and the Chinese world
in particular were in contact throughout their history with the
western and southern parts of the Eurasian continent .. .
[through] the great commercial currents— maritime and car-
avan traffic; ... the great military expansions and of diplo-
matic relations; and . . . the spread of the great religions and
of pilgrimages... The big Chinese cities—especially the
capitals—have always been in every age cosmopolitan cities
(1982: 19-20).
201
Both authors, as well as others, also demonstrate the economic and
commercial basis of many military, diplomatic, and other political,
well as cultural, relations across Eurasia overland and around it
sea. A systematic political, economic, and commercial network is
as
by
at-
tested to by migrations, invasions, alliances, diplomacy, pilgrimages,
technological diffusione (and come attempte to restrain or monopolize
it), tribute payments, caravan and maritime trade, and colonies of many
thousands of both resident and itinerary merchants (and sometimes
the slaughter of some thousands of them recorded in historical doc-
uments) in each of many cities of west, inner, and east Asia (and
cluding China).
in-
This is not the place to retrace the “ins and outs” or “ups and downs”
of these relations between China and other parts of Eurasia (and Af
rica) over the years. However, even Gernet’s summary of these rela~
tions, especially for the medieval period, lends further support to the
thesis that trans-Eurasian relations, as far as China, were systematic
in a single world system.
During the whole period from the seventh to thirteenth centur-
ies the two great civilizations of Eurasia were those of Islam
and China. . . . Contacts between the Islamic world and the
Chinese world began in the Tang age and went up to the Mon-
gol age. . . . The first contacts took place at the time of the Arab
expansion in the area between Mesopotamia and Lake Balk-
has, from about 650 and 750. The whole policy of the T’ang
in this part of the world aimed at opposing the victorious ad-
vance of the Arabs, but China's various alliances . . . were un-
able to hold up their progress. . ..202
Andre Gunder Frank
‘The Chinese T’ang and Sung empires, the former continental
and warlike, the latter maritime and commercial, were contem-
poraneous with the Ommayad and Abbassid empires, and be-
long to the same period of Eurasian history. East Asia and the
Islamic world even seem to have evolved in the same way. . . .
Relations across central Asia were contemporaneous with the
maritime expansion of the Islamic world into the Indian Ocean
and as far as East Asia after the foundation of Baghdad in
762... . The Chinese expansion in Asia in the seventh and
eighth centuries resulted in increasing influence of T'ang civ-
ilization in all neighboring lands: central Asia, Tibet, Trans-
oxiania, Korea and Japan, and the countries of South-East Asia.
Certain elements of Chinese culture spread to the Turks. . .
Sino-Islamic contacts were to permit the transmission of cer-
tain techniques from East Asia to the Islamic world and thence
to Europe (1982: 287-89).
This expansion of the Chinese world’s relations with South-
East Asia and the Indian Ocean took place in a much wider
context. The Nanking dynasty’s interest in overseas countries
was contemporaneous with the expansion of Indo-Iranian sea-
faring activities and the development of commercial routes be-
tween the Middle East, the Indian Ocean, and South-East Asia.
This is the explanation of [their] progressive conversion to Hin-
duism (1982: 197).
In the opposite direction on the other hand,
the two big currents of civilization flowing from Persia and In-
dia mingled with and mutually enriched each other in the whole
area extending from Afghanistan to the valley of the Amu-
Darya and the oases of the Tarim basin. The most active mer-
chants in central Asia and North China came from Samarkand
(K’ang to the Chinese), Maimargh (Mi), Kish (Shih) and Buk-
hara (An). Their language, Sogdian, an eastern Iranian dialect
spoken on all the roads from the Amu-Darya basin to the Wei
valley [the center of Chinese power], was the great medium of
communication in central Asia. . . . Since the trade routes con-
tinued from Bukhara on to Merv and from Balkh to Herat,5,000 YEARS OF WORLD SYSTEM HISTORY
it is not surprising that Iranian influences penetrated fairly
widely into China. . . . The presence of an embassy from Sas-
sanid Persia in Chiang-an [the T’ang Chinese capital] is men-
tioned as early as 638. . . [and] the last Sassanid sovereign . .
demanded China's help against Arab attacks. . . . Even distant
Byzantium had considered the idea of an alliance with China
(1982: 282-83).
Buddhism conquered the greater part of the Asiatic continent,
by travelling along the commercial routes, carried with the great
flow of trade. ... The first translators of Buddhist texts into
Chinese were not Indians but Parthians, Sogdians and Indo-
Scythinans, or people born in China or the Chinese borderlands
of Sogdian or Indo-Scythina parents. . . . The distant memory
of Greek sculpture preserved in the folds of the drapery, the
poses and faces of certain Chinese and Japanese Buddhist stat-
ues is one of the finest proofe of the unity of our world (1982:
211, 213, 226).
3.3.4. World (System?) History
203
We finally meet the most shining angels in the holiest of my trinity.
I will only “introduce” them here in order to place them on my highest
pedestal. Some certainly need no introduction to historians, and I will
later draw on their work amply myself to make my own proposed
view of world system history.
re-
William McNeill (1964) takes us still further back and farther afield.
His monumental The Rise of the West: A History of the Human Community,
offers a reading of much more of a whole world history than I could
summarize here. I will draw heavily on him for my own historical
re-
view. At this point, I cite McNeill only to illustrate some oft-neglected
events in world system history prior to the Islamic apogee.
By 200 ap, Hellenism had lost almost all of its expansive en-
ergy. ... The Confucian tradition of China also underwent a
similar, though less drastic decay. . . . India, on the contrary,
entered upon an age of remarkable cultural expansion both at
home and abroad. . . . The spread of Mahayana Buddhism over-
land into central Asia, China, Korea, and ultimately to Japan
was the most spectacular demonstration. ... But an almost204
Andre Gunder Frank
equally vast movement followed the scaways .. . and planted
a series of states in southeast Asia and Indonesia . . . (McNeill,
1964: 361).
India therefore appears to have played the leading role in the
entire Eurasian world between 200 and 600 ap. . . . In one re-
spect, the expansion of Indian culture was very different from
the earlier expansion of Hellenism. Military conquest played
almost no part: merchants and missionaries took the place of
the armies of Macedon and Rome. Consequently, Indian ex-
pansion followed trade routes, particularly the sea route to south-
east Asia and the islands of Indonesia, and the overland trail
through the oases of central Asia to northwest China. Com-
munication between India and the lands of southeast Asia was
very ancient (McNeill, 1964: 362, 337).
Farther west, the Sassanian Persian empire replaced the previous
dominance of the Kushan one, which had its center to the northeast.
The other world-historically most significant development, however,
was yet another outward mobilization of central Asian peoples, who
Teft
outer Mongolia in several directions. They occupied northern China,
pushed the Ephthalites into India, engaged Sassanian Persia (which
rivaled Rome, but adapted with greater success), and moved west, push-
ing other peoples through Europe into the heart of the Roman Em-
pire, to whose decline they contributed. The name of Attila the Hun
has become a byword in the West. McNeill observes in a footnote and
concludes in the text that
the remarkable chronological coincidence suggests that displace-
ments of peoples within the central steppe, probably emanat-
ing from the region where the Juan-juan confederacy was coming
into being, pushed both Huns and Ephthalites from their ac-
customed pasture lands. . . . Indeed, the nomad challenge was
sufficiently important that the political history of ecumene may
in large part be understood as the consequences of the shifting
pressures brought from the steppes against the various segments
of the civilized world (1964: 388, 316-17).
The fall of Rome (in the West, since Byzantian Rome continued) was
due to many causes which are subject to many interpretations. Two,5,000 YEARS OF WORLD SYSTEM HISTORY 205
however, are of particular interest from the present perspective of world
systemic history. One is that of the above-mentioned barbaric invasions
from the north and east. However, these must be traced not only to
their central Asian origins as above, but also to the resistance of Han
Dynasty China, which perhaps deflected more migratory invasions west-
ward, and eventually into the Roman empire. Indeed, it may be ar-
gued that Han strength against nomad incursion was itself partly
supported by foreign trade, including that with Rome. Eventually, of
course, both Han China and Rome fell to tribal invaders, but not before
both had in turn been weakened by trade losses to their Indian and
Parthian intermediaries.
Thus, another cause of decline was the connections between China
and Rome themselves, particularly those via India and Persia. Joseph
Needham (1954-84) and others have documented and emphasized the
export of Chinese knowledge and goods to Rome, often via India. Hirth
(1885), Hudson (1931), and more recently Yu (1967), among others,
have documented the far-reaching trade connections. Particularly the
Chinese export of silk and import of woolen and linen textiles, glass,
Baltic amber, and Red Sea corals and pearls, as well as slaves and jug-
glers, from markets in the Roman empire, albeit probably from its
eastern rather than its western part. The Roman historian Marcelli-
nus claimed, perhaps with some exaggeration, that the use of silk “spread
to all classes without distinction, even to the lowest” (Stavrianos, 1970:
100). How else did Rome pay? Taylor summarizes these scholars findings
about
the silk trade with China. The economic consequences for the
‘West were considerable. Because Roman exports of wool, linen,
glass, and metalware to the East in the first two centuries A.D.
did not match in value the empire's imports of silk, spices, per-
fumes, gems, and other luxuries, the West suffered seriously
from an adverse balance of trade. Gold and silver had to be
continually exported to Asia. Late in the first century a.p. Pliny
estimated that India, China, and Arabia drained away annu-
ally at least 100,000,000 sesterces, declaring: “That is the sum
which our luxuries and our women cost us.” The discovery of
large hoards of Roman coins in India supports Pliny’s state-
ment of the drain of specie. Few Roman coins have been found206 Andre Gunder Frank
in China, probably because the Romans purchased the silk pro-
duced in China from Parthians, Kushans, and other middle-
men in western Asia. A British scholar has estimated that
between 31 5.c. and 192 a.p. alone, Rome's trade with the Or-
ient cost her a net money loss of some £100,000,000. (This se-
rious drain took place at a time when the known sources of
gold and silver supply within the empire were being exhausted,
and must be reckoned as a major factor in the economic de-
cline of the Roman world.) (1987: 43).
This account leaves unmentioned the consequences for Rome (not to
mention for her colonial subjects), of the Roman exertions and expen-
ditures to maintain and expand her empire necessary to guarantee,
produce, and drain away the surplus for these remissions to cover its
commercial deficit with the Orient. (Women today might, of course,
not agree with Pliny’s cost/benefit calculation about their female an-
vestors.)
‘The previous period is known in Westernized world history as that
of Hellenic culture in Greece and Rome. According to this latter-day
Westernized Eurocentric reading, this culture and empire developed
in situ, as though it had emerged out of the head of Zeus, like Pallas
Athene. Martin Bernal (1987), having recently unmasked this west Eu-
ropean (re)writing of history, argues for an alternative reading of Black
Athena: The Afroasiatic Roots of Classical Civilization. The historical record
seems to support him, For, as McNeill writes, Greek and Roman art
styles can be traced over the Silk Roads to China, and there were also
longstanding Indian connections, which in turn were tied to Persia.
Of course, Greek culture also spread eastward. Besides, Greek was
an Indo-European language.
McNeill discusses the Hellenic period:
‘The keystone of the ecumene was Bactria, where tendrils from
the Hellenistic, Indian, Chinese, and Iranian styles of civiliza-
tion met and intertwined. . . . A propitious political climate for
the caravan trade . . . came to link India, China, and the Mid-
dle East more closely than ever before (1964: 330, 320).
By the first century a.p. direct voyages across the full width
of the Arabian Sea, from the Straits of Aden to the southern-5,000 YEARS OF WORLD SYSTEM HISTORY 207
most India, had become matters of routine. As a result, trans-
port between the Mediterranean and the Indian Ocean became
vastly cheaper as well as faster; and a greatly enlarged flow of
goods between these regions funneled through Alexandria by
Egypt. This long-distance trade kept Alexandria prosperous even
when its immediate hinterland in the Nile Valley suffered eco-
nomic retrogression under Roman rule.
Similar voyages beckoned farther east; and on the eastern shores
of the Bay of Bengal, enterprising merchants established con-
tacts with other seamen operating from the south China coast.
The result was to link China with India by sea, just as the Med-
iterranean has (perhaps only slightly earlier) been linked to In-
dia. By the second century a.b. (and possibly before) . . . [they
were] merged into one more or less continuous commercial net-
work, spanning the entire breadth of the southern seas between
Africa and south China. Thug, in the latter decades of the sec-
ond century 8.c. China consciously entered into regular con-
tact with other civilizations of Eurasia. Organized trade routes,
both by land and by sea, soon linked the four great cultures
of the continent. In addition, Eurasia’s central sea of grass pro-
vided a third linkage, more sensitive to military than to mer-
cantile enterprise, By the fourth century 8.c. horse nomadry
had spread to all the tribes of the steppe. . . . Thus, the Eura-
sian ccumenc was closed as never before (1964: 296-97).
However, McNeill still stopped short of reading the world history of
the last three thousand years as that of a world system or “central civ-
ilization? as David Wilkinson (see below) calls it. From this perspec-
tive, I still have some important reservations about McNeill’ earlier
reading of world history.
One reservation is that McNeill saw things too exclusively in po-
litical terms. Here and there the economic infrastructure peers through
his account, but it only gets very marginal attention. We need to de-
vote more attention to the world economy of central civilization and
to integrate that with political and cultural history. McNeill writes me
in correspondence that [AD post-1500] world system writers lend too
much weight to economic matters and not enough to political (includ-
ing military) and cultural ones. That may be. But if so, most historians208 Andre Gunder Frank
of the earlier periods make the opposite mistake and pay too little at-
tention to economic forces and relations. More economic records are
available than have been used. Writing was supposedly developed to
keep economic records, and archaeologists have deciphered many. Even
secondary sources on economic relations have been neglected by his-
torians intent on writing political and “civilizational” history. Popula-
tion changes, invasions, migrations, trade, the diffusion of military and
other technology, religious and other cultural exchanges, etc., all con-
tribute to developing the forces and changing the relations of produc-
tion. So do changing ecological “conditions” of production (to which
O'Connor [1988] refers in his work on economy and environment).
I don't see why these matters should be considered separately, rather
than as an essential part of the cement and indeed motor force of the
workings of the system as a whole.
For instance, in his examination of the same Hellenic world and
period, which was reviewed by McNeill above, Hodgson attributes more
importance to trade and other economic relations.
In the sixth century Bc even more lands from the Nile to Oxus
were absorbed into the Archmenid Persian empire, which united
in peaceful interchange the many peoples from the Aegean to
the Indus. Under it Aramaic merchants received full mercan-
tile freedom, and Aramaic became the chief administrative
tongue. . . . The various lands in the region were closely inter-
related; to some degree they underwent a common destiny. I
think this resulted at least partly from the role of the mercantile
classes in the regional life. . . . Increasingly over the centuries
there was opportunity for the high culture focused on the mar-
ket to become more autonomous and even more influential in
the society asa whole. . . . It was doubtless a mercantile poten-
tiality that, at one stage of Oikoumenic development, made it
possible for Greek culture to gain such a hold in the region
(Hodgson, 1974: 118-24).
And in his review of the end of the period, Hodgson writes:
The mercantile development represented in part a response to
the ever quickening pattern of trade throughout the Afro-
Eurasian Oikomene. Direct trade by sea and land between China5,000 YEARS OF WORLD SYSTEM HISTORY
and the Indo-Mediterranean regions opened . . . [and] quickly
became important commercially and financially. Trade elsewhere
in the Southern Seas (the seas of the Indian Ocean and east-
ward) had likewise expanded, as had trade both north of the
Mediterranean in Europe and south towards the Sahara. The
peoples from Nile to Oxus not only took full advantage of their
crossroads, but also helped develop new fields of trade. It was
in this period that Greek, [the] premier commercial language
. . . was displaced in favor of elements from the Fertile Cres-
cent. Along the central Eurasian trade routes, it was Iranian
and Semitic culture and religion that came to dominate locally
even more than Chinese or Indic. . . . In the west Mediterra-
nean and its hinterland, the same elements—‘Syrians” and
Jews—became carriers of mercantile culture ... (1974: I,
142-43).
L. S. Stavrianos (1970) argues that
at the basis of the new Eurasian ecumenism was technological
advance. ... When hoes and axes and plows, as well as weap-
ons, could be made of iron [which was more widely available
and cheaper than bronze], the economic, social, and political
repercussions were immediate and farreaching. This stage was
reached slowly—about 800 8.c. in India, 750 x.c. in Central
Europe, and 600 s.c. in China. In these and other regions the
advent of cheap iron led first and foremost to the cutting down
of heavy forests hitherto invulnerable to stone-edged axes and
wooden plows. But farmers were now able to use their strong
and sharp iron axes and iron-shod plows to extend agriculture
- + + (1970: 88).
This expansion of the frontiers of agriculture made possible
a corresponding expansion of the ecumene of civilization, which
grew more in the half millennium from 1000 to 500 x.c. than
in the preceding three millennia from 4000 to 100 b.c. The
basic reason for this was the tremendous increase in produc
tivity that now took place. ... The jump in agricultural pro-
ductivity meant that surplus was now available for economic
development and state-building purposes . . . (1970: 88).
209210 Andre Gunder Frank
The net effect was a great stimulus for all kinds of commerce,
a corresponding stimulus for manufacturing and agriculture,
and an overall increase in economic specialization with an at-
tendant rise in efficiency and productivity. The manufacturer
of cheap goods now had available for the first time a mass mar-
ket, while the small landholder could turn from subsistence ag-
riculture to specialized farming, whether the mulberry and
silkworm in China or olive oil in Greece. The new iron tools
also made possible the building of better and larger ships, which,
in turn led to longer voyages and to more trade and coloniza-
tion . . . (1970: 89).
Side by side with these economic developments were equally
significant social and political changes. The military aristocracy
. . . was being undermined by the new class of merchants, crafts-
men, and mariners. The old tribal society was being transformed
by monetizatiun, persunal services and allegiances were being
superseded by the exigencies of the market place. Equally dis-
ruptive was the political consolidation made possible by the eco-
nomic growth. Tribal chiefs and their advisory councils and
assemblies were being replaced by kingdoms and then by em-
pires, whether in Italy or India or China . . . (1970: 89).
Such basic and all-inclusive disruption was unsettling and un-
comfortable. It led to soul searching —to the posing of new ques-
tions and the seeking of new answers... The answers
constituted the great philosophical, religious, and social systems
of the Classical Age. It was not happenstance, then, that the
spokesmen for these systems were all contemporaries— Confucius
in China, Buddha in India, Zoroaster in Persia, and rationalist
philosophers in Greece. In all these regions the disruption and
the challenge was the same, but the answers varied greatly . . .
(1970: 90).
Yet, it may also be argued that the religious answers and social solu-
tions to common problems were remarkably similar.
Therefore, another reservation to McNeill is that he wanted to
see four separate civilizations (Chinese, Indian, Mideastern, and Med-
iterranean/European) between 500 Bc and ap 1500. Yet he repeatedly5,000 YEARS OF WORLD SYSTEM HISTORY 2u
shows, as in the preceding quotations, that they were connected all
along. However, McNeill emphasizes and perhaps overvalues the dis-
tinctiveness of these civilizations and insufficiently values the
significance of their connections and indeed their unity. If we rewrote
MeNeill’s argument into a description of a single Eurasian
system/civilization, without thereby denying the civilizational distinc-
tions and cultural varieties, I think we would be a step ahead. As
Wilkinson says, the fact that one or more places declined does not
mean that the whole system disappeared. On the contrary, the com-
ing and going of the parts seems to be an essential mechanism of
the development of the whole.
A third problem with McNeill’s account of world history during
this period is that he virtually left sub-Saharan Africa out of the story.
With our contemporary sources, like Curtin, Davidson, and Fage, which
have become available since McNeill wrote, and with more reliance
on contemporary sources like Ibn Battuta, it is now possible to recon-
struct much more of the cultural, political, and economic participa
tion, and integration of Africa in this common history of a single
Eurasian-African (really Asian- African-European) world system, Mad-
agascar had been populated by Malays. The Axum empire in east Af-
rica had far-reaching international relations across the seas and into
Africa, and became Christianized. Zimbabwe gold was exported from
southern Africa; and west African gold and slaves entered the trans-
Saharan and then Mediterranean trade at least since the Phoenician-
founded Carthage became a turning point in the same. We need sys-
tematic accounts of African participation in the world system, and I
will make an effort in this direction in the proposed historical review.
(The “New World” is not treated here, although contacts during our
periods have been documented. Going farther back in history would
of course include the migration of Asian peoples eastward across the
Pacific.)
The other problem with McNeill’s world was that it had a hole
in it. Every now and then barbarian hordes emerged from the Eura-
sian steppes and McNeill recorded their arrival. But in his account,
they emerged as a deus ex machina from an “uncivilized” no-man’s
land. Yet these “barbarians” transformed McNeill’s four civilizations,
which by Wilkinson’s and my reading formed a single
civilization/system. McNeill repeatedly analyzes the consequences of212 Andre Gunder Frank
central Asian invasions and migration on and in the civilized regions
where they moved. But McNeill did not inquire into the causes which
led to these peoples/forces coming out; he treats them as if they came
from nowhere. In part, this omission may be because they left little
or no written records. However, in part, it probably was also because
McNeill, like most others, is fascinated by “civilization(s)? which ex-
clude the barbarians by definition. The comings and goings, or non-
comings, of the barbarians need to be systematically fitted into a larger
system (bigger than that including only the civilized peoples), within
which all of these interact. Needham’s work on China, and especially
the works of Owen Lattimore and others on inner Asian frontiers, help
make these connections, Therefore, I will make a special effort to in-
clude the central Asian peoples and their movements in my own read-
ing and writing of world system history.
The reason I have written of McNeill in the past tense above is
not that he has stopped writing, but on the contrary, because in re-
cently published articles, as well as in private correspondence with me,
he has been reconsidering his earlier thinking. He is now much more
inclined to look for and see a world system which goes much farther
back in history and includes more peoples, but given his incompar-
ably vaster erudition perhaps more cautiously.
In his article on “The Rise of the West after Twenty Five Years” McNeill
now writes that:
The central methodological weakness of my book is that while
it emphasizes interactions across civilizational boundaries, it
pays inadequate attention to the emergence of the ecumenical
world system within which we live today. . . . Being too much
preoccupied by the notion of “civilization? I bungled by not
giving the initial emergence of a trans-civilizational process the
sustained emphasis it deserved. . . . Somehow an appreciation
of the autonomy of separate civilizations (and of all the other
less massive and less skilled cultures of the earth) across the
past two thousand years needs to be combined with a portrait
of an emerging world system, connecting greater and greater
numbers of persons across civilizational boundaries.
‘To make this a feasible enterprise, one needs a clear and dit
tinct idea of the emergent world system as manifested first in5,000 YEARS OF WORLD SYSTEM HISTORY 213
the ancient Middle East and a second time in the modern world,
and one must reflect on how these intersected with the more
local civilizational and cultural landscapes they impinged
upon. . . . In the ancient Middle East, the resulting interactions
among peoples living in different landscapes, with diverse lan-
guages and other outward signs of civilized diversity, led to the
emergence of a cosmopolitan world system between 1700 and
500 ac. . . . There is a sense, indeed, in which the rise of civ-
ilizations in the Aegean (later Mediterranean) coastlands and
in India after 1500 ac were and remained part of the emergent
world system centered on the Middle East... . All three re-
gions and their peoples remained in close and uninterrupted
contact throughout the classical era. . . . [Moreover,] one may,
perhaps, assume that a similar [to the modern] primacy for ec-
onomic exchanges existed also in earlier times all the way back
[to] the earliest beginnings of civilization in ancient Mesopo-
tamia .. . (1990).
This recent reflection by McNeill, coming after his vast historical work
and reconstruction, can encourage us also to advance further in this
systemic reading of world history. So does the next author mentioned
below.
Alistair Taylor (1987-88) also commented on the early Abu-Lughod,
like McNeill and myself. Taylor also proposes
a different historical perspective — one that takes account of the
existence of a Eurasian commercial system antedating even Clas-
sical times—[with a] . . . continuum in Eurasia (and parts of
Africa) in which societal discontinuities occur periodically with-
out destroying the interregional system of commercial, cultural,
and other linkages. In adopting this counter-perspective, we
envisage any number of dynamic interconnected processes at
work: technological, political, economic, religious . . . (1987-88:
48).
Here, we can take a mere glimpse at the thousand-year history of parts
of this Eurasian system before classical times from 1500 to 500 ac.
Beyond the developments already reviewed by McNeill and Stavrianos,
the two most notable “events” were the two waves of nomadic migra-ait Andre Gunder Frank
tion and invasion between 1700 and 1500 sc and between 1200 and
1000 sc, and their respective impacts on the more settled peoples to
the east, south, and west. The first wave used bronze weapons and
two-spoked horse-drawn carriages to invade China and India. Indo-
European Hittites and Kassites settled in Asia Minor and the Levant,
the Hurrians assimilated in Mesopotamia, and the Semitic Hyksos
conquered Egypt.
The second wave of arrivals rode (by then larger domesticated) hor-
ses and used iron weapons. Again, they moved eastward into China,
southward to India, and southwestward. Semitic Phoenicians settled
on the east Mediterranean coast, while the Aramaenas and Hebrews
settled to the South of them. Other invaders arrived farther east; the
Dorians in Greece, from where they also expanded eastward. Out of
these developments emerged new Assyrian and then Chaldean empires
centered on Niniveh and Babylon, and later a new Persian empire and
Classical Greece.
The Phoenicians became traders par excellence and dominated com-
merce and new settlements over a wide area around the Mediterra-
nean and beyond between 1200 and 700 sc. They founded Carthage,
Marseille, and Cadiz; they settled on Malta, Sicily, Sardinia, and the
Balaeric Islands; they traded in the Atlantic as far as Britain. They,
or their local descendants, also engaged in trans-Saharan trade from
Carthage and Utica, and traded in the Egyptian population and com-
mercial centers. The Phoenicians dominated the regional trade in Iber-
ian metals, west African gold and slaves, and much of the trade in
foodstuffs, textiles, dyes, timber, and pottery. Eventually, the Phoeni-
cians were overtaken by Greek advance, settlement, and commerce,
in the Black Sea region.
David Wilkinson (1987) writes in a more unitary-history sense about
what he calls the “Central Civilization.” It began in the west Asian part
of the Eurasian land mass and eventually encompassed the entire globe.
Central Civilization is the chief entity to which theories of class
society, the social system, world-economy and world systems
must apply if they are to apply at all. A suitable theoretical
account of its economic process does not yet exist; one for its
political process may . . . (1987: 56-57).
Wilkinson’s subtitles indicate his intent and recommended procedure:5,000 YEARS OF WORLD SYSTEM HISTORY 215
Recognizing Central Civilization as a Reality. . .. Recogniz~
ing a single entity in adjacent “civilizations”. . . . Recognizing
a single entity after civilizations collide. . . . Recognizing a sin-
gle entity when “civilizations” succeed each other. . . . Did Cen-
tral Civilization Ever Fall? (1987: 35-39).
Wilkinson’s answer is “no? since its birth was when Sumer and Egypt
joined hands around 1300 sc. However, as Stedman Noble (1989),
among others, suggests, even before that Egypt probably could not
have developed as it did without its ties to Mesopotamian Sumer via
the Levantine shores of the Mediterranean. Wilkinson argues that the
rise and decline of particular regional urban centers and their rural
hinterlands does not spell the rise and decline of what he calls Central
Civilization. It continues and is even enriched by the ups and downs
of its various parts. Thus civilization —historical world system in my
terminology—has continued and developed through relations includ-
ing collisions between civilizations, successions and combinations of
“different civilizations” and the decline and fall of some city-centered
region(s), empire(s) or civilization(s) and its/their replacement by others.
Central Civilization is the chief entity to which theories of class
society, the social system, world-economy and world systems
must apply if they are to apply at all (1987: 56-57).
Inan unpublished manuscript concluding part of the same article, Wil-
Kinson poses “a centralist challenge to centralists” to overhaul thoroughly
pluralist textbooks and curricula. He proposes eighteen successive but
continuous historical processes from “when [Egyptian and Mesopota-
mian] civilizations collide” around 1500 sc to “the present moment and
the prospects for Central Civilization” as the possible skeleton of a course
in world history, comparative civilizations, or world civilization, which
he would call “Central Civilization: The Origins of World Civilization”
Ina more recent paper, Wilkinson offers numerous additional refiec~
tions which can further guide our inquiry into the history of the world
system.
For whatever reason, the Central economy is at all times a mixed
political economy, embodying trade and war, coercion and bar-
gaining, the one-few-and-many. The balance shifts with time,
scale, region and commodity. And possibly other variables. The216
Andre Gunder Frank
determinant of the mix need study. . . . Whenever it is possible
to map the distribution of wealth in Central ilization, in-
equality prominently appears: by city, by region, by political
power, by inheritance, in law, by age and family status, by gender.
‘The several inequalities do not appear to be reducible to any
one fundamental root inequality.
The entry into the Central world-economy of fish, wheat, oil
and wine, suggests mass consumption driven either by political
redistribution (to hire the loyalty of armed men, clients, voters,
etc.) or by markets, probably varyingly by both. Luxury goods
may also have spread more widely through the social struc-
ture. . . . World-economic commodities in Central Civilization
have tended strongly to be elite goods—luxury food, clothing,
shelter, and display items—along with the trade tools of elite-
supporting soldiers and bureaucrats (weapon-metal; paper for
recordkeeping). Elites, dasses, aud the assuciated inequalities
must not be treated as recent phenomena. Early Central trade
in precious metals may, and coinage does, imply the develop-
ment of mobile free persons, merchant classes and economic
vs. politico-military elites, characterized by private property in
portable wealth. These elements of capitalism similarly must
not be treated as of recent vintage.
‘The general trend over time is clearly toward a continuing in-
crease in the number and variety of commodities trades in the
world economy of Central Civilization. Within this trend there
are temporary and permanent commodity dropouts, shifts in
regional contributions, epochs of faster and slower commodity
increase; but the trend remains. Commodities and commod-
ification too precede modernity, and must be attributed to some
early cause, perhaps simply to civilization’s division of labor,
increased scale, and increased population. . . .
The increase in the number and variety of commodities over
time is one piece of evidence for a secular trend to expansion
in Central Civilization over the past 5000 years. . . . There ex-
isted an Old World ecumenical macroeconomy, a multicivili-
zational structure which apparently provided the highest-level5,000 YEARS OF WORLD SYSTEM HISTORY 217
largest-scale economic order until the global reach of Central
Civilization. . . . This economy was larger than any polity.
It may require theoretical treatment as a whole . . . (1988: 51-55).
Thus, Wilkinson takes another big step toward a reading of historical
continuity of a single world system encompassing much of the Eura-
sian land mass. Wilkinson goes to pains to establish what should be
intuitively obvious; that long-lasting recurrent (armed political or eco-
nomic) conflict between and among peoples and “societies” suggests
that they also have long-lasting systemic contact and relations. By the
above-mentioned criteria of system and continuity, not only is “co-
operative” trade and migration evidence of systemic continuity, but
equally or even more so are invasions, defenses against them, and other
recurrent military and political conflict, not to mention cultural diffu-
sion and imposition. For the history of none of the participating peo-
ples and regions would have been the same without this interaction
among them, and the historical course of the whole was, and remains
more than, the sum of its parts.
This also suggests that Wilkinson’s following thesis twelve must also
be applied beyond Central Civilization: “the Central economy is at all
times a mixed political economy, embodying trade and war, coercion
and bargaining, the one-few-many. The balance shifts with time, scale,
region, commodity” Indeed, the whole economic dimension must be
applied to central Asia and its relations with the east-south-western civ-
ilized areas, At the same time, we must agree with Wilkinson's theses
ten and eleven that in this sense “the world economy of Central civilization
has never been fully statist . . . [nor] fully capitalist? This implies not
only his thesis eight “that Central Civilization has never been completely
penetrated by any particular ‘instrument of expansion’ . . . or ‘produc-
tion’ (in the Marxist sense), but also, as I already argued, that we should
abandon most of the Marxist fixation (but de facto also that of the ideo-
logical “free marketers”) with modes of production. I think this fixation
is an ideological snare and delusion which obscures more than it clarifies
social relations. This fixation especially hides the essential continuity
and operation of the historical world system.
To make his case for continuity, Wilkinson uses McEvedy’s (1967)
Penguin Atlases of Ancient and Medieval History to construct turnover
tables of the rise and decline of ancient and medieval cities and their218 Andre Gunder Frank
hinterlands, which demonstrate their system-wide continuity. Of course,
the same procedure can and should be applied to regions, empires,
and “civilizations.” That way we can perhaps demonstrate that the rise
and decline of individual regional centers and empires, and the conflicts
and successions among them, are part and parcel of the evolutionary
course of world systemic history itself, and analyze how this worked.
Wilkinson's emphasis on urban-centered ‘civilization? however, excludes
the nomadic peoples. Thus, his procedure largely omits the important
role of the central Asian peoples, who invaded and migrated into the
“civilized” areas and who at home and abroad served as partners and
way stations along the overland Silk Roads.
UNESCO, at the initiative of Eiji Hattori, has organized a world-
wide research and seminar project to reconstruct the Silk Roads. These
various overland and maritime routes, which for millennia connected
the peoples of the Eurasian and African land mass, along with the mi-
grations and invasions along and across them, may be regarded as the
backbone and rib case of this world system. The ever-changing inter-
action between central Asian and neighboring peoples to the east, south,
and west was the pulsating heartbeat of this (single) system. And the
cultures and great religions of their civilizations may then be regarded
as the ever-renewed mind and spirit of the peoples in this long-evolving
historical world system.
I finally end this (perhaps excessively lengthy) review of the rela-
tions between the views of others concerning world system history and
my own. To do so, I select a long citation whose content, if not alto-
gether the procedure recommended, is closest to mine. Yet even to this
statement, I shall express some reservations and amendments below.
L. S. Stavrianos (1970) begins The Werld to 1500: A Global History with
his “Introduction: Nature of World History”:
The distinctive feature of this book is that it is a world history.
It deals with the entire globe rather than some one country or
region. It is concerned not with Western man or non-Western
man, but with all mankind. The viewpoint is that of an observer
perched on the moon, surveying our planet as a whole, rather
than that of one who is ensconced in London or Paris, or for
that matter, in Peking or Delhi. [Stavrianos himself is now in
San Diego, California].5,000 YEARS OF WORLD SYSTEM HISTORY
The global approach to history represents a new departure in
modern historiography. . . . In recent years, interest in world
history has been growing largely as a reaction to the manifestly
global sweep of contemporaneous events. . .. World history is
manifestly essential for the understanding of a world that has
become “one” in reality as well as in rhetoric.
This utilitarian function, however, is not the only reason for
turning to world history. Equally important is the fact that the
story of man from its very beginnings has a basic unity that
must be recognized and respected. Neither Western nor non-
Western history may be properly comprehended without a glo-
bal overview encompassing both. Only then is it possible to per-
ceive the interaction amongst all peoples at all times, and the
primary role of that interaction in determining the course of
human history. . . .
World history is not the sum of histories of the civilizations of
the world, in the same manner that Western history is not the
sum of the histories of the countries of the West. . . . [Similarly]
the geography of world history is . . . [not] the sum of the con-
tinents comprising the earth’s surface . . . [which would be] as
mechanical and as misleading as the former. . . . For the same
reason that the structure of world history requires focusing on
historical movements that have had major influence on man’s
development, so the geography of world history requires focus-
ing on those regions that initiated those historical movements.
When this is done, one land unit stands out uniquely and un-
challengeable: Eurasia, the veritable heartland of world history
since Neolithic times. . . . To an overwhelming degree, the his-
tory of man is the history of these Eurasian civilizations.
‘The Central Eurasian steppes comprise the endless grasslands
stretching from Manchuria in the east to Hungary in the west
and provide an overland channel of communication among the
centers of civilization strung out in Eurasia’s periphery. . . . Thus
the history of Eurasia was to a great extent molded by this inter-
action between nomadic tribes and sedentary civilizations. . . .
The ancient, the classical, the medieval periods of pre-1500 Eu-
219220 Andre Gunder Frank
rasian history . . . were heralded by major turning points pri-
marily attributable to these nomadic invasions (1970: 3-6).
Thus, Stavrianos still gives less emphasis than I would like to world
systemic structure and to simultaneous interconnections in the world-
historical process. Therefore, this reading of world history also sees
less cycles and crisis-generated critical turning points in the historical
process than I suspect further inquiry might reveal. However, it does
stress the essential unity of the world-historical process, the centrality
of central Asia, and above all that the whole is more than the sum
of its parts.
3.3.5. Extending World System History in Space and Time
‘To extend world system history still further in space and time, as
suggested above, we will have to draw on diverse historical and arche-
ological evidence. However, three well-known authors, Owen Latti-
more, Gordon Childe, and Philip Curtin, can serve as guides. Their
readings of history imply that following Fairbank’s second rule will
bear additional earlier fruits.
Owen Lattimore (1962) built much of his reputation studying the Jn-
ner Asian Frontiers of China and the relations of nomad tribal and more
sedentary civilized peoples across them. For my present purposes of
including central Asia and its peoples in the world system, it will be
useful to cite Lattimore on three points: great walls, interaction across
them, and the cycle of the same.
The walls were manifestations of the constant, or at least recurrent,
relations among Asian peoples. Lattimore (1962: xliv-xlvii), citing the
Russian writer Barthold and others, argues that there were many “great
wall” fortification systems other than the well-known Chinese one. They
were built at one time or another in many places along the whole stretch
of central Asian borders from the Yellow Sea to the Black Sea. The
great empires had extended their agricultural base northward, and were
dependent on it, and tribute from it, for their survival. Therefore, Lat-
timore argues, “it was profitable for many individual subjects of the
empire to cross the line of demarcation and to throw their lot in with
the barbarians” (1962: xlvi). In particular, the rulers of the civilized
empires to the south sought to avoid the export of grain staples to their
independent barbarian nomad neighbors to the north. The “‘great walls?5,000 YEARS OF WORLD SYSTEM HISTORY 2ai
therefore, had a function of keeping in the subjects of the empires as
well as in keeping out the barbarians” (Lattimore, 1962: xlvi).
Indeed, Lattimore lays great stress on the interaction, and its cy-
clical recurrence, between inner Asian nomadic peoples, especially those
bordering the great civilizations, and these latter. Ending his intro-
duction to the second edition of this book, Lattimore says:
It was the interaction of conditions that favored the barbarians
. . with the conditions that favored the great civilized empires;
it was this interaction, and not the fluctuations of climate, that
for a whole age of history set the rhythm of the pulse of Asia
and Europe (1962: xlvii-xlviii).
At the conclusion of the book, he suggests:
There remains a further point to be considered: whether the
two interacting cycles [among barbarians and among civiliza-
tions] have independent phases, were independent in origin.
Probably not. Inasmuch as the evolution of the Chinese agri-
culture and society, by the pressure it put on the people of the
steppe margin, helped create the true steppe society, the no-
mad cycle was at least in part a product of the Chinese cycle.
Once established, of course, the nomad cycle acquired a vigor
that enabled it to interact on the cyclical history of China with
independent force (1962: 550-51).
Huntington’s books, Pulse of Asia and Civilization and Climate, had pre-
viously attributed these recurrent waves of tribal migrations to 640-year
climatic cycles in Asia, which sent the barbarian hordes outward, and
of course, each people forced the next farther out. As against
Huntington's climatic explanation, Lattimore, as cited above, and Jo-
seph Needham (1954-84) advance more social-interactive explanations.
Either way, or in any combination of these explanations, however, the
fact remains that this interaction in and across Asia was important
for all its peoples. So important was this interaction that none of the
Asian peoples’ history could or would have been the same without it.
That is, they were in the same system.
Similarly, the above-cited Gernet and Eberhard repeatedly empha-
size the importance of inner Asian interactive relations in the devel-222 Andre Gunder Frank
opment— indeed in the very formation—of both China and its neigh-
bors. Gernet argues that
it would be simplistic to see in the Great Wall a sharp divide
between the world of the nomadic cattle-raisers and that of the
Chinese farmers and townspeople. The northern frontiers of
the Chinese world formed a zone where the opposing modes
of life of the farmer and the herdsman mingled and com-
bined. . . . Just as certain tribes of herdsmen changed over to
agriculture, so some Han adopted the nomad’s mode of life. . . .
‘The Great Wall formed only one element in a much vaster com-
plex: allied tribes who collaborated in the defense against in-
cursions; outposts, forts, and advanced garrisons; military
colonies; lands developed by deported populations; horse breed-
ing, and so on... . assimilation, diplomatic combinations, and
commercial exchanges (1982: 122-23).
The contributions of the steppe, of Sino‘Tibetan borders, and
of South China were crucial in the formation of the Chinese
civilization. The influence of neighboring cultures made itself
felt in every domain— methods of harnessing horses, use of the
saddle and stirrup, [and much military technology]. . . . [Be-
yond borrowings, in particular the steppe peoples contributed
significantly to] the centralizing, statist tendencies which seem
to be bound up with the problem of defence against incursion
from the steppes, but also with the needs of colonization, dis-
tribution of land and irrigation of dry zones . . . (1982: 195, 174).
Gernet (1985) and Eberhard (1977) also demonstrate how, repeatedly,
(cyclical?) developments in the steppe regions brought on reactive Chi-
nese offensive and defensive alliances in the north and west, or retreats
and expansions southward, and greater thrusts into or retreats from
maritime expansion. However, “the inverse phenomenon was almost
certainly just as important” (Gernet, 1982: 199). Indeed, the Manchu,
Mongol, Chin, and many earlier rulers and dynasties in China came
from the steppes—and were Sinicized. The very names of China and
Cathay were derived from other Asian peoples.
‘The systemic interrelation of the tribal-nomadic and sedentary-
civilized peoples is further elucidated by McNeill and others, albeit5,000 YEARS OF WORLD SYSTEM HISTORY 223
from a civilizational perspective. McNeill associated cach wave of no-
madic invasion with a new military, technological, and organizational
innovation (chariots the first wave, iron the second, etc.), which made
the invaders irresistible. Of course, we may well wonder what made
all these nomad peoples “barbarians” if they were superior in (at least
some) technology, military prowess, and the therefore necessary social
and political organization, as to win against the “civilized” empires time
and again. Moreover, each barbarian influx was rapidly integrated and
“civilized” some time after their arrival.
McNeill also observes that, after each invasion, there occurred a
new civilizational flowering, often at the edge of the “system,” in part
as local reaction (Toynbee’s challenge and response?), and presumably
in part as a result of the new economic, political, cultural, and blood
infusions from the invaders. These (challenge and?) response civiliza-
tional developments include the original Mesopotamia-Egypt fusion
in 1500 nc (central to Wilkinson’s argument); Assyria after 1000 Bc;
Confucian Han China, Buddhist India, Sassanian Persia, classical
Greece, etc., around 500 sc; China, Kushan, Rome at (Christian time)
0; China, Persia, Islam, Byzantium, and even Dark Age western Eu-
rope after av 500; Ming China, Akbar Mogul India, the Ottomans,
Safavid Persia, southwest Europe after the Mongols, etc. The “barbar-
ian’-“civilizational’ relations in all of these cases require incorporation
into a continuous and systematic reading of civilizational history and
systemic continuity. The resulting, or related, “local” regional ups and
downs should perhaps be read as shifts in the core/center within a sin-
gle world historical system.
Other recent studies by Tom Hall, Philip Kohl, and Christopher
Beckwith, which I found since writing the above, reinforce my con-
cern with inner and central Asia, or central Eurasia, as Beckwith calls it.
It is this latter point that is crucial: the development of core areas
cannot be understood apart from the interconnections with its peripheral ar-
eas... . It follows from this redefinition of civilizations as the
entire field of interaction, that “barbarians? typically nomads,
played a role in the evolution of civilization. When one adds
that the Mongol Empire was one of the largest . . . empires
in human history, the importance of nomads in civilizational
change becomes inescapable. . . . Nomadic groups have had a224 Andre Gunder Prank
major influence on the course of civilizational change. The habit
of pushing barbarians beyond the pale . . . of intellectual con-
sideration is at best misleading, and profoundly disastrous to
understanding the processes and variations of civilizational ev-
olution . . . (Hall, 1989: 2, 15; emphasis in original).
Hall emphasizes the essential role of Osman nomads in the formation
of the Ottoman empire and civilization. More generally, Hall also cites
Philip Kohl on “The Balance of Trade in Southwestern Asia in the
Mid-Third Millennium s.c.” (1978: 489):
One purpose of this study of long-distance trade in southwest-
ern Asia was to show that even the earliest “pristine” example
of state formation cannot be explained entirely as an internal
process of social differentiation but must be viewed partly as
the product of a “world-economy” at different levels of devel-
opment which stretched at least from the Nile Valley and south-
eastern Europe in the west to Soviet Gentral Asia and the Indus
Valley in the east (Haal, 1989: 5).
Beckwith adds a reflective Eurasian world-embracing “Epilogue” to
his study of medieval Tibet and writes:
The evidence shows that, during the Early Middle Ages, the
Tibetan Empire and Frankish Western Europe were integral
parts of a civilized world which included the Islamic caliphate
and T’ang China and was “focused” (to adapt Pirenne's usage)
on Central Eurasia . . . (1987: 196).
Central Eurasia was the overwhelming focus of Arab and Chi-
nese foreign policy and the source of strong cultural influences
on them, There should be little doubt that Central Eurasia was
a most influential factor in the history of the medieval empires.
Scholars of earlier generations had good reason to look to that
region for their “missing link” in world history . . . (1987: 193-
94).
The history of Central Asia is of fundamental importance for
an understanding of Eurasian history (1987: 229).
Not surprisingly, Beckwith (1987: 194n, passim) also takes the strong-
est exception to “the widespread and frequently deliberate use of the5,000 YEARS OF WORLD SYSTEM HISTORY 225
pejorative term ‘barbarians’ for the peoples of Central Eurasia” I hope
that Beckwith would agree also to extending these reflections back-
ward in time to periods before the medieval one.
V. Gordon Childe (1942) stands out among the authors who long ago
pointed us to an earlier beginning of the world system in his Man Makes
Himself and What Happened in History. In his chapters on “Early Bronze
Age Civilization in Egypt and India” and “The Expansion of Civili-
zation? Childe repeatedly emphasizes that long before 1500 uc these
civilizations were essentially dependent on long distance trade—in “lux-
uries” as well as in necessities. For one place and time after another,
Childe speaks of the organized, regular, and also long-distance market
trade of raw materials in exchange for manufactures. Similarly, he in-
sists on the spread of the money economy, and the sale and purchase
of land like any other commodity. Of course, he is particularly con-
cerned with the development and spread of productive technology.
Moreover, he sees these processes as quite widespread early. Thus, he
suggests such connections between China and western Asia before 2000
ac. More recent archeological research and publication tends to re-
confirm many of Child’s claims. Some of these are reviewed by Robert
M. Adams in his “Anthropological Perspectives on Ancient Trade” (1974).
Moreover, the (cyclical?) rise and fall of civilizations, and of their
intra- and intercivilizational trade and political relations, was also
already related to “barbarian” and other migrations and invasions before
1500 ec. This date is only 3,500 years ago. Yet, the quote from Wil-
kinson above refers to 5,000 years, Childe to earlier than that, and
Lenski and Lenski (1982) write of 10,000 years back. For all these rea-
sons, and perhaps others, any provisional 1500 2c, or alternative cut-
off date, may be convenient, but wrong. Perhaps we will need to con-
tinue following Fairbank’s second rule and study history, including that
of the world system, much further backward still!
Philip Curtin (1984) recently assembled and reviewed some other-
wise scattered references to “Ancient Trade” in his Cross-Cultural Trade
in World History. He suggests that trade and exchange are as old as hu-
man society. In the Mediterranean, people began moving about by
boat by 7000 ac. Sea trade in the eastern Mediterranean expanded
in about 3000 2c, and fourth-millennium Egypt depended on trade
with the coast of present-day Lebanon. Curtin joins other historians
in supposing that the commonalities among the Indic, Mesopotamian,226 Andre Gunder Frank
and Egyptian civilizations between 2500 and 1500 sc can only be ex-
plained by their contacts with each other. He notes that, according
to some archaeologists, sea and land commercial links may have been,
not only the result, but the very cause of Indus Valley urban civili-
zation. Others suppose greater influence coming across the Indian
Ocean from east to west.
Curtin discusses evidence of long-distance trade over at least 1,500
kilometers and its significance in these areas between the fifth and third
millennia. This trade involved particularly metals, their ores, and other
minerals, These in turn were associated with specialization in skilled
metal-working, which often took place far from the site of the mines
in Anatolia and elsewhere. During the rule of Saragon of Akkad, and
the extension of his empire circa 2300 sc, trade extended at least from
Oman, Bahrain, and Anatolia to India. On the one hand, “Saragon
sent out military expeditions to and maintained fortifications in An-
atolia to protect his commercial interest” On the other hand, “once
again the long-distance traders appear to have been from the less multi-
functional society, carrying their goods to the economic centers of that
time” (Curtin, 1984: 66). In these Mesopotamian centers, recent re-
search has downgraded the supposed dominance of the temple econ-
omy and has shown
that most land actually belonged to private individuals, includ-
ing women. Private people had controlled land from the be-
ginning of Mesopotamian history down through the Third
Dynasty of Ur. State control over the economy becomes im-
portant only in later times. Finally, new investigations showed
that markets and fluctuating prices could be found in Meso-
potamia at least as early as the end of the fourth millennium
a.c. even though redistributive and reciprocal arrangements
may also have existed. . . . [About 2600-2500 s.c.] other evi-
dence suggest{s] an elaborate network of trade, probably car-
ried by professional traders and operating through market
exchange, though some of the merchants may have acted as
agents from temples or for political authorities, who also wanted
to profit from trade (Curtin, 1984: 64-65).
The Kultepe tablets [from Kanesh in Anatolia] portray an eco-
nomic order where merchants were extremely sensitive to small5,000 YEARS OF WORLD SYSTEM HISTORY 227
changes in price, where prices fluctuated rapidly, and where
political authorities interfered in trade as much as they could
to suit their own particular interests—but never to the extent
of successfully suppressing the underlying play of supply and
demand, though here as elsewhere in the history of commerce,
protection costs were an inescapable cost of doing business (Cur-
tin, 1984: 70).
Some Implications. These discoveries suggest several propositions which
are relevant to my general thesis about world system history.
—A\ll the political-economic modes of Marx, of Polanyi, and
of others, and particularly markets and tributary redistribution,
have a much longer history and wider prevalence than previ-
ously thought.
—Long-distance trade began much earlier and was much
more important for social development than was previously,
that is recently, thought possible.
—Both local and long-distance trade and their associated
other manufacturing and agricultural economic activities and
political structures and processes combined to signify much more
capital accumulation, and much more significance of the same,
than we have as yet considered.
These political-economic structures and processes were
much more central to the rise and decline of states, empires,
and civilizations, and to changes in relations among them, than
isolated political-cultural factors, which have received prefer-
ential emphasis by many other analysts.
These and other processes and their analysis suggest much
longer and broader systemic connections in world history than
most suspected, and they justify any attempt to do a world sys-
tem history over several millennia.
—Further research is likely not to confine, but further to
expand, the time depth and geographical and social range of
this historical continuity, and its political-economic connections
in the world system, perhaps to include east and southeast Asia
in one direction and Africa and Europe in the other.228 Andre Gunder Frank
IV. CONCLUSION: HISTORICAL THESES,
CONCEPTUAL PROPOSALS, AND RESEARCH AGENDA
The foregoing introductory and theoretical review, along with my
recent related work with Barry K. Gills, lead me to the following ten-
tative historical theses, conceptual proposals, and research agenda. I
begin with the most concrete historical ones and progress to the more
abstract theoretical ones. However, this conclusion is intended as a
(not the) conceptual guide to my (and I hope others’) writing of a his-
tory of the world system in which European/American hegemony is
only a late arrival and perhaps a temporary phase. The following par-
agraphs incorporate the introduction to, and a summary of, the com-
plementary paper by Gills and Frank (1989).
1. World System Origins
1.1. The origins of our present world system can and should be traced
back at least 5,000 years to the beginning of the relations between Mes-
opotamia and Egypt. The world system developed out of several con-
fluences like a major river system (e.g., the Mississippi-Missouri). One
such major confluence was that of Mesopotamia and Egypt, which
Wilkinson dates around 1500 sc. However, there also were other im-
portant contributory developments reaching back to around 2500 Bc
or earlier. Relations of Mesopotamian Sumer with Anatolia on the
one hand, and with the Indus civilization on the other, also contri-
buted to the formation of this world system. However, the expansion
was not really like a river that flows through a preexisting bed; rather,
the world system expanded into and incorporated other areas like a
glacier, in part adapting to preexisting topology and in part itself re-
structuring this topology. We may also use the analogy of an inkblot
that spreads, or to two or more inkblots, which merge. Thus, even-
tually China and southeast Asia converged into this developing world
system. Then, the “discovery” of the New World, and later Oceania,
were major instances of incorporation.
1.2. The ecological basis of the world system accounts for its origins
and much of its subsequent historical development. The rise of spe-
cialization, and the consequent exchange among urban civilizations
with their agricultural peripheries, and with tribal and nomadic hin-
terland suppliers, of essential raw materials was ecologically determined5,000 YEARS OF WORLD SYSTEM HISTORY 229
and led to the genesis of the world system. Ecological-economic-social
interactions and, especially, constraints have accompanied the histor-
ical development of the world system ever since and lately are
(re)impressing themselves on popular consciousness.
1.3. Economic connections among various parts of the world system be-
gan much earlier, and have been much more prevalent and significant
than is often realized. Economic connections through trade and mi-
gration, as well as through pillage and conquest, have been much more
prevalent and much wider in scope. They have also gone much farther
back through history than is noted in most studies. So have manufac-
turing, transport, commercial, and other service activities, and their
relations with agriculture and livestock. These economic connections
have not received nearly as much attention as they merit per se. Even
more, the far-reaching importance of these trade connections in the
social, political, and cultural life of “societies” and their relations with
each other in the development of the world system as a whole have
been totally neglected. Even those who do study the trade connections,
such as Philip Curtin in his study of trade diasporas (1984), neglect
the systematic study of the world system complex which these trade
connections promote.
The more one looks, the more widespread, integrative, and influ-
ential does this world system political-economy appear. An instance
(reviewed by Lombard) is the “medieval” Islamic-centered single world
economy, which affected countless millions of people spread through-
out northern, western, southern, and eastern Europe; west, north, and
east Africa; and west, south, central, east, and southeast Asia. None-
theless, this does not mean that a particular political-economic pro-
duction is determinant for, or coterminous with, the world economic
system. It also does not imply that solely economic or even political-
economic relations are the “ultimate” determinate, let alone the be all
and end all, of social life or even of the world system. The example
of Islam itself disproves this. We return to this matter below.
Other earlier examples of economic connections are the China-
India-Persia-Rome-Africa trade at the beginning of the Christian era,
and the Asian and African connections, if nor origins, of the previous
Hellenistic world. Already long before classical times, Egypt, the Le-
vant, Anatolia, Mesopotamia, and the Indus Valley engaged in long-
distance trade with each other. This trade was so systematic—or so230 Andre Gunder Frank
systemic—as to have played probably essential roles in their respective
economic, social, political, and cultural formation, and thus interre-
lated their civilizations. Yet, these in turn were dependent on their
political-economic relations with tribal and other “external” peoples.
1.4. World system extension grew to include most of the Asian-African-
European ecumenical (“Eastern” hemisphere) land mass and its out
lying islands by 600 sc. It incorporated much of the “New World” (the
Western hemisphere) by ap 1500. However, there is increasing evidence
of earlier trans-Atlantic and trans-Pacific contacts, even if between two
or more systems rather than within one. Systemic economic and po-
litical connections did develop much earlier among Africa-Egypt-the
Levant-Anatolia-Mesopotamia-south Asia, and it in turn with east and
southeast Asia, and by all of them with inner and central Asia. In east
Asia also, first economic and cultural connections, and then political
connections, developed among Shang, Chou, Korean, and Japanese
peoples in the north and east. Similar connections developed among
Thai and other southeast Asian peoples, and those in the south of what
would become part of China. “China’ itself may be dated perhaps from
Shang times already before 1500 nc, and certainly from 1000 sc in
the Chou dynasties in the north. Systematic north-south connections
developed between then and 500 Bc to form the China we know with
its dynastic ups and downs.
Economic, cultural, and political connections developed and sys-
tematized both through incorporation of “outer” areas and through their
fusion, as between two expanding inkblots. How far “out” did the
world system extend at any particular time? For Wallerstein, large
parts of the world remained “external” to the world-system for long
periods even after 1500. This is because these areas and their peoples
did not satisfy his (I now believe excessively) stringent criteria for
system inclusion—until they came to do so through “incorporation”
into the system. However, long before ap 1500, and even before 1500
Bc, we can identify “hinterland” regions and (often migrant) peoples,
who for long periods maintain trade, migratory/invasory, political,
and socio-cultural contacts with the world system peripheries—and
thereby also with the system core(s). All sedentary “civilized” and
most nomad “barbarian” peoples on the Afro-Eurasian ecumente land
mass have jointly participated in a single—even if multicentric—
world system for several millennia.$4000 YEARS OF WORLD SYSTEM HISTORY 231
2. World System Routes and Nexuses
2.1. Maritime routes furthered economic and other connections among
many parts of the world system and contributed to its expansion in
important ways. The advertising “blurb” of The Sea-Craft of Prekistory,
by Paul Johnstone (1989), reads: “The nautical dimension of prehis-
tory has not received the attention it deserves. .. . Recent research
has shown that man travelled and tracked over greater distances and
at a much earlier date than has previously been thought possible. Some
of these facts can be explained by man’s mastery of water transport
from earliest times.” The sea routes were cheaper and often favored
over the overland ones between east, south, and west Asia, as well as
to and from east Africa and the southern and northern shores of the
Mediterranean. These and other sea routes mentioned below also
formed part of the network of “Silk Roads”
2.2. The Silk Roads, over both land and maritime routes, formed
a sort of spinal column and rib cage of the body of this world system
for over 2,000 years. These “roads” extended overland between China,
through inner and central Asia, to the above-mentioned “Middle East-
em” (west Asian) nexuses. From there, they extended through the Med-
iterranean into Africa and Europe. However, this overland complex
was also connected by Mediterranean, Black Sea, Red Sea, Persian
Gulf, and numerous river maritime “roads.” Moreover, this overland
Silk Road complex was complemented by a vast maritime network cen-
tered on the Indian Ocean through the Arabian Sea and Bay of Ben-
gal, and on the South China Sea. These in turn were connected overland
across the Kra narrows of the Malay Peninsula, as well as by ship
through the straits between it and Sumatra, etc.
The Silk Roads, of course, derive their name from China's prin-
cipal export product to the West. However, the trade extended far be-
yond silk alone. Indeed, the silk had to be paid for, and complemented
by, a large variety of other staple and luxury goods, as well as money
and services, including the services of slaves and other people who per-
formed them. Thus, the Silk Road also served as the trade routes, the
loci of urban and administrative centers, and the military, political,
and cultural sinews of a vast and complex division of labor and cul-
tural diffusion.
2.3. Central Asia has been a much-neglected focal point of world232 Andre Gunder Prank
system history, both as a logistic nexus among its regions to the east,
south, and west, and through the recurrent pulse of its own waves of
migration and invasion into these regions. Pursuing the interactions
and the cycles among them along the central Asian trails of Owen Lat-
timore, Aurel Stein, and others like Marco Polo before them would
help fill the hole left by McNeill and the omission of these peoples by
Wilkinson. Was central Asia—and not primarily the “civilized” areas—
something of a motor force of change in the whole system? Whatever
the causes and explanations of the central Asian cycles, it seems es-
tablished that movements of invasion and migration, each lasting some
200 years, occurred in roughly half-millennium intervals. There were
waves of invasions in 1700-1500 sc, 1200-1000 sc, around 500 8c,
around the year 0, ap 400-600, and ap 1000-1200/1300. Each inner
wave created outer waves, except the last one of Genghis Khan and
his successors up to Tamerlane, who overran everything themselves.
2.4, Logistical nexuses in what is now called the Middle East, in inner
Asia, and through some sea straits acquired critical roles as bottleneck
control-points along these routes. Therefore, these regions and corri-
dors or nexuses also played an especially significant role in the devel-
opment of the world system. They were magnets which attracted the
flow of migrants and invaders and the attention of hegemonic powers.
However, they were also magnets because of their market demand and
their supply of precious metals and other raw materials. Therefore,
they have also been especially significant fulcra of religious and other
cultural diffusion.
‘The Afro-Eurasian (more properly Asian-Afroeuropean) ecumene
has traditionally had a few outstanding nexuses, which have repeatedly
attracted such expansion and control, and have served to cement the
world system. We can identify three obvious trade routes/corridors con-
necting the Mediterranean and its European and African outlyers to
the west, central, south, and east Asia: (1) The Nile and Red Sea route
(with canal or overland connections between them) to and from the
Indian Ocean and beyond; (2) The Tigris and Euphrates-Persian Gulf
route to and from the Indian Ocean and beyond (whose topology and
climate gave it competitive advantages over the rival first route); and
(3) the Dardanelles-Bosphorus-Black Sea and then overland “Silk Road”
route to Asia.
Of course, each of these three main routes had competing alter-5,000 YEARS OF WORLD SYSTEM HISTORY 233
native variants and feeder routes of its own. For instance, there were
several Silk Roads between east and west. There were also different
feeder routes in east and central Asia, and to/from south Asia, as well
as routes connecting northern and western Europe and through the
Baltic Sea via Dnieper, Don, Volga, and other Ukrainian and Russian
routes. Of course, within and between Europe, Africa, south Asia, cen-
tral Asia, and east Asia, there were additional alternative competing
routes and bottlenecks of trade, migration, invasion, and cultural diffu-
sion. Topological factors favored some inland places, like Samarkand,
over others in relations within and among east, central, and south Asia.
Ceylon was an important locus of east-west transshipment, as was Kra
on the Malay Peninsula, the Straits of Malacca, and others.
Significantly, these nexuses repeatedly were the objects, and some-
times the subjects, of especially important economic, political, military,
migratory, and cultural processes—and of competition for the access
to and benefit from the process of accumulation, which their privileged
position afforded. It is no (historical) accident that the most privileged
of these nexuses to and from the Persian Gulf has been a major mag-
net of military attention (still for U.S. and other fleets in 1988). Nor
can it be a historical accident that most of the past and present major
world religions were born in and around this region and either died
out (like Zoroasterism) or spread around the world, like Judaism, Chris-
tianity, Islam, Buddhism, and Hinduism. Eurocentrism has posed an
impediment to the identification and study of these other important
nexuses in the world system—and for that reason, among others, to
the appreciation of the world system as a whole.
3. Infrastructural Investment, Technology, and Ecology
3.1. Infrastructural investment and capital accumulation accompanied and
supported most parts of the world system from its beginning and
throughout its historical development. Capital accumulation can be
shown to have been a constant process (albeit an uneven or cyclical
one) for millennia. Capital accumulation went into the construction
of, and (re)investment in, massive and extensive productive facilities
and public works in the following areas: agriculture and manufactur-
ing; the development of facilities for transport and trade; defensive
and offensive military measures; the provision and use of financial in-
stitutions; and yes, the construction of ideologically legitimizing mon-234 Andre Gunder Frank
uments. All of these represented accumulation, then and now.
Moreover, many historical-economic, political, and cultural/
ideological events, processes, and institutions can be clarified and “fall
into place” if they are seen as instrumental to promoting the process
of capital accumulation and sharing in, or disputing, its benefits. Not
only trade and migration, but also much political power and rivalry,
as well as ideological legitimization and cultural diffusion, are assoc-
iated with “economic interest” in the generation and extraction of eco-
nomic surplus and its investment in capital accumulation. On the other
hand, to what extent did the associated cost of accumulation through
military (over)extension, la Paul Kennedy (1987), contribute to, or
even generate, the decline of hegemonic powers and the exhaustion
of nomad invaders? Moreover, this process of economic surplus man-
agement and capital accumulation is so interregional and inter-‘societal”
as probably to have constituted a process of world accumulation in
a world system over the millennia.
3.2. Technological innovation also played a similar and related role
throughout the historical development of the world system and me-
diated in the competitive economic and military conflicts among its
parts. Technological change was much more pervasive and important
than is often acknowledged. Technological process in military and civ-
ilian techniques of production, organization, and trade has long played
an important, and oft-neglected, role in the history of the world sys-
tem and in the changing relations among its parts. Technological ad-
vance and advantage has been crucial throughout history in offensive
and defensive armaments, shipping, and other transportation equip-
ment and facilities, as well as in construction, agricultural, metalwork-
ing, and other manufacturing methods and facilities. Progress, and
leads and lags in all of the above, have significantly contributed to the
creation of, if they might not be said to have caused the relations of
inequality within, the world system, including core-peripheral relations
and the existence of hegemonic powers.
3.3. Ecology, however, has always exercised an essential influence
and constraint on this world system development. Indeed, much tech-
nological development itself has been in response to ecological
conditions — and has in turn had ill effects on the ecology. So technol-
ogy again responds to ecological constraints.5,000 YEARS OF WORLD SYSTEM HISTORY 235
4. Surplus Transfer and Accumulation Relations
4.1. Surplus transfer and interpenetrating accumulation among parts of the
world system are its essential defining characteristics. Capital accu-
mulation, and the associated process of the generation, exploitation,
transfer, and investment (or, in short, the management) of economic
surplus, has been woefully neglected in historical and social-scientific
studies. To talk of transfer of surplus from one “society” to another is
to say that neither would be as it was and is without its relations with
the other. The capture by elite A (with or without its redistribution)
of part of the economic surplus extracted by elite B connects not only
the two elites, but also the economic, social, political, and ideological
organization of their “societies.” Thus, this transfer, or sharing, of sur-
plus between “societies” structurally and systematically interrelates their
institutions, processes of surplus management, and their class and gender
structures of exploitation and oppression. Of course, if part of the sur-
plus of elite A is also traded, through equal, or more usually unequal,
exchange, for part of the surplus from elite B, the systemic connections
are even more systematic. Similarly, if elite B has a relation of surplus
trade or extortion with elite C, as it has with elite A, then A and C
are also systematically connected in the same system.
The interconnection and interpenetration of economic surplus and
accumulation beyond and between polities can thus become an (the?)
extended and/or concretized criterion of inclusion in, and bounding
of, the world system. If different “societies,” empires, and civilizations,
as well as other peoples, exchanged surplus, then they participated in
the same world system, since the organization of “society” A could and
would not be the same as it was in the absence of its contact with B,
and vice versa. An example of this was the connections between the
imperial elites and the productive/exploitative structures of the Roman-
Parthian-Indian-Han Chinese empires. Therefore, in a socio-political
economic sense, trade in high-value luxury items, not to mention pre-
cious metals, for elite consumption and accumulation, may be even
more important than low-value staple trade for the expansion and re-
production of the world system as a whole.
4.2. Conter-Peribhery- Hinterland (CPH) complexes and hierarchies among
different peoples, regions, and classes have always been an important