[ Pobierz całość w formacie PDF ]

methodology.
Chapters 3 through 6 expand one aspect of the argument with illus-
trations from historical examples. Chapter 3 looks at the weaknesses of
recently independent states and their impact on lengthy civil wars. I
describe the international processes that led to the dismantling of the
colonial empires and the creation of large numbers of weak states, as
well as the ways in which state weakness affects the duration of wars.
Chapter 4 deals with the counterargument that lengthy civil wars are
due to hatreds bred of ethnic conflict. I criticize the notion that  eth-
nic or  identity-based civil wars are different from other types of war.
Instead, I argue that identity-based civil wars are merely one variant of
weak-state civil war, and are based in the same general processes as
other civil wars of this era.
Chapter 5 examines the effect that the Cold War has had on pro-
longing civil wars. Rather than taking the communist/Western aspect
of civil wars as a given, I argue that international actors constructed and
imposed Cold War dimensions on many civil wars. Once a civil war
32 Neverending Wars
was defined as communist by the superpowers, Cold War frameworks
provided material and cognitive resources that increased the duration
of these wars.
Chapter 6 discusses the impact that interstate interventions have had
on the lengthening of civil wars. One major source of intervention
has been the Cold War conflict, which encouraged partisan interven-
tion in Third World conflicts. Additionally, international interventions
are commonly perpetrated by states other than the superpowers, for
reasons ranging from support of weak-state sovereignty to carrying out
interstate aggression to policing civil unrest in the intervening country.
These interventions, which often take place on both sides of the civil
war, have contributed to the substantial increase in the duration of civil
wars after 1945.
In Chapter 7 I reflect on the importance of recognizing the role
played by the international system in lengthening the civil wars in this
period, and speculate on ways in which the international community
might constrain civil wars in weak states in the future.
2
World Patterns in
Civil War Duration
the statistical analyses in this chapter test the argu-
ments I have advanced thus far specifically, the impact of a changing
international ecology of states on civil war activity over the last two
centuries. I also examine more targeted hypotheses regarding weak
states, ethnic conflict, the influences of the Cold War, and the conse-
quences of interstate intervention. To foreshadow the outcomes briefly,
the statistical findings do strongly support the general argument that
the international system has played a critical role in lengthening the
civil wars in the late twentieth century.
The claims developed earlier are very broad in scope, addressing a
worldwide shift in civil war activity over the last century. Any given his-
torical case (or handful of cases) provides an insufficient basis for draw-
ing conclusions regarding global trends of this scope. While statistical
models do not capture the historical and political complexity of indi-
vidual civil wars, they are able to provide valuable information on such
broad empirical trends. A statistical approach has the advantage of
highlighting general trends without being overly swayed by a few ex-
ceptions. Subsequent chapters look to specific cases for greater detail
and give a more fine-grained treatment of arguments.
This study requires clear conceptualization and measurement of civil
wars that is not biased across region or time. The general public re-
ceives more news and scholarship about some regions of the world than
33
34 Neverending Wars
others. For instance, the news coverage of the conflict in Northern Ire-
land has been far greater than that of the civil war in the Sudan. Thus it
is a shock to realize that the casualty figures for the thirty years of
 troubles in Northern Ireland are approximately 3,000, while the ca-
sualties for the twenty-year old civil war in the Sudan run in the mil-
lions (Brogan 1998; Sivard 1996). Moreover, there is a political aspect
to the labeling of a conflict as a civil war, rather than a  police action
or  control of foreign elements. Since the results of the analyses can
only be as reliable as the data that went into them, care must be taken
to ensure that the definition and measurement of civil wars is consis-
tent and unbiased.
The data on civil wars that is used in this book comes from the
Correlates of War dataset (Small and Singer 1982; updated in Sarkees
2000). The chief advantage of this dataset is that it uses explicit and
objective criteria for defining wars, in contrast to other sources of civil
war data that disproportionately emphasize certain regions or certain
time periods (for instance, see Luard 1972; Richardson 1960;
Clodfelter 1992). The Correlates of War define civil wars according to
three criteria. First, the war must be between the state and a societal
opposition group, rather than among societal groups. Secondly, both
groups of combatants must actively participate in the civil war. This
criterion distinguishes between a war and a massacre in which one
group is slaughtered without significant resistance. Third, the yearly
toll of casualties must be greater than a thousand battle deaths. This
criterion limits the list to large-scale civil wars. The battle casualty cri-
terion also provides dates on the start and end of civil wars, which is
used to calculate the duration of the war. While a Correlates of War
civil war may be preceded or followed by years of lower-intensity con-
flict, the battle death criterion includes only years of high-intensity
conflict. By using these criteria I do not mean to imply that lower-in-
tensity conflicts are unimportant or that they cannot have significant
political or social consequences as the examples of Northern Ireland
and South Africa demonstrate. But the objective criteria of the Corre-
lates of War dataset provide comparability across countries and histori-
cal periods that is critical to this study.
The criterion of 1,000 war-related casualties per year is not as arbi-
trary as it might seem at first. Most civil wars have well over a thousand
casualties, and most events we would intuitively consider riots or assas- [ Pobierz całość w formacie PDF ]