Citation
Cultural repression and secessionist conflict

Material Information

Title:
Cultural repression and secessionist conflict utilizing relative deprivation theory to explain anti-regime violence
Creator:
Macrina, Joseph Norman
Publication Date:
Language:
English
Physical Description:
x, 104 leaves : ; 28 cm

Thesis/Dissertation Information

Degree:
Master's ( Master of Arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Sociology, CU Denver
Degree Disciplines:
Sociology

Subjects

Subjects / Keywords:
Deprivation (Psychology) ( lcsh )
Intergroup relations ( lcsh )
Separatist movements ( lcsh )
Political violence ( lcsh )
Culture ( lcsh )
Culture ( fast )
Deprivation (Psychology) ( fast )
Intergroup relations ( fast )
Political violence ( fast )
Separatist movements ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 98-104).
General Note:
Department of Sociology
Statement of Responsibility:
by Joseph Norman Macrina.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
57662245 ( OCLC )
ocm57662245
Classification:
LD1190.L66 2004m M32 ( lcc )

Full Text
CULTURAL REPRESSION AND SECESSIONIST CONFLICT:
UTILIZING RELATIVE DEPRIVATION THEORY TO
EXPLAIN ANTI-REGIME VIOLENCE
by
Joseph Norman Macrina
B.A., University of Colorado at Denver, 2002
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Masters of Arts
Sociology
2004


This thesis for Master of Arts
degree by
Joseph Norman Macrina
has been approved
by


Macrina, Joseph Norman (M. A., Sociology)
Cultural Repression and Secessionist Conflict: Utilizing Relative Deprivation Theory
to Explain Anti-Regime Violence
Thesis directed by Assistant Professor Yili Xu
ABSTRACT
The ongoing use of income and other economic indicators to explain
collective frustrations that lead to the onset and rise of secessionist conflicts and other
forms of political violence has continued to produce unreliable and erratic empirical
results. The empirical inconclusiveness of Gurrs income and economic based
Relative Deprivation Theory (RDT) is overcome in this study by examining, not
merely economic indicators of frustration, but rather the impact cultural deprivation
has on violent secessionist movements. Collective cultural discontent as a product of
the repression of a groups cultural capital is argued to be a significant determinate of
violent political rebellion. Past studies on the subject of political violence have
simply failed to take into account the role that the repression of a groups cultural
existence plays in the dynamics of secessionist conflict. Propositions about violent
political rebellion as being the result of frustrations arising from cultural repression
are examined in this study by using multiple regression analysis. Additionally, this
study is exceedingly unique whereby a cultural capital index was created from
variables gleamed from the cross-national Minorities at Risk data set, and consisted
m


of such cultural indicators as language, religion, customs, culturally unique dress, and
loss of cultural autonomy. This index, along with other cultural repression variables
such as restrictions on movement, and forced resettlement, indicate that, when
controlling for general racial, economic, and political variables, there exists a
statically positive relationship between cultural repression and violent political
rebellions.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed
IV


DEDICATION
I dedicate this thesis to my mother Darlene, my sisters Christy and Melissa, and
Lindsey Sabec for their unwavering support and steadfast encouragement. These four
women have been a loving source of understanding and patience to which I am
forever indebted.


ACKNOWLEDGEMENT
My special thanks to both Dr. Candan Duran-Aydintug and Dr. Yili Xu for their
guidance, patients, and insightful views. Without their amazing sacrifices and long
hours working with me, this work would not have been possible. Their support and
dedication has served has a source of inspiration and motivation for me for the past
several years.


CONTENTS
Figures....................................................ix
Tables.....................................................x
CHAPTER
1. INTRODUCTION............................................1
2. REVIEW OF LITERATURE AND THEORY........................12
Rational Choice Theory..............................13
Game Theory.........................................25
State Power Theory..................................35
Resource Mobilization Theory........................46
Relative Deprivation Theory.........................51
3. METHODOLOGY............................................63
Data................................................64
Sample..............................................65
Variables and Indicators............................68
Method of Analysis..................................74
Limitations.........................................75
Other Limitations.............................78
4. DATA ANALYSIS..........................................80
5. CONCLUSION.............................................84
vii


APPENDIX
A. VARIABLE DEFINITIONS AND CODING................90
B. DESCRIPTIVE STATISTICS FOR VARIABLES...........95
BIBLIOGRAPHY...........................................98
viii


FIGURES
Figure
2.1 Structure of a Simple Game Tree, or Extensive Form Model.........28
2.2 Normal Form Model................................................32
2.3 Movements of Crisis and Affluence................................51
IX


TABLES
Table
4.1 Regression Analysis for All Independent Variables.................81
4.2 R-squared Model...................................................82
x


CHAPTER 1
INTRODUCTION
Violent political conflict has tended to exist as an unfortunate byproduct of
human interaction for millennia. Since the Napoleonic Wars, civilization has fought
an average of six international wars and six civil wars per decade (Waller, 2002).
Additionally, since the birth of the 19th century, civilization experiences, at any given
time, three high-fatality political conflicts (Waller, 2002). However, it was believed
that the end of World War II would bring about a new era in world history where
civilization would enter into newfound peace and security (Wallerstein, 1974)- this of
course was not the case. The five decades following the end of WWII saw 150 wars
and only 26 days of world peace (Eckhardt, 1992). During this time, it is estimated
that well over 100 million people have been brutally slain at the hands of fellow
human beings (Eckhardt, 1992). That is more than five times the number from the
nineteenth century and more than ten times the number from the eighteenth century
(Westing, 1982).
The importance of understanding violent secessionist conflict is self-evident
and should be in doubt to no one. The world is becoming increasingly interconnected
and as a result, the worlds culturally distinct groups, once separated from one
another by geography, are being thrust together by advances in technology,
transportation, and population dynamics. In the last decade in particular, the world
1


has become more economically, politically, and culturally interdependent.
Transnational and international organizations have begun to play significantly larger
roles in the cultural dynamics of all nations. At the same time, nations split, dissolve,
and multiply (Williams, 1994), arms procurement among weak and developing
nations increase rapidly to extraordinary levels (MWEAT, 1999), and political
instability has become endemic throughout much of the developing world (Williams,
1994). When these elements are mixed into a countrys sociopolitical climate that is
wrought with intolerance, oppression, and tyranny, cultural clashes tend to produce
bloody internal conflicts and even civil wars.
To begin to understand how political violence operates, one first needs to be
cognizant of just how culturally diverse the world truly is. First, Grimes (1988)
reports that the worldwide prevalence of culturally distinct groups and ethnic
diversity is demonstrated by the existence of about 6,000 different languages, and
somewhere between 900 and 1600 culturally distinct groups (Levinson, 1993).
Williams (1994) explicates this point by showing that in the former Soviet Union
alone, there resides within its border some 130 officially recognized nationalities.
Secondly, Williams (1994) gives the example of the 183 countries that are
members of the United Nations (as of June 1993). Of these nations, only a few are
homogeneous, where the vast majority of these U.N. countries are in fact very
multiethnic. The third, and more problematic illustration of the worlds multiethnic
character, demonstrates that in a world sacked by conflagration and discord,
2


culturally related conflicts account for the vast majority of violence and intra-national
war (Brogan 1990). As such, Williams (1994) goes on to assert one half of the
worlds [nations] have experienced significant ethnic conflict since WWII [and that]
about 80% of deaths in warfare during that period have been internal to national
states, and much of that total has come from ethnic violence (p. 50). Although
studies that have attempted to report estimates for total deaths attributable to ethnic
violence since 1945 tend to vary significantly in their reported numbers, conservative
estimates are usually placed anywhere betweenl 1 to 20 million (Williams, 1994;
Nagel, 1980; Brogan, 1990 Russett & Starr, 1989).
Such saddening events as the bloody dissolution of Yugoslavia and the
ethnic cleaning policies against Muslim and Croat civilians, periodically violent
ethnopolitical campaigns of secessionist terrorism and state sponsored repression in
Chechnya, and the civil war bloodbath of Rwanda in which Hutu interhamwe
slaughtered hundreds of thousands of Tutsi civilians in just a matter of months have
brought about new academic attention to the problem of political violence. These,
and other such disheartening events of just this past decade, have worked to reshape
and renew public and scholarly awareness to this lamentable social phenomenon.
Brubaker & Laitin (1998), note that this renewed academic attention to cultural and
nationalist violence is profoundly important to deal with this new world disorder (p.
424).
3


Past research into the field of violent political rebellions has often been
dominated by theories focusing on macro-societal processes (Williams, 1994) such as
split labor markets (Bonacich, 1976), economic competition (Banton 1983; Olzark,
1992), and income inequities (Gurr, 1968). Additionally, other types of theories have
taken shape within the relatively short period of just the last few decades. Such
structural approaches as explicated by state power theory and even by theories on
modernity, international influences, and historical colonialism (Hachter, 1975), have
all attempted to explain political violence as being the product of large morphological
phenomena that seem to be divorced from group-level processes. Thus, most critics
of these theories tend to view them as being to macro in concept and are removed
from the actual group-level mechanisms that truly influence the decision making
process. Jenkins & Schock (1992) give an example of such structuralistic theories:
Dependency, world-systems and global hegemony theories of
domestic conflict have paid special attention to weak authorit-
arian or elitist states in the world-system periphery; the prima-
ry mechanisms posited are class polarization and capitalistic
intrusion upon traditional societies, (p. 55)
However, in an attempt to reconcile overly structuralistic macro-level theories
with individual based micro-level processes, the academic community has recently
produced what might be considered antithetical approaches to the problem of
secessionist violence (Williams, 1994). Some of these included sociobiological
perspectives (van den Berghe 1991), and primordial solidarity theories (Stack 1986).
Nevertheless, most of the conventional studies produced today tend to rotate around a
4


select assemblage of theoretical paradigms. Most of these theories have a propensity
to be unidimensional in their scope, and primarily only concentrate on one or two
causal factors of violent political rebellions. However, this study constitutes a move
away from such unilateral approaches and toward a more complex analysis of cultural
capital as a whole. Unlike previous studies that attempted to test relative deprivation
hypotheses using only one or two indicators of frustration such as religion (Fox,
2002) , land (Midlarsky, 1988), international support (Saideman, 1997), or income
equality and political opportunities (Schock, 1996; Canache, 1996), this study
examines a complex amalgamation of general cultural indicators and political and
economic factors.
Before I continue, I believe that a clarification of terms is in order. Terms
such as political violence, secessionist, rebellion, and revolution each have
specific meanings, and therefore it is important that one fully understands what is
meant when these terms are being used. Throughout this study, the words violence
and political rebellion will be used interchangeably (if the term political rebellion is
not used in conjunction with the word violent, one can assume that the former is in
fact a violent act unless determined otherwise). This is because that which is being
tested, the dependent variable- anti-regime conflict index- is an explicit measurement
of violent forms of political rebellion (see Appendix A). According to the rebellion
index, only the most serious manifestations of rebellion is coded (Davenport et al.,
2003) . Consequently, even a low score of 1 on the rebellion index scale (political
5


banditry and/or sporadic terrorism) indicates a significant degree of violence.
However, this is not explicitly a study on ethnic violence. The reason for this is
because the cultural capital index that was created for this study does not include
ethnicity as one of its indicators. Moreover, ethnicity, although part of ones culture,
is a complex and intricate phenomenon that typically necessitates its own research
(Lake & Rothchild, 1996; Tishkov, 1995; Horowitz, 1971).
The violent distinction is of incredible importance to this study because I
explicitly attempt to measure violent forms of rebellion, not nonviolent forms.
Nonviolent forms of political rebellion do exist; mainly where the political structure
of a country is such that it allows for groups to peaceably voice their grievances
without fear of governmental reprisals, but they are not the focus of this study. Such
forms of nonviolent political rebellion are peaceful protests, sit-ins, boycotts, and
other such forms of demonstration.
Political rebellion and secessionist conflict are phrases that are used
interchangeably throughout this study. Both connote a meaning in which a group
attempts, through violent means, where peaceful avenues of social change are absent,
to secure for itself political, cultural, and economic independence from the nation in
which it currently exists. One must keep in mind that the word revolution can also be
used, which indicates a substantial precipitation of violence into organized internal
war to bring about significant social change (Russell, 1974). The scaling of the
dependent variable additionally justifies the use of this word because the highest
6


score on the scale, 7, indicates protracted civil war fought by rebel military units
with base areas (Davenport et al., 2002).
The use of the word political is additionally important because one must also
keep in mind that this is inherently a political process, whereby political borders are
attempted, through forceful means, to be redefined by seceding groups. While I do
not focus on political causes of violent rebellion per se, and although the causes of
political rebellion might not always be political in nature, as this study is attempting
to demonstrate, the result is inherently political- the creation or attempted creation of
a new, autonomous and independent region or state. The collective desire to achieve
culturally distinct independence from a repressive regime is an enterprise whose
desired logical consequence is the obtaining of some form of politically recognized
entity- an independent region or state.
This should not be confused with irredentism, where an international
boundary divides members of a single, culturally distinct group, or when a country
incorporates a region, thus splitting cultural groups into separate countries. An
example of this was when Germany was divided into two separate nations following
the end of WWII. If annexation is likely to occur by a neighboring state that is
viewed less desirable than membership in the original state, then that group might
otherwise be less inclined to separatism (Horowitz, 1981). Separatism being defined
as a groups attempts at securing a separate, yet independent region within an existing
state. Secessiortism, on the other hand, is a groups attempt at securing a completely
7


autonomous and politically sovereign nation. There have been many examples of
secessionist and separatist groups such as the Palestinians, the Kurds, the Tamils, the
Chechens, and the Basques. These groups are often a blend between separatists and
secessionists organizations because rebel forces often do not always maintain a
coherent and clearly defined objective. Within one particualar miliatant organization
will often exist a myriad of goals, whereas some will demand complete national
sovereignty, while others merely wish for regional autonomy. Complicating the issue
is that irredentism often comes into play because some groups like the Kurds, the
Basques, the Tamils, the Hungarians, the Bakongo, the Balochi, and others, are
culturally distinct groups that are split among two or more states.
Although separatism and secessionism are nuanced terms that maintain
definitions which are slightly different from one another, they can still be used
interchangeably because the dependent variable in this study does not require a
distinction between the two. Whether it is violence used to bring about separatism or
secessionism, violent conflict still occurs in both instances, and in both instances the
outcomes are still intrinsically political in nature. Moreover, Horowitz (1981) claims
that the use of secessionism, separatism, and secessionist can be used by researchers
of political violence interchangeably due to the similar nature of their definitions.
Before I proceed, I would like to briefly give a general overview of what
relative deprivation theory (RDT) is and how it compares to other theories. In 1968,
then an assistant professor of political science at Princeton University, Ted Robert
8


Gurr, proposed what would become the most wildly analyzed and thoroughly
critiqued theory on secessionist conflict in modem history. Gurrs RDT presented an
understanding of political violence as being the manifestation of collective
psychological frustrations stemming from a groups perceived gap between what they
have and what they think they should have. However, since Gurr presented his
theory, he and other Relative Deprivation theorist have become the target of
considerable criticisms as to the actual predictive value of this theory. Most
opponents argue that the economic and financial nature of RDT creates inadequate
indicators of collective discontent.
This study, however, is an attempt to overcome these criticisms by looking,
not merely at attributes of economic discrimination as being facilitators of collective
aggression, but at the repression and deprivation of a groups cultural existence. This
is an extremely novel approach to this subject in that most studies have failed to
analyze what effects the repression of a groups culture has in facilitating secessionist
violence.1 Most studies have focused mainly on political or economic forms of
repression, and thus have neglected to take into account how cultural repression, in
conjunction with political and economic factors work to advance collective rebellion.
Even though Gurrs RDT does have some significant shortcomings, that I
indicated before, it is one of the few theories today that actually takes into account
collective psychological inclinations. Violence and conflict are inherently
1 For a general overview of past research in the field of political violence see Burbaker & Laitin, 1998.
9


phenomena that are the manifestations of psychologically produced emotions and
frustrations. This is a fact that is unremovable from our psychical constitution. In
fact, Websters Dictionary defines violence as being a vehement feeling or
expression; an instance of such action or feeling (2004). Although there is some
debate over the degree to which frustration actually leads to aggression (a point that I
will revisit later), it is difficult to logically assert that such things as terrorism and
violent banditry are somehow not functions of fervent emotional mental states.
When violence, anger, animus, and aggression reach a point were people
begin to kill one another, it is all but impossible to assume that these actions are not
the result of some form of emotional sentiment. Although past studies utilizing RDT
have tended to misrepresent what in fact causes these affective behaviors, this theory
nevertheless is the only one that rightfully takes into account the manifestations of
violent behavior as being caused by collective psychological dispositions. Collective
violence is a pure expression of affective emotional states, and as such, collective
psychological emotions, such as anger, rage, and even hatred are necessary conditions
for there to be collective violent behaviors.
People do not kill one another because they are victims of some abstract
structural phenomenon. They kill one another because they posses affective laden
cognitions. Virtually all other theories on this subject have failed to fully
acknowledge the psychological aspect of political violence. Other theories such as
resource mobilization and rational choice are fairly good at accounting for and
10


creating predictive models about nonviolent movements, but violence is something
different altogether. This is a profound distinction that past non-relative deprivation
theorists have glossed over.
11


CHAPTER 2
REVIEW OF LITERATURE AND THEORY
The social sciences have long sought explanations to what causes groups of
people to rebel and to seek out new areas of cultural and political autonomy. The
conflicts that arise from these deadly movements are often a complex amalgamation
of state stability, class inequality, resource mobilization, international
interventionism, and cultural differentialization (Williams, 1994). Therefore, this
chapter will be devoted to examining and critiquing the conceptual underpinnings of
the most preeminent theories and studies as they relate to violent political
movements. One must keep in mind, however, that although these theories are
relatively dissimilar in their basic theoretical postulations, they are not wholly
mutually exclusive from one another. There does exist some degree of conceptual
overlap from one theory to the next; thus one can always, through hyper-analysis,
extrapolate similarities between any given theories. Nevertheless, this is not the
purpose of this study. Generally speaking, these theories, and the hypotheses derived
from them, are categorically distinct from one another, and that is what this chapter is
designed to demonstrate.
12


Rational Choice Theory
Rational Choice Theory (RCT) presupposes that actions taken by a particular
individual or group are fundamentally the result of a careful analysis between costs
and benefits. Consequently, if the costs associated with possible action exceeds the
benefits then action is less likely to take place. Conversely, if the foreseeable benefits
produced from a given action exceed its cost, then action is more likely to occur.
However, because there are so many different social, political, cultural, and
environmental variances that exist within any given nation, all of which working to
create a complex matrix of cost-benefit calculations for particular groups, nations
tend to differ greatly in their rates of domestic political violence (Muller & Weede,
1990). Even though the intrinsic and rational interest of all sovereign nations is the
maintenance of governmental and political stability, violent secessionist conflict
and/or civil war still tend to threaten the validity and societal balance of many
regimes (Muller & Weede, 1990).
The process of secessionist conflict is not free from outside influences or
extenuating circumstances. Consequently, according to RCT, the decision making
process, as it pertains to secessionist movements, is one in which rational actors will
take into account the costs, or negative consequences, that might go along with
rebellious activities. This theory, therefore, presupposes that decisions to rebel are
not made capriciously or myopically, but are rather the result of careful deliberation
and examination of ones options. If the costs of rebellion outweigh the benefits, or if
13


there is peaceful means through which to secure positive political change, then
rational actors will decide that violent rebellion is ineffective, and it therefore will be
avoided.
In 1973 Hibbs conducted a study in which he attempted to measure a wide
array of deprivation hypotheses. What he found was that it was difficult to find
significant evidence to support the concept that collective frustrations, caused by
relative deprivations, lead to political violence. His findings, however, would also
prove to be incongruent with how rational choice theories conceptualize the notion
that the immediacy of costs plays an important role in the decision making process.
Hibbs model showed that what was significantly responsible for political violence
were variables that took into account current levels of negative governmental
sanctions to suppress dissent. Such things as previous violence and governmental
acts of coercion where shown to have a significant impact on a groups immediate
decision to engage in rebellious political behavior.
This is problematic for RCT because this study indicates that when actions
taken by a government to suppress dissent are in the short term, levels of political
violence where actually exacerbated (Muller & Weede, 1990). Thus, even in the face
of experiencing extremely high costs in the short term, Hibbs study indicates that
groups were still likely to engage in political violence. Moreover, the lagged effects
of the current negative sanctions indicator is also inconsistent with the RCT
perceptive because the presence of negative sanctions actually contribute to high
14


levels of political violence (Muller & Weede, 1990). According to RCT, what one
should expect to find is that governmental coercion and political violence yield an
inverted U-curve relationship between one another. What this means is that,
according to the rational choice argument, when there exists either very high or very
low rates of governmental sanctioning of dissenting groups, then the likelihood for
political violence will be low. As such, political violence is mostly likely to occur
when rates of governmental sanctions are at intermediate levels (Muller & Weede,
1990).
This inverted U hypothesis squares with the rational choice paradigm whereby
rational actors will be less apt to participate in violent political rebellion when the
threat of governmental reprisals is very high. In this instance the regime can be
characterized as being totalitarian and relatively powerful. Examples of this sort
would be Stalins Soviet Union, or Saddam Husseins Iraq. In either of these cases,
an attempt at collective political change would have been seen as suicide in which the
government would have been capable of total and complete elimination of the threat.
Conversely, when the political landscape of a particular country is such that the
government exhibits relatively low levels of governmental sanctioning of dissenting
groups, then there usually is no reasonable need for political violence. In these
situations, there typically exists peaceful and more democratic means by which to
petition the government for grievances.
15


Gurr had originally formulated the inverted U hypothesis but according to
some (Hibbs, 1973; Snyder, 1976) the indirect nature of the indicators used
constitutes an ineffective testing method. Gurr, himself the creator of relative
deprivation theory, had originally hypothesized that political violence might be the
product of rational action by rational actors. However, Gurr never came out and
proposed a rational choice hypothesis, rather he disguised the rational component of
this original theory under the term utilitarian justifications (Muller & Weede,
1990). Nevertheless, he did eventually begin to move away from its rational choice
underpinnings and RDT became more of a theory that explained political violence in
socioeconomic terms.
In studies that have initially focused on testing RCT via the inverted U
hypothesis, there has been some support found between the repressiveness of a
political regime and the rate of political violence (Muller & Seligson, 1987). Where,
as stated before, the likelihood of collective political rebellion is contingent on the
degree of governmental sanctioning that exists within a particular state. If extreme
repressiveness makes collective mobilization all but impossible then rebellion will be
less likely to occur. In this instance, the omni-repressiveness and the omnipotence of
the state become an invaluable deterrent in preventing mass-rebellions. However, if
the political structure is semi-repressive, where peaceful means of social change are
fundamentally ineffective, and where collective mobilization is not completely
blocked by the regime, then political rebellions are possible.
16


The nature of this curvilinear relationship between political rebellion and the
degree of state repressiveness is exhibited in the fact that states working to keep
rebellions at a minimum and thus maintain political stability, should always be
cognizant of the consistency of their approach (Muller & Weede, 1990). Lichbach
(1987) reminds us that governments wanting to avoid mass-political rebellions should
always maintain policies that are all or nothing- either completely repressive or not at
all. Regimes that are somewhere in the middle almost always become victims of
mass-rebellions and even revolution. About this Lichbach contends constant
government accommodative and repressive policies reduce dissent while inconsistent
polices increase it (1987, p. 287).
One must keep in mind that the inverted U hypothesis is the cornerstone of
RCT and is thus intrinsically a product of rational choice theorizing. This is so
because the feasibility of political violence is contingent on a groups ability to
rationally survey the political landscape and make logical decisions on the
probabilities for success.
Rational choice theorists claim that they do not necessarily disregard all
elements of RDT, but rather assert that frustrations and deprivations do play a role in
collective decisions to engage in political conflict. However, these decisions are
perceived by RDT as only relevant when the cost-benefit calculations tend to favor
political rebellion over inaction or apathy. The fundamental difference between RDT
and RCT is that the former views violence as being the impetus for political rebellion-
17


that it has utility and a primary purpose. In other words, RDT views violence and
collective discontent as being the foremost motivating factor in facilitating political
violence, while RCT understands violence and collective discontent as merely being
secondary or tertiary factors in the overall decision making process (Gurr, 1970).
One of the significant problems with RCT is its empirical testability, in that
the only thing that can logically be tested for is the existence of the inverted U
hypothesis. These tests typically only included measuring the degree to which a
regime is either repressive, semi-repressive, non-repressive. Most notably is RCTs
inability to actually test and measure the intricate complexities involved in the
collective mechanisms used to formulate and make calculations regarding violent
political rebellion.
Most empirical studies on RCT are therefore simply methodologically ill
equipped to measure the fundamental concept that the theory advances- the expected
benefits or costs that a group might incur during a political rebellion. This fact is
bared-out when one recognizes the fact that there simply have been very few
empirical studies created to test the primary concept of the rational choice approach.
Moreover, most studies on political violence, operating under the rational choice
paradigm, have merely only tested regime repressiveness as an indicator for
rational actions and cost-benefit analyses.
Most importantly is the fact that RCT has a difficult, if not impossible, time
dealing with individual-level perceptions. This is embodied in the types of research
18


that has been conducted on political violence using RCT. Most of these studies
require the measuring of not rational decisions per se, but rather the constructing of
studies designed to test the inverted U-curve hypothesis. As such, degree of
repressiveness becomes the main independent variable used to test political violence.
This is, of course, based on the assumption that the political behavior of individuals is
contingent on the opportunities or constraints that are manifested through the political
structure of a given regime (Muller & Weede, 1990).
This indirect approach to testing political violence is dependent on ones
ability to transcend objectivity and formulate subjective assumptions about how, or if,
regime repressiveness is an appropriate indicator of collective rational action. Some
have argued, however, that this approach is not necessarily problematic, whereby,
one always needs to make auxiliary assumptions about how to measure deprivation
or the presumed costs and benefits of resorting to the use of political violence
(Muller & Weede, p. 646,1990). Regarding this Simon (1985) asserts:
The rationality principle, as it is incorporated in theories of substantive,
rationality, will provide us with only limited help in understanding pol-
itical phenomena. Before we apply the methods of economic reasoning
to political behavior, we must characterize the political situation, not as
it appears subjectively to the actors. We can only select the appropriate
model to adaptation after we undertake the requisite empirical study to
determine this subjective representation both of goal and of the situat-
ion. (p. 297-298)
The prerequisite that RCT be tested by the use of auxiliary assumptions
because of its fundamental difficulty in testing the actual micro-level mechanism of
19


human rationality, is an issue that the social sciences have been grappling with for
decades, if not centuries. In a more philosophically abstract sense, the basic
underlining principles of scientific inquiry as a intellectual concept is, in itself, ill
equipped to directly test worldly phenomena (Lakatos, 1968-1969). As such, the
consequences and explanatory value of scientific research is heavily dependent on the
researchers ability to formulate and construct a logical set of auxiliary assumptions
(Lakatos, 1968-69; Muller & Weede, 1998).
Even though both RDT and RCT require the researcher to make auxiliary
assumptions about the fundamental explanatory value of the theory, I will argue that
the assumptions made in this study require a significantly smaller leap in logic than
what RCT, or previous studies on RDT, demand. However, one must keep in mind
that the general validity of both RDT and RCT approaches are no less valid because
of these assumption (Muller & Weede, 1998), but rather their validity is a function of
what specific types of auxiliary assumptions the researcher uses.
Another significant point of departure between RCT and RDT is that the
former is most useful in only explaining what I will classify as, top-down rebellions.
This is when social change, via collective rebellions, is decided upon by those that
have the most likelihood of bringing about that change. An illustration of this is
presented by Weede & Muller (1998) by hypothesizing that if a colonel in charge of a
tank unit decides to side with the rebels and attack the current regime, he will have a
greater amount of influence on the probability for success than the impoverished
20


farmer whos resources are virtually nonexistent, and whos actions will have limited
effect on the success of the rebellion.
Therefore, the accumulation of resources, by way of elite status, is viewed as
being a significant determinate of ones decision to participate in rebellious activities.
If resources (economic, social, intellectual, human, tangible, etc.) are understood as
being facilitators of rebellion success, then logically, one that possesses such
resources is the one that would view rebellion as being an enterprise whose
probability for success is high.
This, however, begs the question- if one is experiencing the privileges of a
sociopolitical hierarchy that is directed in ones favor, why then would that individual
demand social change? It seems very unreasonable and contradictory, and thus
contrary to RCT, for one to demand social change, even when that very person would
seem to be negatively effected from a change in the current status quo.
However, the cornerstone to RCT is built upon the notion that even if an elite
is subject to negative consequences from a failed rebellion, the benefits are much
more significant if that rebellion is to succeed. This is the fundamental premise of
RCT; simple cost-benefit calculations that are designed to dictate ones actions. With
the phenomenon of political rebellions, it is reasonable to expect that the richest
rewards go to the most resourceful actors contributing to the transfer of power
(Weede & Muller, p. 46,1998).
21


According to this perceptive, one should expect to find instances in which
political rebellion is a function of elite resourcefulness, and elite mobilization. Thus,
it is hypothesized that elite rebellions (top-down) are more likely to occur than mass
rebellions (Weede & Muller, 1998). The latter, what I classify as bottom-up
rebellions, are defined as being rebellions where there exists an absence of elite
leadership and are perceived to be rare occurrences.
Although there does exist some degree of literature in the filed of elite-lead
political violence in revolutionary studies (Brinton, 1965; Skocpol, 1979; Pipes,
1990), these tend to be singular case studies that do not represent the totality, or even
the general nature of political violence. Most of these studies were explicitly
designed to document specific cases of elitist revolutions, like France, China, and
Russia, and were not interested in bottom-up revolutions or violent rebellious that
occur from state-sponsored repression of a specific culturally distinct group. Rational
choice theorists cannot take into account such non-elitist, or bottom-up rebellions
such as those that have occurred among Southern Sudanese, Karens in Burma,
Muslims in the Philippines, Muslims in Chad, Kurds in Iraq, Nagas and Mizos in
India, Muslims in Thailand, Bengalis in Pakistan, and must and recent violence that
has erupted in Haiti (Horowitz, 1981). Such cases are more exemplary of RDT where
one would argue that resourcefulness is negatively associated to frustrations or
deprivations. In other words, according to RDT, when one is better off, he or she is
less likely to feel deprived and frustrated, and is therefore, less likely to rebel.
22


Another significant problem with the RCT model of elitist rebellion is its
testability. Russell (1974), Walt (1992), and Zimmermann (1990) have compiled lists
of successful and attempted mass rebellions over the last century. Of these most mass
rebellions failed to succeed in overthrowing the ruling class or the established
regime (Weede & Muller, 1998). The problem being that because this is a rational
choice conceptualization of political violence, these studies only tended to look at the
successfulness of a revolution, thus falsely assuming that if a revolution did not
succeed then it was generally dismissed. Most of these studies only concerned
themselves with the frequency of successful revolutions, where unsuccessful
revolutions were discounted (most of which were bottom-up in nature). This
approached is grounded in RCT, whereby assuming that rational actors will only
partake in predictably successful enterprises. This study, however, is not concerned
with the success of a rebellion, just whether or not one actually occurred. It is thus
designed to analyze how a specific societal phenomenon (cultural repression) sets the
process of rebellion into motion.
Moreover, there are significant problems with the validity of rational choice
theorists constructing studies using mere lists of civil wars and mass rebellions. Not
only are these types of lists hard to come by, but, by in large, they were not compiled
solely for the use of testing rational choice hypotheses. Small & Singers (1982) list,
for example, codes such events as attempted irregular executive transfers (Weede &
Muller, p. 47,1998), which, at best, can only be used as proxies for elite rebellion
23


(Weede & Muller, 1998). Even the Journal of Peace Research expresses concern
over the validity of using irregular transfers as an indicator for elite rebellion (Weede
& Muller, 1998).
The notion that some rational choice theorists advance, that revolutions are the
manufacturing of the few elitists that have the means to do so (Muller & Weede,
1990, Weede & Muller, 1998; Tullock, 1974; Russell, 1974), is contradicted outright
by Horowitz (1981). Here Horowitz contends that by far the largest number of
secessionists can be characterized as [oppressed] groups living in [oppressed]
regions (1981, p. 173).2
His studies indicate that oppressed and frustrated groups will generally realize
that their collective wellbeing is of little concern to the dominant regime, and they
will eventually rebel. He goes on to state that when the denial of such opportunities
is coupled with clear signs that the state is dominated by members of other groups,
[oppressed] groups in [oppressed] regions choose to opt out (Horowitz, 1981, p.
173). Furthermore, he argues that this degree of oppression and overwhelming lack
of opportunities will be so profound that the group will likely secede even in the face
of negative costs and significant collective sacrifice. Horowitzs point that
oppression and discrimination play a significantly larger role in secessionism than
2 Here, Horowitzs esoteric use of the word backward is substituted with the word oppressed. This
was done merely for the sake of clarity, and the two words have the same conceptual meaning.
24


resource mobilization and rational choice decisions. This is central not only to RDT
in general, but is also to this study specifically.
Game Theory
Although game theory maintains some fundamental similarities to rational
choice theory, there do exist some key differences in the way it conceptualizes and
models conflict. As a theory designed to model the complex paths that are exhibited
in both the decisions people make, and the reactions people have to those decisions, it
has gone though ebbs and flows of popularity within the academic community. When
John Von Neumann and Oscar Morgenstem wrote their groundbreaking book Theory
of Games and Economic Behavior, most believed that this theory would be capable of
describing virtually all types of human interaction. From sociology, to psychology,
and from economics to political science, game theory quickly became viewed as
having unlimited explanatory value, especially in the area of conflict and mediation.
Problems soon arose, however, when some began to demonstrate that game
theory was too unrealistic and did not accurately represent the way in which people
truly experienced reality. Most opponents see human interactions as being less rigid
and more flexible- that the fluidity of interpersonal experiences were impoverished by
such overly inelastic theorizing. Controversy surrounding this theory increased
during the Cold War when game theory became the theory of choice among many
social scientists proposing hypotheses designed to model international relations
25


between the United States and Russia. Although game theorists have worked their
way back into mainstream academia within the last decade or so, there still exists
some strong opposition by those who demand more realistic models (Bennett, 1995).
Many people are often adverse to the use of the word game in game theory-
believing that it connotes a lacking in seriousness, and that it trivializes the intricate
and complex nature of human interaction. However, many game theorists contend
that the use of the word game is merely designed to explicate a situation in which at
least two interacting parties, pursuing their own idiosyncratic goals, have influence
over the situational reality of one another (Bennett, 1995; Lax & Sebenius, 1986). In
other words, the way in which a person, group, community, etc., reacts to a particular
situation will have consequences on the way in which a competing person, group,
community, etc., reacts to the same situation and so on.
However, because no actor has complete control over what other people will
do or how they will react in any given situation, the actor must always be cognizant
of what might be the possible actions of others, and must, therefore, act accordingly
(Bennett, 1995). For example, if actor A knows that the most probable chance for
success is contingent on what actor B will do, then actor A will attempt to influence
actor Bs decisions. However, actor B is in the same situation- success is contingent
on what actor A will do, thus hoping to influence actor As decisions. Knowing that
both actor A and actor B are aware of each others intentions, and that their decisions
26


are dependent on what the other one will do, there arises a number of possible
outcomes.
Although conflict does not always have to be the absolute consequence, it is
without question a possible outcome. Despite the possibility of there being mutual
threats, deceit, bluff, and counter-bluff, there is also a possibility of cooperation
arising; where all parties achieve some degree of benefit (Bennett, 1995). Such
collaborative advantage (Huxham & Macdonald, 1992), tend to operate in
conjunction with threats because particular goals by actors tend to bifurcate, whereby
creating a situation where conflict and cooperation become intertwined (Bennett,
1995; Lax & Sebenius, 1986).
Game theorists utilize models to demonstrate, in visual form, the intricacies
that exist between parties during the decision making process. Figure 2.1 below
illustrates the different paths and the different sets of possible contingencies that are
exhibited between two actors. This simple game tree, or extensive form model is
designed to provide an analytical guide through all possible outcomes of moves and
counter-movers between two actors (Bennett, 1995).
Not unlike a game of chess, one players movements and decisions are
dependent on anticipating what will be the moves of the other players. The player
will thus be forced into acting accordingly to the way the other player reacts to the
original situation. As one can imagine, the concepts of game theory are highly
dependent on ones use of strategy. This is simply ones ability to foresee all possible
27


moves and counter-moves by ones counterpart. It also requires one to modify ones
own moves based on anticipating what the other actor will do.
Figure 2.1. Structure of a Simple Game Tree, or Extensive Form Model. Player A moves firs,
choosing between moves a and b. B then chooses p or q. If B chooses q, A has a further choice either
between x, y, and z (if initial choice was a), or just between x and y (if initial choice was b) (Bennett,
1995).
An example of strategy would be if actor A has a specific goal that must be
met, but whose outcome is dependent on actor Bs ability to intervene and/or block A
28


from achieving its goals. Actor As strategy would be to choice path a or path q or
p depending on what actor B will do. Game theory proposes that the use of strategy
is fundamentally ones ability to use foresight to have a back-up plan in achieving
ones goals, regardless of the decisions the other actors make.
One of the primary aspects to game theory is the aims, or goals, that actors
have. In some respects, one players goals might come at the complete loss of the
other actor. Thus, when the achievement of ones goals comes at the expense of the
other actors gains, this is typically referred to as a zero-sum game. This is
characterized by ones aims being in direct contradiction to those of the other actors.
Because ones losses are another persons gain, these interactions are expressed by
utility scales, and the total utility is always the same, (Bennett, p. 22,1995), and is,
therefore, expressed as a constant. This is because the constant is measuring the
losses relative to the gains of the other actors; it is thus categorized as being zero-
sum.
However, this occurrence only constitutes one end of the game theory
spectrum. At the other end of the spectrum is when both actors desire the exact same
outcome. Labeled pure-coordination, (Huxham & Macdonald, 1992), these
interactions still constitute a problem among actors in the game. The problem arises
from the fact that both sides have to determine what the best path is in achieving what
both commonly desire. In this situation, it is not necessarily the ends that are at issue,
but rather the means that are employed to secure those ends.
29


Most common game-type interactions are what are called mixed-motive (Lax
& Sebenius, 1986). In this situation, the desired outcomes of all actors are disjointed,
and are incongruent from one another. For example, one group might desire free
international trade without governmental regulation, while another group desires free
trade, but within the confines of fairness policies, yet another group might desire
governmental sanctions imposed on environmentally unsound nations. Although all
groups fundamentally desire trade, each group nevertheless has conceptually different
means by which to secure those ends. In this type of situation, groups are usually tom
between conflict and cooperation.
Most game theorists believe mixed-motive games to be closer to the real word
interactions of groups (Lax & Sebenius, 1986; Huxham & Macdonald, 1992; Bennett,
1995). Bennett (1995) asserts, mixed-motive games are much closer to the real
world of partial conflicts, threats, bargaining and negotiation, dispute settlement, and
conflict resolution (p. 23). These are also seen as instances in which responses and
counter-responses might set into motion a sequence of events that inextricably lead to
conflict or resolution (Brams, 1985).
In figure 2.2 below, Bennett (1995) presents one of the single most popular
illustrations of game theory. The Cuban Missile Crises provides game theorists an
unequally clean and relatively simplistic model of brinksmanship. In this model, one
side sees how far they can push the other right before the point of conflict. However,
if neither side backs down, both will be forced with the worst of all possible
30


outcomes; in this case, nuclear war. Figure 2.2 demonstrates that in box 1,1 where
both sides, at least theoretically get what they want and stand firm, then they both
experience the worst possible situation- war.
The figure also illustrates each players possible strategies are shown along
one side of the matrix: one player controls the rows, the other the columns (Bennett,
1995). Each cell in the matrix represents the possible outcomes for each actor in
this case either the White House, or the Kremlin. The number indicate the degree
of choices each player has ranging from 1, which is back down, to 4, which is
stand firm. In this scenario, 4 is the best possible choice for that particular player,
where 1 is the worst choice. However, this is not a straightforward zero-sum game
because if each side were to get what they want, then the result would actually be the
worst of all possible outcomes for both sides. Bennett (1995) goes on to assert,
analyzing the game reveals two Nash equilibria, 4,2 and 2,4, which is a win for
either side.. .however the game is symmetric between he two players, so there is no
way of telling which side will establish a winning position (p. 24).
The popular prisoners dilemma is not too dissimilar from the above example.
The term is derived from a hypothetical situation in which two prisoners, held in
different interrogation rooms, have one of two options- either squealing, or remaining
silent. The dilemma arising from the fact that neither actor is completely sure what
the other will do, but that only two outcomes are possible. Not unlike figure 2.2
31


below, both actors can either win or both can lose depending on the decisions each
other make.
Back down
White House
Stand firm
Figure 2.2. Normal Form Model. Cuban Missile Crisis modeled as a game of chicken, showing
matrix (Bennett, 1995).
The premise behind the prisoners dilemma is that this type of situation is not
dissimilar those that might lead to either political violence, or political compromise.
If, for example, a group is being oppressed, that group might conclude that violence is
the only viable option- believing that the response by the government will only be
met with more repression. In this situation, either the group will do nothing and
continue to be repressed, or they act out in an attempt to bring about social change.
In the latter case, they have nothing to lose being that they are already repressed.
Strategy would also be employed whereby a repressed group might create a
back-up plan to a given set of outcomes. For example, if petitioning the government
for change does not work, then lesser forms of violence might be utilized in affecting
change. If the government acts by giving into the demands of the group, then the
Kremlin
Back down Stand firm
3,3 2,4
4,2 1,1
32


violence will be called-off. However, if the government reacts militarily to crush the
rebellion, the rebels might either decide to step-up their campaign of violence, or
scale back and give into the governments force. As one can see, the tree of
outcomes begins to bifurcate and create new paths with every move and counter-
move made by each group.
As indicated in the introduction of this section, there exists a very strong
opposition to game theory. Most of the criticism levied against this theory are based
on the idea that its modeling is to unrealistic and that it is incapable of taking into
account the eloquent ebbs and flows of human interaction (Ward, 2002). In many
instances human interaction exhibits unforeseen trajectories that can either cause
violence to be avoided outright, or have it spiral out of control (Young, 1996).
Furthermore, because this theory is grounded in the rational choice paradigm
(presupposing that humans are rational actors whose inclinations are the product of
careful and logical deliberations) it becomes subject to the same types of criticisms
brought against rational choice theory. Not the least of which is the criticism that
humans do not always act in accord with reason and logic (Horowitz, 1981). This is
exemplified in game theorys own model. In figure 2.2 and in other matrixes that
model the prisoners dilemma, one notices that the most reasonable of situations-
when one is acting to maximize ones benefits and minimize ones costs- tend to
exhibit the worst outcome (Jervis, 1976). This unreasonableness is manifested in the
sense that, especially when dealing with political violence, groups tend to act
33


emotionally and without a great deal of thought and mental calculations of costs and
benefits (Horowitz, 1981; Waller, 2002). In many instances, groups will often
participate in actions that might actually harm the well being of the group (Waller,
2002). Along these lines Neuberger (1987) actually contends that many times there is
actually is no rational basis to explain the actions of groups pushed to their limits
(P-94).
Another problem with game theory is that it fundamentally assumes that all
parties are aware that the game actually exists. Modeling of a game would make
little sense if there was only one player and that the other parties were completely
unaware that a game is even being played. For there to be a game under the game
theory concept, all sides must share the same view of what the issue(s) is, who the
players are, and what ones strategies are. The problem being a cursory examination
of real life suggests that this is seldom so (Bennett, 1995 p. 27). Therefore, most
people actually have no idea what the game is, who the players are, and what their
strategies would be. This fact undermines the entire premise upon which game theory
is grounded (Axelrod, 1984; Holsti, 1962; Jervis, 1976).
When related to political violence it is easy to notice that there exists a
significant degree of uncertainty and ambiguity as to what the intentions are of either
side. Moreover, because both sides typically are not even fully sensitive of who all
the players are within the totality of the game itself, there is no possible way for
34


actors to formulate and construct moves and counter-moves, thereby, negating the
entire presupposition of the theory.
State Power Theory
State power theory or state-centered theory is a worldview that expresses
political violence, and specifically secessionism, as being products of how vulnerable
a nation is. This is more of a politically centered theory that contends if a nations
political structure is characterized as being weak and unstable, then it is hypothesized
as being susceptible to political rebellion and even revolution. There are many sub-
fields to this theory that can be approached by many different angles. Therefore, this
section is designed to explain the overall ideas that are encapsulated within this
theory.
State power theory has evolved from early Marxist theories of class struggles
and proletarian revolution. According to this paradigm, revolutions are understood as
being the products of a ruling bourgeois facing increased pressures from a
discontented under class, or proletariat (Collins, 1993). Under Marxian philosophy of
social change, ruling ideas were those of the ruling elite, and as the rift between the
ideas of the bourgeois and the proletariat grew, challengers emerged whose change in
consciousness worked to mobilize the oppressed for the coming revolution (Collins,
1993).
35


However, according to Randall Collins (1993), in recent decades, the old
class-conflict paradigm of revolution has broken down; vestiges remain in current
thinking about ideology, rather like a flywheel spinning after the drive shaft has
broken off (p. 118). State-power theory, has been the new paradigm that has stepped
up to fill the space left empty by Marxian theory of class-conflict. One of the
fundamental points of departure of state-power theory from Marxian theory is that the
former views revolutions as being a breakdown in the upper echelons of government,
rather then from insurgencies from the bottom levels of society.
Goldstones (1991) Revolution and Rebellion in the Early Modern World,
articulates the position that it would seem illogical for one to assume that revolutions
come from the bottom-up because the disenfranchised simply do not posses the
means in which to bring about social change via revolution. It is explained that as the
ruling class possesses the military and police apparatuses, oppressed groups will
essentially have to bide their time and wait for the state to weaken at the top. Tillys
(1978) research on this matter has presented a position in which revolutions and
political rebellions are constrained by the sociopolitical environment in which they
take place. If there exists a strong sociopolitical environment that a group is
attempting to rebel against, a revolution will be all but impossible (Tilly, 1978).
Goldstone (1991) postulates that for there to be an opening of the door for
political rebellion to move through, there first must exist three primary components to
indicate a breakdown in the state. Fist, there must be fiscal stain, whereby the ruling
36


elite must be unable to finance their benefactors and, primarily, their military and
police apparatuses (Goldstone 1991; Collins, 1993). Secondly, there must exist intra-
elite conflict and antagonisms that render the decision-making process of the ruling
elite ineffectual, which incapacitates coherent governmental interdiction of rebel
forces. Lastly, there must be, in conjunction with the other two processes, popular
revolt and a collective desire to overthrow the current regime (Goldstone, 1991;
Collins, 1993).
One of the main avenues of approach under the umbrella of state power theory
is the vulnerability issue and now it relates to either facilitating or deterring
secessionist conflict. Most vulnerability adherents contend that states that are
susceptible to secessionism are less likely themselves to support or otherwise
condone secessionist activities outside their own nation (Saideman, 1997; Treisman,
1997; Hill & Jewett, 1994).
The key premise to the vulnerability hypothesis has best been articulated by
Zartman (1966), who said the greatest deterrent to territorial revisionism has been
the fear of opening a Pandoras Box... If any one boundary is seriously questioned,
why not all boundaries (p. 109)? As a result, concerns over the vulnerability of
ones nation should therefore deter policy makers from supporting secessionist
activities in neighboring regions. Additionally, fears that external unrest might have a
cascading effect into ones own nation, typically causes heads of states to work to
strengthen international alliances and territorial boundaries. Such groups as
37


Organization of African Unity (OAU) are the byproducts of these collective fears,
and are designed to quell political unrest throughout Africa and work to prevent its
spread (Buchheit, 1978; Foltz, 1991; Saideman, 1997).
Separatism is perceived as being a serious threat to the stability of nations
within the African continent. Several studies on this phenomenon have demonstrated
that political rebellion seems to be endemic and moves from one country to the next
(Neuberger, 1987; Kier & Mercer, 1996; Saideman, 1997). This is exasperated in
Africa because there exists so many culturally distinct groups, and as a result, leaders
fear that successful rebellions cause some groups to question the validity of politically
defined national borders (Neuberger, 1987; Kier & Mercer).
However, if the vulnerability hypothesis was truly valid, then the last decade
would have proved to be much more peaceful (Saideman, 1997). Secessionist
inclinations in Kosovo, and in the Sandzak region should have deterred the actions of
Serbia. Additionally, secessionist effects of Croatias Serbian minority should have
inhibited Croatias ambitions in Bosnia (Saideman, 1997 p. 722). The another real-
world example that demonstrates the weak explanatory value of the vulnerability
hypothesis is that the disintegration threat should have prevented Russia from
supporting secessionist activities throughout the former Soviet Empire, but this, too,
was not the case (Hill & Jewett, 1994).
The vulnerability hypothesis has an even more difficult time in explaining the
very part of the world that it was explicitly designed to analyze. Secessionist
38


conflicts throughout Africa exhibits a pattern of activity that can be best characterized
as being the antithesis of the vulnerability hypothesis. Most analysts (Jackson &
Rosberg, 1982, Jackson, 1990; Bull, 1997) advance the idea that African nations that
are themselves susceptible to secessionist activities within their own boarders, should
not only be deterred in supporting secessionists in neighboring regions, but in many
instances, actively work to prevent such movements. Moreover, being that state
stability is in the natural interest of all nations, one should expect to find that this
common interest of state stability should compel all nations in Africa, and
elsewhere, to prevent secessionism.
The problem with this analysis is that it relies on the notion that all nations are
either cognizant of such common interests, or that they care about them in the first
place. Saideman (1997) claims, the problem is that in each African secessionist
crises, most of the secessionists important supporters were facing their own ongoing
separatist movement (p. 722). In other words, contrary to what the vulnerability
hypothesis posits, most secessionist movements in Africa garnered support from
nations who themselves were subject to internal secessionist movements.
Another one of the glaring problems with the vulnerability hypothesis is that
its a complete abnegation of the domestic policy issues that exist within any given
nation. The hypothesis merely presupposes that vulnerable nations will not support
and/or will actively work to prevent regional secessionist movements without taking
into account the domestic political landscape of a given country. It is merely a
39


hypothesis that allows simplistic geopolitical calculations to override the complex
nature of domestic political strategizing.
The term state stability may seem like a straightforward and unsophisticated
term, however it is multi-dimensional and operates on more than a geopolitical level.
Even though the leaders of nation A, for example, are themselves susceptible to
internal discord and rebellion, one should not automatically assume that country A
will participate in a wholesale admonishment of secessionist movements, even those
within its own region. The leaders of country A have to be mindful of the demands of
their own constituency, and if actually supporting a secessionist movement of a
neighboring state will curry support from their citizenry, then the notion of regional
common interest is generally discarded.
Saadia Touval presents the point that the vulnerability of African nations, due
to their own internal political conflicts, create situations in which they tend to not
support secessionist movements or engage in irredentism (Saideman, 1997). Touvals
studies on Africa demonstrate that nations are generally weary about supporting
secessionism due to the endemic nature of political violence (Touval, 1972). In this
respect, the validity of nationally recognized boundaries and the admonishment of
secessionism and irredentism would be in the best interest of African nations.
Touvals studies have pointed to Somalia as being the instantiation of her belief that
African nations tend to not be in favor of recognizing secessionist movements for fear
40


of themselves being subjected to such conflict (Touval, 1972). However, Saideman
(1997) contends that Somalia has supported regional secessionist activities and that:
the rise of clan conflict in the early 1990s and the de facto secession
of Somaliland (northern Somalia) demonstrate that Somalia is current-
ly vulnerable to ethnic conflict, an examination of its history indicates-
that secessionism and ethnic conflict have plagued Somalia since it be-
came independent in 1960, thus challenging Touvals assertion, (p. 723)
Jackson & Rosbergs (1982) studies also indicate that there are norms that
exist within the African continent which implicitly castigate nations from attempting
to support regional secessionism and thus work to pursue state stability. The
influential nature of Africas international community, therefore, is a facilitating force
in the preservation of national sovereignty and postulates a position in which nations
operate toward the common interests of all African states (Jackson & Rosberg, 1982).
Jackson & Rosberg (1982) go on to claim that there is a common interest in the
support of international rules and institutional and state jurisdictions in the African
region that derives from the common vulnerability of states and the insecurity of
statesman (p. 18). Yet, their notion of common interest is still conceptually ill
equipped to explain why is it that nations support secessionist movements and
suggesting only that a state that was invulnerable to secessionism could do so if
desired (Saideman, 1997, p. 724). Hence, because nations do both condone and
actively support secessionist movements in some cases, the vulnerability hypothesis
has very limited explanatory value (Saideman, 1997).
41


Another sub-field that has garnered a significant deal of attention under the
state-power paradigm is that of decolonization (Bergesen & Schoenberg, 1980;
McGowan, 1985). Although this hypothesis advances the same fundamental
philosophies as those presented under the vulnerability hypothesis, specifically, and
state-power theory, generally (that weak state tend to be more susceptible to political
rebellions) it approaches the phenomenon from a slightly different angle.
Focusing on the intricacies and the dynamical behavior of decolonized regions
by mostly Western democracies, the decolonization hypothesis posits that the process
of colonization and decolonization inexorability leads to dependency and the
emergence of weak nations (Strang, 1990). These weak nations are, therefore, unable
to neither construct a strong and viable sociopolitical infrastructure, nor are they
capable of putting-down political rebellions once they have taken shape. In either
respect, weak nations, created by decolonization and dependency on core nations,
often become subjected to coup d'etat, revolution, or civil war.
A number of empirical studies have worked to advance the understanding of
decolonization and what affects it has on the creation of weak nations (Bergesen &
Schoenberg, 1980; McGowan, 1985; Boswell, 1989). Most of these studies postulate
that since the end of World War II, Western empires began to disassemble and pull
back from their far-flung colonies (Strang, 1990). This emergence of newly
constituted independent nations worked to redefine the international economic and
political structure.
42


Under the world economic structure perceptive, some propose that
decolonization creates a system in which international exchange of goods became
hierarchically situated into a ridged division of labor (Wallerstein, 1974; Chase-Dunn
& Rubinson, 1979). Under this new system evolved capitalism which is, for the first
time, a truly global system held together, not so much with political or military ties,
but by economic and technological ones (Wallerstein, 1974). This has made possible
the rapid and almost total expansion of the modem capitalist world economy-
specifically by the core nations of Western Europe and the United States. According
to Wallerstein (1974), different parts of the world-system maintain different functions
and thus specialize in different areas. The world, as opposed to just a particular
nation, is therefore, broken-up into a division of labor.
The relative underdevelopment of the poorest nations in the world constitutes
the periphery nations. It is these areas of the world that cheaply supplies the raw
materials to the worlds core nations for development and enterprise. The core
nations therefore, are made-up of the developed and modernized areas of the world,
whose powers of technological manipulation, economic exploitation, and
militarization work to maintain the power structure throughout the world. One
critical economic component of this division of labor would be, for example, foreign
investment, which results in comparatively high rates of profit earned by core
investors.
43


Most importantly, Wallerstein (1974) further argues that because of this
socioeconomic dynamic in the modem world-system- where the core nations are able
to exploit the resources of periphery nations- core nations have the sociopolitical
capacity to create strong nation-states, where peripheral nations have weak ones.
This phenomenon of weak states creates even further problems for the periphery
nations insofar as they become subject to possible external invasion or internal strife.
In the core nations, economic prosperity, created, in part, by growing bureaucracies
and a tremendously strong force structure, generates social stability, which in turn
creates political stability. However, the capitalist world-system is one of economic
integration, and it is through economic mechanisms that the core controls the
periphery (Wallace & Wolf, p. 161).
Just as the world economic structure changes due to decolonization, so to does
world polity. The institutionalized structure that evolves from decolonization
becomes characterized as being power of domination (Strang, 1990; March & Olson,
1984). This dynamic of external constraint by core nations also involves the spread
of political structures (Strang, 1990).
The first step in the decolonization process involves the imposition of the
nation-state concept by Western powers onto their dependencies (Fieldhouse, 1966).
Next, the nation-state concept is galvanized into the collective consciences of the
masses by means of the imposition of a nationalistic belief system. However, having
44


a strong nationalistic sociopolitical system can, itself, lead to internal power struggles
brought about by militant movements of nationalism.
The problem with the decolonization hypothesis is that it has yielded, at best,
weak empirical correlations between decolonization, weak nations, and political
rebellion (Chase-Dunn & Rubinson, 1979; Strang, 1990). There has been even less
empirical support derived from the testing of the most fundamental of decolonization
concepts. The argument that the core nations control the periphery, and that the core
works to expand its hegemony, tends to be in serious empirical doubt (Strang, 1990).
Strang (1990) states that it is difficult to empirically distinguish the impact of
hegemony and global economic cycles over the twentieth century (p. 857). In other
words, empirical tests have been inconclusive regarding just how much core-nations
economically control periphery nation. If it is the case that core nations do not
economically control the periphery, then the notion of dependency (at least economic
dependency) is questioned. Therefore, this throws into doubt what in fact can
account for the emergence of weak nations that are unable to defend themselves from
internal rebellion.
There tends do be, however, some empirical support surrounding the world
polity argument (Goldsworthy, 1971, Strang, 1990). This notion contends that
decolonization is facilitated by the transmission of the nation-state model from
Western powers to their dependencies (Strang, 1990, p. 847). However, this
45


hypothesis only helps explain the phenomenon of decolonization, and does not
necessarily explain the creation of weak states, or the rise of political rebellions.
Resource Mobilization Theory
Resource mobilization theory (RMT) is an approach that had evolved from the
social movements that had characterized the late 1960s and early 1970s. In the
1980s, however, it began to adopt a new emphasis in analyzing and explaining
political movement that might lead to violent political rebellions and even
revolutions. Although RMT has produced some empirical studies that tend to
support its basic premise, there are significant disagreements over how well suited
this structurally orientated theory is in explaining violent political rebellions (Jenkins,
1983; Kerbo, 1982; Tilly, 1978).
In the 1980s RMT became a new paradigm that took into account the totality
of the movement, including the institutionalized actions, the rationality of movement
actors, the strategic problems confronted by movements, and the role of movements
as agencies for social change (Jenkins, 1983, p. 528). This perspective of social
movements was designed to question the individually based, social-psychological,
and frustration orientated relative depravation theory presented by Gurr some years
prior. Resource mobilization theorists countered Gurrs theory by claiming that
because deprivation seems to be almost omnipresent throughout the entire world, and
because there are obviously not revolutions occurring everywhere, stress and strains
46


need not be a primary element in political movements (Kerbo, 1982; McCarthy &
Zald, 1977).
Instead of relative deprivations, RMT advances the notion that social
movements are the results of the ability of social movement promoters to procure and
manipulate recourses of power (Kerbo, 1982). Such recourses of power included
ones ability to organize, to recruit members from existing voluntary pools of
association networks, and provided individual incentives or coercion in motivating
participation in social movement activities (Kerbo, 1982, p. 646).
However, there exist some of fundamental empirical and theoretical problems
with RMT that calls into question the validity of the theory. The primary problem
with the theory is that there simply are far too much historical data to support the
notion that the level of a groups deprivation and discontent are in fact the primary
motivating factors in the precipitation of political violence and even revolutions
(Salert, 1976; Kerbo, 1982). Salert (1976) criticizes the basic contention of most
recourse mobilization theorists (Olson, 1965; Jenkin, 1983; Goldstone, 1980) that
collective action functions as a deliberate and intricate process of collation formation,
incentive distribution, and power brokering. Rather, Salert (1976) explicates
collective action as being spontaneous acts of mass participation, which is at the
center of many revolutionary events, not the least of which included the early stages
of the French Revolution of 1789.
47


A widely used example of resource mobilization operating to create
revolutionary events is the Russian Revolution of 1917 (Olson, 1965; McCarthy &
Zald, 1977). About this McCarthy & Zald (1977) argue that the Bolsheviks had
prevailed because of the large amounts of resource flows they experienced into their
organizational center. However, because many workers believed that it would not be
rational to risk their lives and/or their resources in a revolution against the bourgeois
government, the October Revolution of 1917 had to rely instead on a small collection
of conspirational elites (Kerbo, 1982; Olson, 1965). These conspirational elites
were able to utilize their collection of resources, and ability to recruit, to take
advantage of a weak government during a period of social disorganization.
However, according to Kerbo (1982) this begs the question of what brought
about the conditions of anger, hostility, frustration, and social disorganization that led
to mass rebellion and the successful overthrow of the government by the
conspiratorial elites in the first place? Skocpol (1978) answers this by stating it was
not that a communist party created popular organizations.. .rather, it first allied with
spontaneously created or popularly supported bodies (p. 17). The key issue that
resource mobilization theorists have to contend with is that movements such as these
are initiated by massive levels of societal discontent borne out of situations of
unemployment, shortages in food and energy, and other such deplorable conditions
that characterize impoverished nations. About this Kerbo (1982) puts forth one of the
most logical and concise criticisms of RMT by asserting dismissing such factors as
48


important underlying precipitants for social movements, as the recourse mobilization
thesis does, leads to an adequate explanation of only some types of social
movements (p. 652, emphasis added).
Kerbo (1982) goes on to make the important distinction between movements
of crisis and movements of affluence. Not necessarily proposing a wholesale negation
of RMT, Kerbo (1982) contends that relative deprivation theory, with its emphasis on
strain, anger, stress and other psychological factors, is theoretically better equipped to
deal with movements of crisis while, RMT is theoretically better equipped to deal
with movements of affluence.
The difference between movements of crisis and movements of affluence
represent fundamental distinctions between what RMT is equipped to explain and
what RDT is equipped to explain, and is thus the backbone of this study. Since this
study is explicitly designed to test what affects cultural deprivations have on violent
political rebellions, and hence, movements of crisis, the use of RDT provides a better
conceptual justification then the use of RMT. Kerbo (1982) goes on to conceptualize
social movements as operating on a continuum of movement characteristics.
Figure 2.3 below illustrates that on one side of the continuum there are movements
that are characterized as being the result of life-threatening sociopolitical situations.
These are what he hypothesizes as being the types of movements that are better
explained with the utilization of RDT, and on the other end of the continuum there are
movements of affluence. These are characterized as being movements of collective
49


action, in which the major participates do not experience, nor are they motivated by,
immediate life-threatening situations of political or economic crisis, but rather, have
their basic needs of life met, or even in abundance (Kerbo, 1982, p. 654).
Ironically, RMTs theoretical foundation is preconditioned on the notion that
those that facilitate and are active in movement activities are the very ones that can
do so because their basic life needs are met. These people therefore, possess the
surpluses in both energy and resources, such as money, time, social networks, and
intelligence, to participate in such activities (McCarthy & Zald, 1977). Using
McCarthy & Zalds studies on social movement in the United State of the 1960s,
Kerbo (1982) proposes that these are best described as movements of affluence, and
are generally not characteristic of mass violent political rebellions. This is so
because, as Kerbo (1982) states:
movements of crisis are made up primarily of beneficiary members
individuals experiencing life-threatening conditions, while move-
ments of affluence may have a higher number of conscience memb-
ers individuals motivated primarily by ideology and moral issues-
who have surplus resources to devote to causes from which they
benefit less directly, (p. 657)
50


Movements af Crisis
1) Preconditions found in times of
life threatening political and/
or socioeconomic crisis
2) Participants primarily
beneficiary members
3) Participants primarily
movement specific
4) Early stage of the movements
usually spontaneous and
relatively unorganized
5} Collective violence and hostile
outbursts more often involved
(at least in early stages)
6) Less systematic use of
Individual incentives to
motivate movement participation
(at least in early stage)
Movements of Affluence
1) Preconditions found in times of
affluence O'1 relatively iOod
political and economic periods
2) Participants often include a
greater number of conscience
members
3) Participants often multi-
muvement oriented
4) Movements more likely to begin
with a social movement
organization and leadership
structure
5) Less likely to involve collective
violence and hostile outbursts
6} The systematic use of individual
incentives to motivate movement
participation
Figure 2.3. Movements of Crisis and Affluence: Continuum of Differing Movement Characteristics
(Kerbo, 1982).
Relative Deprivation Theory
Although relative deprivation theory (RDT) has been discussed throughout
this study, this section will be dedicated to discussing its theoretical roots, evolution,
limitations and advantages, and its application to this study. As being one of the first
pure theories to explain political violence, it has experienced both significant
criticisms and praise. It has been tested and retested, yielding both sound and
ambiguous empirical results. However, as indicated numerous times, this particular
study is an attempt to move beyond the previous conceptual boundaries that have
been constructed, and approach the phenomenon of secessionist conflict from a
different relative deprivation angle.
51


The core premise of RDT is centered on the frustration-aggression
hypothesis, which seeks to explain the relationships between the origins and objects
of aggression. Frustration-aggression is viewed as being the result of accumulated
frustrations arising from the experiencing of extreme stresses and strains that build-up
to produce aggressive behavior. Therefore, aggression which is produced by the
conjunction of different sets of frustrations but directed at fewer than all of the
frustrating groups (usually just one) is called cumulative aggression (Horowitz, 1973
p. 2). It is thus theorized that the intensity of cumulative aggression is a function of
the magnitude of accumulated frustrations (Horowitz, 1973). When applied to
political violence, one can see that where societies are sharply divided into casts of
cultural groups, these societies will tend to exhibit higher levels of violent aggression
caused by collective frustrations over cultural repression.
The classical definition of the frustration-aggression hypothesis asserts that
aggression- a term that includes violence and other attempts to inflict injury- is the
result of anger produced by the thwarting of a goal-directed activity (Horowitz,
1973, p. 2). The interference of this goal attainment activity is committed by a
thwarting agent, or instigator of anger (in the case of state-sponsored repression this,
of course, would be the government), which becomes the source of frustration. In the
absence of internal or societal restraint, this frustration is converted into aggressive
and violent behavior that is often directed at the thwarting agent. The basic premise
of frustration-aggression is relatively simple and straightforward. If, for example, one
52


is actively prevented from achieving ones goals, or worst yet, is kept horribly
repressed, one is likely to experience frustration, even to the point where aggression
becomes a viable option to get what one wants (Horowitz, 1973).
Originally created by Dollard et al. (1939) the frustration-aggression
hypothesis became a good conceptual fit to apply to political violence. Because it is a
phenomenon that is understood as being a goal-directed behavior, frustration-
aggression is relatively easy to conceptualize as being responsible for mass
movements of political violence. This is consistent with Mandlers (1975) theory of
emotion, whereby when goal-directed behavior is interrupted or blocked, and
therefore one cannot fulfill ones goals, the result becomes frustration-laden arousal.
This is because when ones goals are being actively blocked (especially in the case of
governmental repression) then one becomes filled with a sense of danger because his
or her wellbeing is being threatened (Mandler, 1975).
Moving on to the essence of RDT itself, one has to keep in mind that this is an
inherently psychological concept which was defined by Ted Robert Gurr in the late
1960s as being the gap between what one has and what one expects. Gurr wanted to
look at the social-psychology of the situations in which aggressive behaviors occur.
He contended that the potential for political violence within a given society operates
as a function of how much discontent is experienced by members within that society
(Gurr, 1968). The gap between what people have and what they believe they should
53


have functions as a motivating factor as advanced by the frustration-aggression
hypothesis.
Into the 1970s, Gurrs RDT became a widely accepted theory that would be
used by almost all social scientists investigating violent political movements (Brush,
1996; see also Runciman, 1989). The notion that political revolutions were the result
of frustrations arising from collective expectations was not necessarily a wholly
unique concept. De Tocquevills writings on the French Revolution were grounded
in the idea that the masses became frustrated and therefore aggressed against the
government because their expectations were never fulfilled. The saying revolution
of rising expectations became the essence of Gurrs theory, and was also
summarized by Davies (1962) J-curve (Brush, 1996). With the J-curve, Davies
(1962) demonstrates how a rapid drop in the actual need satisfaction causes it to lag
much farther behind the continually growing expected need satisfaction, thus creating
an intolerable gap between what people want and what they get (p. 6).
The main thrust of relative deprivation studies focus on the extent that
economic inequalities facilitate collective sentiments of relative deprivation, which in
turn, leads less well-off members of society to aggress against those that are
responsible for their disposition (Schock, 1996). Gurr, therefore, postulates that
relative deprivation is a necessary condition for there to be political violence.
Although Gurrs theory also encompasses such things as normative justifications and
cultural threats that legitimize violence, his, and other relative deprivation theorists,
54


primary focus has been analyzing what effects economic conditions have in
predicting violent political discontent. As such, other situational factors, such as
culture and political context, have generally been seen as secondary factors, or are
generally left unstudied (Schock, 1996).
However, there do exist some problems with RDT, and one of the significant
problems lies with the frustration-aggression hypothesis, which is the foundation of
the theory. This problem is that widespread frustration may in fact lead to violence,
but such aggression should logically be directed at a verity of available targets and
should result in random outbursts of individual aggression (Newman & Erber, 2002).
Thus, collective states of frustration should not necessarily lead to a coherent mass
movement against the source of the frustration, and as a matter of fact, the aggression
cannot, by its nature, be directed at its source (Newman & Erber, 2002). This is
because, the classical interpretation of the hypothesis proposes that aggression is
displaced because it is unacceptable to aggress against the real source of the
frustration, requiring the aggressive energy to be expended in another way (Newman
& Erber, 2002, p. 122).
Another problem with the frustration-aggression hypothesis is that is was
primarily developed to explain individual-level behavior, and studies that have been
produced to test this hypothesis have generally focused on single subjects (Horowitz,
1993). As a result, there has been some difficulty in applying the theory to explain
collective behaviors. Additionally, just as with the case with individual violence,
55


mass political violence too might be the result of careful calculation and planning,
and not the product of spontaneous fits of anger and frustration. However, others
assert that violent secessionist conflict does tend to function as the byproduct of anger
produced by stresses and frustrations that arise within a situation of state-sponsored
repression (Horowitz, 1973; Gurr, 1968).
The way in which RDT has traditionally been tested, tends to constitute a
separate problem in itself. There is little doubt that the use of income as an indicator
for relative deprivation has by used ad infinitum.3 Most of these studies tend to be
somewhat supportive of the idea that as relative income inequality increases, so too
does the likelihood of political violence. However, Hardy (1979) and Weede (1987)
show that when studies control for the level of relative economic condition, the
affects of income on the dependent variable all but disappear (Dudley & Miller,
1998). Horowitz (1981) presents a scathing condemnation of why one is best served
to not utilize factors of income to measure relative deprivation:
I mean to reject direct causal relationships between regional econo-
mic disparity and secession. If degree of regional economic dispar-
ity alone determined the emergence of separatism, it would be reas-
onable to expect the preponderance of such movements in those sta-
tes occupying the middle income levels, for in such states regional-
economic disparities seem to be greatest. But no such tendencies
can be identified. Secession is attempted in low-income states.. .and
needless to say, it is also an issue in a number of economically deve-
loped countries, (p. 172)
3 For a general overview of studies that have used income as a primary indicator of relative deprivation
see Sigelman & Simpson, 1977; Muller, 1985; Gurr, 1985; Muller & Seligson, 1987; Boswell &
Dixon, 1990; Schock, 1996; Canache, 1996; and Dudley & Miller, 1998.
56


Moreover, there exist both conceptual and empirical problems with using
income as an indicator of relative deprivation. Conceptually, relative deprivation and
income inequality are two different phenomena. Income inequality is an objective
concept, whereas relative deprivation is an inherently subjective concept which is
premised on the collective perceptions of what a group has and what that group
believes it should have (Dudley & Miller, 1998). Therefore, using income inequality
to analyze the degree to which a group believes to be experiencing relative
depravation directly confounds equality with equity (Dudley & Miller, 1998). This
produces inconstancies in empirical studies that use income inequality to test for
relative depravations, whereby, utilizing an objective phenomenon to measure a
subjective concept. Discussing the differences that exist between equality and equity
Bronfenbrenner (1973) explains:
The terms equality and equity are widely confused. Despite their
phonetic similarities and philological connections, they are quite
distinct. The equality of a distribution of income or wealth is ba-
sically a matter of fact and is, therefore, basically objective. The
equity of the same distribution is basically a matter of ethical jud-
gement and is therefore, basically subjective, (p. 9)
This point is furthered by the research that has been conducted by Pen (1971)
who demonstrates that, based on his surveys of ethical belief systems, people perceive
income inequality in a verity of ways. He found that there where 21 different
ideologies about how people perceived income distribution, ranging from Socialist
notions of income redistribution, to lazier fare concepts of free markets and
57


unregulated competition. This indicates that the use of income to measure relative
depravations presents some very problematical issues with respect to how people
perceive their socioeconomic realities.
The use of income inequality introduces some empirical problems as well.
About this issue Dudley & Miller (1998) assert that the use of income ignores
noneconomic sources of relative deprivation, it excludes the potential effects of other
equally important concerns, such as social and cultural rights (p. 81). This
distinction is fundamental to the study at hand, in that I attempt to overcome these
conceptual and empirical problems by analyzing, not just factors of income as
causing relative deprivations, but also cultural factors. I contend that these problems
are transcended because the cultural capital index (see Appendix A) that was
constructed for this study is designed to measure inherently subjective phenomena,
namely, the way in which groups perceive themselves to be repressed culturally.
Therefore, this study reconciles the disjointed character of past studies by putting in
line and making consistent the subjective nature of the independent variable (cultural
repression), and the subjective nature of the dependent variable (relative depravation
leading to violent political rebellion).
The use of culture as an indicator for relative deprivation, and thus, violent
secessionism, is grounded in the notion that cultural identity is understood as being a
more accurate predictor of political violence than class identity (Schock, 1996).
Marxist based theories about revolution, which constitute the preponderance of
58


empirical studies (see footnote 3), argue that political conflict and rebellion are the
products of a collective response to economic discrimination. As we have seen,
Gurrs RDT, with its heavy emphasis on economic factors, is not too dissimilar from
Marxist theories of class-conflict. In fact, both neo-Marxist theories and traditional
relative deprivation theories, assert that expanding divisions in income distribution
constitute the single most significant impact on the emergence of violent political
rebellion and civil unrest. These approaches, unfortunately, treat income as
something that is unaffected from other social phenomena. By analyzing how
income, by itself, facilitates political violence, researchers have treated income as
operating in vacuum; unaware of the fact that income inequality is only a symptom of
a greater social pathology.
As such, this study argues that cultural identity is a more accurate predictor of
political behavior than class differentials. This approach is presented because it is
hypothesized that indicators of cultural repression are more directly related to violent
rebellions than economic repression. This is so because the use of economic
repression to test for political rebellion is indirect in the sense that the aggregate
economic differences between the core and periphery are directly related to their
culturally based differences (Schock, 1996, p. 103). Therefore, economic
differences are almost always associated with, or contingent upon, cultural
differences.
59


Consequently, groups are more likely to be repressed culturally, then merely
economically because economic discrimination is dependent upon ones cultural
affiliation (Hechter, 1975; Schock, 1996). Therefore, when studies have traditionally
used economic factors to test for political violence, whey have mistakenly not taken
into account that income disparities typically fall along cultural lines (Hechter, 1975).
To get at the core of the problem, one thus needs to analyze factors of cultural
repression as being the encapsulation of other forms of repression.
Economic repression is just one problem that faces those that are repressed
generally. Those that are repressed culturally are often repressed against
economically as well. This is so because the state creates a cultural division of
labor, whereas occupants within the higher statuses tend to be reserved for members
of the dominate culture, whereas members of minority culturals tend to occupy
positions at the lower end of the stratification hierarchy (Schock, 1996, p. 103). In
other words, cultural repression is a necessary condition for there to be income and
political discrimination, and by merely testing the latter (like so many relative
deprivation studies have done) one is unwittingly not taking into account the former,
which is a necessary condition of the latter to exist. Here, Schock (1996) articulates
that it would be difficult to imagine a situation in which a group is not economically
and/or politically repressed because of some cultural characteristic they posses.
Therefore, cultural repression is the underling factor that best explains, not only other
60


manifestations of repression, but also best explains collective frustrations that stem
from cultural repression and lead to violent political rebellions.
A culturalistic perspective, which this study advances, understands that
repression of a group because of their culturalistic identities, creates extraordinary
amounts of fear and frustrations. These fears and frustrations go beyond those which
are associated with being economically or politically repressed. Being repressed
because of ones culture creates a feeling of demonization and dehumanization that is
unique to this type of repression (Brubarker & Laitin, 1998). Because ones cultural
beliefs such as language, religion, cultural ceremonies, and culturally distinct dress
are all held to a sacred standard (Brubarker & Laitin, 1998) repression of ones
culture should make frustrations and aggression much more likely than economic
repression by itself.
Culture tends to be a constant throughout human conflict (Henderson, 1997).
Therefore, it seems remarkable just how little relative depravation theorists have paid
attention to this issue. Additionally, this neglecting of cultural factors as being
facilitators in violent political rebellions is even more confounding when one
recognizes that many have argued that cultural conflict is the worlds greatest killer
(Kegley & Wittkopf, 1995, p. 498). Even Gurr (1994), the founder of RDT, estimates
that the 26 million refugees that where fleeing the 50 major conflicts that were
occurring in 1993 and 1994, were displaced because of cultural antagonisms. It is
also argued that that violent internal wars and rebellions occur along religious and
61


linguistic lines, which are two inherently culturalistic phenomena (Huntington, 1993;
Henderson, 1997).4
Consequently, this study focuses on cultural repression as a precondition to
collective political violence brought about by collective sentiments of relative
deprivation. Moreover, this approach is conceptually justified in the understanding
that cultural discrimination is more likely to produce frustrations that lead to violent
political rebellions than simplistic analyses grounded in income inequalities. The
hypothesis that cultural repression is at the center of secessionist movements
transcends the problematic methodological and conceptual underpinnings of precious
studies that have utilized RDT to test this phenomenon.
4 The importance of what role religion and language play in political violence cannot be overstated.
This is why they have been included into the cultural capital index of this study (see Appendix A).
62


CHAPTER 3
METHODOLOGY
To assess the effect that cultural repression has on violent secessionist
conflict, this study investigates how the collection of theoretically justified indictors
of cultural repression work to explain violent political rebellions. Thus, it is
hypothesized that the repression of ones cultural existence will have a positive
relationship with the likelihood of there being violent movements of secessionist
conflicts. In other words, as the degree, or the extent to which, a group experiences
cultural repression increases, so too, will the occurrences of violent political
rebellions.
On each dimension of cultural repression, as the score increases, so too, does
the intensity of relative deprivation. Therefore, each measure on the ordinal scale of
relative deprivation, there should exist a positive relationship to the occurrence and
severity of rebellion. The theoretical justification for this is premised on the notion
that repression of ones group will lead to frustrations caused by what the group has
and what the group believes it should have relative to those around it. Therefore, as
cultural repression becomes more severe, the likelihood that the repressed group will
agrees against the object of its frustration will increase as well.
63


Data
The statistical models constructed for this study were drawn from Phase IV of
Gurr, Scott, & Davenports (2003) Minorities at Risk Data Generation and
Management Program (MARGene). This is an independent, university-based
research project that monitors and analyzes the status and conflicts of politically-
active communal groups in all countries in the world with a current population of at
least 500,000 (Gurr et al. 1993). MARGene is supported by the Center of
International Development and Conflict Management (CIDCM) and the University of
Maryland with funding from the National Science Foundation, the United States
Institute of Peace, the Hewlett Foundation, the Carnegie Corporation, and the State
Failure Task Force.
MARGene was initially created by Ted Robert Gurr in 1986 and has been
based at the Center for International Development and Conflict Management,
University of Maryland since 1988. In 2000, MARGene Directorship was given to
Christian Davenport. To help provide guidance and assistance on the selection
process of groups, construction and application of new indicators, and general data
quality control issues, a national advisory board was established for the project in
1999, and currently the creation of an international advisory board is being
formulated
The MARGene cross-national dataset had been developed over the four
distinct phases. Phase I covered 227 communal groups which met the criteria for
64


classification as a minority at risk for years 1945-1989. Phase II covered 275 groups
from 1990-1995, Phase III covered 275 groups from 1996-1998, and Phase IV
covered 285 groups from 1998-2000. This was a cumulative process, and as such,
over time some groups were added to the dataset and some groups were dropped as
groups either lost or gained their at risk status (Davenport et al., 2003). Data were
collected about at risk groups at more or less regular intervals. Significant effort
was given to collect data on an annual basis, however, at times, data were collected in
five-year intervals.
Davenport et al., (2003) asserts that in an initial analysis of the datasets inter-
coder reliability and internal consistency, informal results indicate that coding
reliability and validity is satisfactory. No actual coefficients were given however,
Davenport et al., (2003) suggests that some revisions to coding rules and procedures
have been made, but generally these concerns and adjustments have been minor.
Sample
A collection of qualitative and quantitative information concerning all
communal groups which meet the criteria for inclusion as a minority at risk have
been included into this study. The at risk classification refers to an ethnopolitical
group, or non-state communal group, that, collectively suffers, or benefits from,
systematic discriminatory treatment vis-a-vis other groups in a society and
collectively mobilizes in defense or promotion of its self-defined interests (Gurr,
65


1993, p. 3). Additionally, the groups must have had at least 100,000 members or
have constituted at least 1% of the population of the country in which they reside and
must reside within a country with at least a total population of 500,000.
The Phase IV MARGene dataset, which is being used for this study, covers
285 at risk groups. Gurr (1993) defines groups as psychological comminutes:
groups whose core members share a distinctive and enduring collective identity based
on cultural traits and lifeways that matter to them and to others with whom they
interact (p.3). The Phase IV database includes all of the conflict, context,
intervening, and response variables from the previous datasets (Phases I through III),
and it includes an expanded set of communal groups (Davenport et al., 2003). This
increased number of communal groups covered in the MARGene dataset, and thus
used in this study, was the result of both the discovery of a few previous
unidentified communal groups, and the emergence of new politically active
minority groups (Davenport, et al., 2003).
The relative dimension of this dataset, with respect to its theoretical
justifications, are such that group discrimination and bias are only included into the
dataset as they exist relative to other groups within the country. This is of paramount
importance to this study because the relative deprivation theory that is being tested
posits that deprivation operates as a function to what people have and what they
believe they should have relative to the immediate social reality in which they live.
For example, the denial of allowing one specific group the right to practice its
66


religion is relevant only if other groups are allowed to practice theirs. If no group is
granted the right to practice its religion then there exists no immediate bases for
groups to gauge their deprivation. Therefore, the focus of the indicators that are used
in this study, and are drawn for the MARGene dataset, measure the relative status and
condition of the group, not the groups absolute condition (Davenport, et al., 2003).
Moreover, the identity of the group must be in some way relevant to its
discrimination or disadvantage (Davenport et al., 2003). Additionally, if the group is
mobilized, the mobilization must center on groups issues vis-a-vis other groups or
the state.. .mobilization as part of some non-ideological campaign, doses not qualify
as group mobilization in the context of this project (Davenport et al., 2003, p. 7).
Again, the importance of having the relative dimension built-into the coding
and the selection of discrimination indicators cannot be overstated. Without this,
measurements would be absolute values, therefore, making conclusions and
judgements deduced from the theoretical framework of relative deprivation all but
meaningless. Studies that have not used the MARGene dataset to test relative
deprivation theory have been forced to create completely separate indicators to
measure the relative dimension of the theory (see Canache, 1996; Midlarsky, 1988).
Doing this confounds and complicates the methods; yielding questionable results and
producing concerns about the internal validity of the study.
67


Variables and Indicators
After careful review of both the variables given in the dataset, and the
literature pertaining to the manifestations of cultural repression, variables were
selected based on their validity in being conceptually accurate indicators of the
phenomenon. The dependent variable was also selected due to its high validity in
measuring violent forms of political rebellion, and that it has been a widely used and
recognized as being a valid and accurate measure of political violence (see Dudley &
Miller, 1998; Harff & Gurr, 1998; Gurr & Moore, 1997; Fox, 2002).
The dependent variable, anti-regime rebellion index (REBEL), was selected
because it is a direct measurement of violent political rebellion, this has high
construct validity (Davenport, et al., 2003). Because it is hypothesized that relative
cultural deprivations produce frustrations that lead to violent mass movements of
secessionist conflict, this variable is a good theoretical fit to RDT. The violent aspect
of this study vis-a-vis the use of cultural deprivations to test for political rebellions is
in line with how the dependent variable is coded because only the most serious
manifestations of rebellion is coded for each of the five-year periods (Davenport, et
al., 2003, p. 90).
The use of the anti-regime conflict rebellion index is valid because it takes
into account the amount of people involved in various types of violent rebellions.
When one is attempting to measure the degree of rebellion there must be a threshold
68


of participation in place (Weede & Muller, 1998). Nevertheless, when one is
constructing indicators of rebellion based on the type of violence and number of
participants, one cannot completely avoid arbitrariness (Weede, & Muller 1998).
Davenport et al., (2003) in the Minorities at Risk project has, in addition to numerous
other criteria, set the number at 1000 to indicate the upper level of rebellious
participation. According to the anti-regime rebellion index scale, a score of 6, for
example, large-scale guerrilla activity, has the criterion of requiring more than 1000
armed fighters (See Appendix A). This number typically squares with past research
on this subject such as Singer & Smalls (1982) correlates of war project. However,
there tends to be no theoretical reason why 1000 is a better number than either 897
or 1039- or, of course, [better than] any other number in between, say 500 and 1500
(Weede & Muller, 1998).
The values coded for the REBEL variable are based on the highest observed
level of rebellion in the given period (Davenport, et al., 2003). For example,
rioting, although a violent act, is operationalized in the dataset as being a
spontaneous escalation of an otherwise non-violent protest, and is coded, therefore, as
a form of protest and not a form of anti-regime rebellion (Davenport, et al., 2003).
This is because protest is seen as lacking the necessary premeditation and intent that
is associated with acts of violent rebellion (Davenport, et al., 2003).
The quinquennial rebellion scores used cover the time period 1945-1999.
These scores are based on actions initiated by members of the group on behalf of the
69


groups interests and directed against those who claim to exercise authority over the
group (Davenport et al., 2003, p. 89).
The anti-regime rebellion index is an ordinal level scale that measures
rebellion from 0 (none reported) to 7 (protracted civil war) (see Appendix A for
complete scale). This variable also exhibits good distribution, whereas issues of
skewedness are not of major concern (see Appendix B for distribution table). Even
though there are some missing values for this variable, the very high N value of 4121
reported cases still yielded high overall levels of reporting (see Appendix B for
descriptive statistics).
The independent variables are a collection of carefully selected indicators of
cultural repression. These variables were not selected haphazardly, but rather
premised on past studies acceptance of them as being good indicators of profound
manifestations of cultural repression5. In other words, the literature on this subject
has shown that particular forms of cultural repression are more aggression-causing
then others (Brubaker & Laitin, 1998). For example, a group would tend to feel
higher levels of frustrations if that group where prevented from practicing their
culturally-specific religion than if that group were prevented from attaining some of
its culturally-specific food. Although food is an important aspect to many cultures, it
5 For the importance of autonomy issues see Dudley & Miller, 1998; Harff & Gurr, 1998; Weede &
Muller; Gurr &Moore, 1997; Williams, 1994; Collins, 1993. For religious issues see Fox, 2002;
Henderson, 1997. For language issues see Henderson, 1997; Horowitz, 1981; Lake & Rothchild, 1996;
Johnston, 1995. For forced cultural resettlement/segregation and movement issues see Brubaker Laitin,
1998; Williams, 1994. For ritual and custom issues see Brubaker & Laitin, 1998.
70


tends to not be as important as religion, language, customs, etc (Brubaker & Laitin,
1998).
A cultural capital index was created consisting of variables of cultural
repression. This index contains cultural variables measuring loss of cultural
autonomy (AUTOLOST), discrimination against language (CULDIFX2),
governmental policy restrictions on religion (CULPOl), ceremonies (CULP04), and
appearance (CULP05) (see Appendix A for variable definitions and scaling; see
Appendix B for descriptive statistics of each variable). These variables were
collapsed into one index for the purposes of easy statistical management within the
regression model (see Appendix B). Additionally, all of the independent variables
have little of their variability explained by the other independent variables in the
model, thus demonstrating that multicollinearity is not a problem (see Collinearity
Diagnostics in Appendix B).
Additionally, two general repression variables were added to test the effects of
cultural repression on the occurrence of anti-regime conflict. These are general
observations of repressive policies targeted at group members (see Appendix A for
coding; see Appendix B for descriptive statistics). The restrictions on movement
(REP11) and forced resettlement (REP 12) variables where not included into the
cultural capital index because they are not a measurement of a specific manifestation
71


of state-sponsored repression directed at a cultural artifact per se. Insofar as the
variables included into the cultural capital index are manifestations of repression
directed at specific elements of ones culture (i.e. religion, dress, language, rituals),
the variables listed under the government repression of group heading are forms of
repression targeted at a culturally distinct group in general. These are polices of
repression directed at the totality of the group, and not just at one aspect of the
groups culture.
To analyze the effects of political and economic repression, one political
repression variable (POLDIFX1) was added into the model along with one economic
variable (ECDIFX1). POLDIFX1 is a political inequality variable that measures
group access to political power, and ECDIFX1 is economic inequality variable that
measures income inequality. These variables were selected because of them being
non-specific in nature, and are therefore, good overall indicators of both political and
economic inequalities (see Appendix A for coding; see Appendix B for descriptive
statistics). These measurements are also meant to reflect the general socioeconomic
and sociopolitical landscape of a given nation. Because it is hypothesized that
cultural inequalities are better indicators of violent anti-regime conflict then economic
and political inequalities, I should expect to find that these variables have little overall
effect on the dependent variable. Additionally, they were added into the model
because view studies in the past have look at how three general social structures
(culture, politics, and economics) work to explain anti-regime conflict.
72


The dummy variable of RACESUB, was used to measure what possible effect
race might have on the dependent variable. This variable was recoded from the
original RACE variable into one dichotomous variable. RACESUB is coded as 1=
different racial stock from the dominate group with substantial intermixture, and 0=
all others. This variable was added to test how significant racial relationships are in
either facilitating or deterring secessionist conflict. In line with past studies, (see
Rummel, 1997; Horowitz, 1981; Saideman, 1997) it is expected that groups with have
significant intermixing with the dominate group will be less likely to participate in
violent political rebellions.
It must be noted, however, that there are some issues concerning the types of
missing data. An important distinction must be made between the lack of
applicability for an indicator versus the lack of a basis for coding an indicator. The
method employed in the collection of data for this dataset holds that in cases where
the variable is not applicable- for example, codings of political restrictions of
Russians in the Newly Independent States prior to the dissolution of the Soviet Union
are not applicable (Davenport, et al., 2003, p. 7)- the variable is left blank. In cases
where observation is possible, but where there exists no basis upon which to make
judgements due to the lack of reliable information, a missing code was used. The
missing information was coded as -99, and was applied any time the coder believed
that there was an inadequate amount of information for a proper judgment to be
formulated about the indicator (Davenport, 2003).
73


Method of Analysis
It is hypothesized that the repression of ones cultural existence will be a
better predictor of violent secessionist conflict than indicators that have been used in
the past, such as sociopolitical and socioeconomic indicators. As such, this study is a
systematic analysis of how group deprivations brought about by differentially applied
governmental policies against cultural identifiers and general communal existence
lead to anti-regime rebellion. Nevertheless, unlike past studies, this analysis is an
examination of all three social structures taken together (cultural, political, and
economic).
In my basic model I examine the relationship between cultural, political,
economic and racial phenomena. This was done by the use of such variables as loss
of cultural autonomy, governmental restrictions placed on culturally specific
religions, dress, languages, and ceremonies, and general governmental policies
restricting group movement, and/or state-sponsored forced resettlements. The
inclusion of a political inequality variable (governmental policies restricting access to
political power) and an economic inequality variable (governmental policies of
discriminatory practices toward income distribution) were both added into the model
to control for their effects.
If my hypothesis is supported, I should expect to find the economic and
political variables to have little overall impact on the dependent variable.
Additionally, being that race is of often considered an important aspect of social
74


interaction, a race variable was also added into the model to test its possible effects on
the dependent variable.
By running a multiple regression analysis, I am able to use two general
cultural repression variables and one cultural index to analyze the strength of the
relationships between these variables and the dependent variable of anti-regime
conflict. Additionally, by utilizing a multiple regression analysis, I am able to
determine the direction of the relationships exhibited by the independent variables on
the dependent variable. According to my hypothesis, I should expect to find that the
cultural repression variables and the cultural capital index have a positive relationship
to the dependent variable. Therefore, as repression on the cultural variables
increases, so too, should the severity of violence on the anti-regime rebellion index.
Using this model I am also able to assess what possible relationship might exist
between the political, economic, and racial variables and the dependent variable.
Limitations
The first methodological limitation that needs to be discussed concerns the
nature of the data that are being used. In the area of political violence there has
traditionally been a heavy reliance on analyzing secessionist conflict via case-based
patterning (Brubaker & Laitin, 1998). This means that researchers have sought to
extract patterns of social behavior from specific regions or countries. These case
studies include, for example, areas like Northern Ireland, Yugoslavia, Sri Lanka,
75


Rwanda, and Spain just to name a few. In these cases of secessionist violence, a
country is treated as a case when the violence is treaded as a set of separate instances.
This, therefore, allows for controlled comparisons to other areas. Some argue that
cross-national data (like the type used in this study) are unable to make the same
comparisons from one country to the next (Alker, 1994; Brubaker & Laitin, 1998).
Moreover, some have expressed doubts has to whether general models of
political violence can be constructed from cross-national datasets (Alker, 1994;
Canache, 1996). The thought is that case-studies and micro-analyses maintain greater
explanatory potential for yielding more specific understandings of political violence.
It is believed (Alker, 1994; Canache, 1996) that large cross-national datasets are
simply unable to measure the micro-level processes of particular social-psychological
phenomena such as frustration, aggression, and violence. As such, cross-national
datasets require one to make a leap in logic, whereby, equating general social
conditions with individual-level psychological states. Additionally, cross-national
studies tend to be unable to provide in-depth analysis that is often associated with a
detailed examination of a particular countrys history, demographics, sociopolitical
structure, etc (Schock, 1996).
However, while case-studies may be good in producing rich analyses of a
particular countrys sociopolitical structure, they are nevertheless limited in their
ability to formulate general causal patterns because of their specific and narrow focus.
If, for example, one particular nation, or region, or group is the sample of a certain
76


study, it is difficult to generalize from that sample to other such cases in other areas;
one is uncertain how well that sample represents other groups throughout the world.
The examination of a large number of nations allows one to extrapolate general
patters that might not otherwise be revealed if one where to analyze only a few cases
(Schock, 1996). Moreover cross-national studies are probabilistic rather than
deterministic and permit the analysis of interaction effects and multiple causes.. .thus
cross-national research is conducive to determining whether or not probabilistic
theoretical propositions can be supported (Schock, 1996, p. 110).
Because the purpose of this study is to examine the existence of general casual
patterns as they relate to cultural deprivation and secessionist conflict, micro-level
approaches in methodology for this particular analysis are inappropriate, unnecessary,
and too constrictive. Cross-national data allow the researcher to analyze general
patterns and variations that are exhibited throughout nations; yielding explanations
that describe the overall character of the phenomenon. Critics of this approach claim
that because one is measuring an individual-level mechanism (in this case frustration
and aggression caused by cultural discrimination), cross-national data sets and macro-
level analyses are ill equipped to measure and test individual specific behaviors
(Alker, 1994; Canache, 1996). However, macro-level analysis has a distinct
advantage over micro-level analysis in that the former allows the researcher to
explicitly take the role of certain social structures into account (Muller & Weede,
1990).
77


Although it is generally recognized that testing theories of rebellious conflict
cannot necessarily be achieved by using countries as the unit of analysis, indicators
pertaining to the national variation of the magnitude of political violence can be
evaluated as to whether it is consistent or not with hypotheses derived from causal
mechanisms assumed to operate at the level of individual behavior (Muller &
Weede, p. 625,1990). Therefore, the empirical modeling of causal relationships
between structural mechanisms, governmental behavior, and collective secessionist
motivations demands the utilization of cross-national comparisons (Muller & Weede,
1990).
Other Limitations
Although the Minorities at Risk dataset is rich with information, its format at
time makes quantitative analyses difficult. Most of the problems revolve around the
time in which information was gathered about at-risk groups. Although this
particular study is not an examination of trends that exist over time per se, the way in
which the variables were coded and gathered makes the analysis of the variables
somewhat problematic. Because some of the variables themselves have a temporal
dimension built-into them, it makes merging typical pooled-cross-national or time-
series datasets difficult. This is due to the fact that the dataset has no explicit time
variable by itself.
78


Another problem with the data that more directly effects this study is that the
MARGene data have tended to be somewhat inconsistent in terms of the frequency of
data gathering on different variables over time (Davenport et al., 2003). Although
significant effort was given to be as consistent as possible, some data have been
collected at five-year intervals, while other data were collected annually. In most
cases, because the nature of the variables are such that they tend to change slowly,
data needed to be collected at wider intervals then annually or biennially
(Davenport, et al., 2003).
Because observations are not uniform, sampling patterns tend to cause
difficulties in analysis. For example, a five-year interval in which a group faced
certain conditions should count as five times more important than a one year interval
of some other condition (Davenport et al., 2003). However, this problem is overcome
because the MARGene software allows users to create interpolations of the data
within these larger intervals automatically, leading to a uniform and standard
annualized dataset appropriate for statistical analysis (Davenport, et al., 2003, p. 2).
Interpolation was used in this study when extracting the data from the software into
SPSS.
79


CHAPTER 4
DATA ANALYSIS
The first objective of this analysis was to create a multiple regression model to
examine the strengths of the relationships that exist between the independent
variables and the dependent variable. Additionally, the veracity of my hypothesis
was contingent upon these relationships being positive. My hypothesis could not be
supported if it is found that as cultural repression increases, anti-regime rebellion
decreases. Therefore, the direction of these relationships, especially with regards to
the cultural variables, is a very important aspect to my model.
The next objective was to include into the model political and economic
variables to analyze what effects they might have on the dependent variable. This
was done because past studies in this field have claimed that the political and
economic characteristics of a particular nation play a significant role in the dynamics
of collective rebellion. If this is truly the case, one should expect to find these
variables to have a significant impact on the dependent variable. However, I
hypothesized that one should expect to find indicators of cultural repression to have a
greater effect on anti-regime rebellion then economic or political indicators of state-
sponsored repression.
One can see that, according to table 4.1 below, the standardized beta
coefficients indicate that all of the cultural variables (CULCAP, REP11, and REP 12)
80


are significant in influencing anti-regime rebellion. Additionally, it is also important
to point out that with regard to the cultural variables, all of them exhibit a positive
relationship to anti-regime rebellion. Therefore, one can infer from this that as
repression of ones culture increases, so too, does the likelihood of there being violent
secessionist conflict.
Table 4.1 Regression Analysis for All Independent Variables
Independent Variables Standardized Coefficients P Significance Levels
Cultural Capital Index .094 .001
Restrictions on Movement .236 .000
Forced Resettlement .177 .000
Access to Political Power .054 .059
Income Differential .021 .459
Different Racial Group w/ Substantial Intermixture -.089 .001
Dependent Variable: Annual rebellion index
This model provides support to my hypothesis whereas it was asserted that the
repression of ones culture facilitates collective frustrations that inexorably lead to
anti-regime conflict beyond other manifestations of state-sponsored repression.
Because the economic and political variables yielded low beta coefficients, one can
81


further argue that repressions targeting a groups economic and/or political wellbeing
are insignificant predictors of anti-regime conflict.
Additionally, according to the model in table 4.1, the restrictions on
movement variable (REP11) exhibits the strongest effect on the dependent variable
with a beta coefficient at .236. The R-squared in table 4.2 below, illustrates that the
cultural variables taken together are significant in accounting for the variability in
anti-regime conflict.
Table 4.2 R-squared Model
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .429(a) .184 .179 1.54474
a Predictors: (Constant), RACESUB, Forced resettlement, Access to power, RACEDIF,
CULCAP, Income, Restrictions on movement
When the political and economic variables were added into the model, their
overall impact in explaining the variability of the dependent variable was negligible.
Moreover, the beta coefficients for access to political power and income
inequality at .054 and .021 respectively, suggests a weak influence on the dependent
variable. This furthers my contention that economic and/or political indicators of
state-sponsored repression are not as strong of predictors of secessionist conflict as
past studies would have one believe.
According to the model in table 4.1 above, the economic variable has no
significant influence at .459 and the political variable is only slightly more significant
82


at .133. However, with a small beta coefficient of .054, the political variables effect
on the anti-regime is still insignificant. This, therefore, indicates that a groups
political and/or economic reality does not exert a statistically significant effect on
anti-regime violence.
It should be kept in mind that the validity of the results are contingent upon
the relative perceptive of the groups included into the sample. Because the variables
are gathered and coded as indicators of ones relative sociological situation,
inferences cannot be made about how the absolute conditions of repressed societies
influence the likelihood for there to be secessionist conflict. Nevertheless, because
this study is a test of relative depravations, inferences can be made about how relative
situations of repression facilitate secessionist conflict.
The race variable was recoded into a dichotomous variable which yielded a
beta coefficient of -.089 for different racial group with substantial intermixture.
This beta coefficient provides support for the idea that groups that are racially similar
to the dominant group(s) are less likely to participate in anti-regime violence against
that group(s). In other words, substantial racial intermixing will the dominant group
will likely reduce the likelihood for secessionist conflict. This supports past research
that in the field of how racial relations operate to either reduce or facilitate political
violence (see Rummel, 1997; Horowitz, 1981; Saideman, 1997).
83


CHAPTER 5
CONCLUSION
Since the time of Aristotle, philosophers and social scientists alike have
attempted to formulate theories to explain the rise and onset of political violence.
From Marx to de Tocqueville, and from Gurr to Tilly, researchers have constructed a
wide array of hypotheses that have sought to describe and elucidate the dynamical
processes involved in anti-regime violence. These theories have argued that this type
of violence is the product of anything ranging from rational choice decisions, to a
collective awaking of class-consciousness, to unified efforts of resource mobilization,
and to collective accumulations of psychological frustrations.
As a contemporary issue for social scientists, there has been a multitude of
empirical studies conducted that have attempted to analyze the existence of causal
patterns in collective behaviors that might lead to secessionist conflict. Most of these
studies have been grounded in a wide verity of theoretical paradigms. Some of the
more accepted of these within the scientific community include, rational choice
theory, game theory, state power theory, resource mobilization theory, and relative
deprivation theory. In this study, I have made an effort to critique each theory
separately, contextualize them against one another, and analyze relative deprivation
theory independently.
84


Relative deprivation theory was used in this study in what can be considered a
new and unique way. Rather than utilize economic indicators, such as income
inequality, to investigate collective relative deprivations, which had been the standard
in past studies, cultural phenomena were examined. As has been the case with so
many relative deprivationist studies, economic and political forms of repression have
been evaluated both separately from one another, and from the greater social-
structural framework. This study transcends these shortcomings by investigating, not
only the effects caused by manifestations of economic and political repression, but
more importantly, by including forms of cultural repression into the analysis.
It is important that one be cognizant of just how important cultural factors are
and why they cannot be divorced from political and economic structures like has
tended to be the case in past research, I contend that the effects that are demonstrated
in the statistical results indicate that culture has a profound underlining influence on
group behavior. I assess the results of this study as illustrating the significant role
that culture plays in predicting anti-regime violence.
However, an argument can be made that, if taken alone, political and/or
economic variables may produce the same results as the cultural indicators. This may
be the case, but one has to keep in mind that culture, itself, has not been tested in this
manner in past research. Indicators of cultural repression have generally not been
analyzed as being facilitators of relative deprivations. Moreover, it is my contention
that economic and political manifestations of state-sponsored repression usually come
85


down on the cultural characteristics of a particular group. Because political and
economic forms of repression do not exist by themselves, but are interconnected to a
deeper cultural meaning, not taking into account the cultural context of repression,
relative deprivations, and secessionist conflict is a misrepresentation of how
secessionist conflict comes into being.
Past studies that have examined economic and political facilitators of relative
deprivation, have generally used indirect indicators to understand the phenomenon of
anti-regime conflict. Economic inequality and political discrimination are merely
symptoms of a greater illness of cultural repression. Therefore, by using forms of
cultural repression as indicators of anti-regime conflict, one is more able to get at the
heart of the issue. As stated previously, it would be hard to imagine a situation in
which groups of people are discriminated against because of them merely being poor,
but not because of any cultural traits they possess like the religion they practice, or
the culturally distinct clothes they ware, or the language they speak, etc.
Looking only at income indicators of relative deprivation, one must be moved
to ask, why is that group poor in the first place? The answer cannot merely be,
because they are poor; this leads to the problem if infinite regress. There has to be a
deeper issue at play. If one is hoping to study income inequality as being the
facilitator of relative deprivation, one must ask oneself if income is truly at the root of
relative deprivation.
86


Moreover, one must keep in mind that the variables used in this study are
indicators of a groups relative reality. Therefore, in the terms of this study, it would
be illogical for one to postulate that the reason the group is poor is because every
group in that particular society is poor and uniformly subjected to the same unjust
economic polices of a despotic or totalitarian regime. The variables used in this study
only include explicit situations in which a group in a particular nation is repressed
while other groups in that same nation are not. Therefore, is would be unreasonable
to assert that groups within a particular nation are economically or politically
repressed simply because all groups within that nation are repressed. Consequently,
because of the relative nature of the variables, the need for ascertaining the deeper
and more profound issues of repression and secessionist conflict is dealt with in this
study.
Merely testing for relative deprivations via the use of policies of economic
and/or political repression is an attempt to get at and understand anti-regime violence
in an indirect manner. Schock (1996) described this situation succinctly when
discussing how economic repression comes into to being, the aggregate economic
differences between the core and periphery are directly related to their culturally
based differences (p. 103). As such, economic and political forms of state-
sponsored repression are deeply interconnected with, and indivisible from, deeper
issues of culturally based manifestations of repression. By divorcing economic
and/or political factors from their culturalistic underpinnings, past studies of relative
87


deprivation and political violence have ill-advisedly strayed away from the essential
nature of the problem.
Cultural repression is a precondition for there to be other forms of repression,
such as political and economic. Generally speaking, when regimes create policies of
repression, those polices often come down on cultural lines. When a government
implicitly or explicitly designs policy that represses one group over another, they
usually do so because of the cultural differences that particular group embodies. This
is because the dominate group creates a cultural division of labor, whereas occupants
within the higher statuses tend to be reserved for members of the dominate culture,
whereas members of minority cultures tend to occupy positions at the lower end of
the stratification hierarchy (Schock, 1996, p. 103).
Violent political conflict has manifested itself since the dawn of civilization.
Responsible for the deaths of countless human lives, sociopolitical discord and
discontent have wreaked havoc on humankind, and has become a problem that has
only escalated with the passage of time (Eckhardt, 1992; Westing, 1982; Waller,
2002). The dawn of a new century demands that we construct new approaches and
new paradigms in a hope to understand and deal with this problem more effectively.
I contend that this study is the first step in a new direction of comprehension and
analysis. Looking at attributes of culture in their totality, and how their repression
might lead to secessionist violence, more aptly allows us to analyze the essential
social structures that drive repression, discontent, and collective violence.
88


Advancing the idea that culture tends to be the single transcendent force
behind many forms of human conflict, Henderson (1997) articulates a premise that
has seemingly gone unnoticed by most relative deprivation theorists. Although there
are limits to, not only this study specifically, but to relative deprivation theory in
general, taking into account of how cultural relationships interact within the overall
dynamics of collective frustrations and anti-regime violence, will help pave new paths
into this complicated and often misunderstood phenomenon of secessionist violence.
89


Appendix A
Variable Definitions and Coding
Dependent Variable: Quinquennial rebellion scores, 1945-1999
Only the most serious manifestation of rebellion is coded for each of the five-year
periods.
REBEL Rebellion Index: 1945-99
Missing Values: -99
Value Label
0 None reported
1 Political banditry, sporadic terrorism
2 Campaigns of terrorism
3 Local rebellions
Armed attempts to seize power in a locale.
If they prove to be the opening round in
what becomes a protracted guerrilla or
civil war during the year being coded, code
the latter rather than local rebellion. Code
declarations of independence by a minority-
controlled government here.
4 Small-scale guerrilla activity
All of the following must exist:
1) fewer than 1000 armed fighters;
2) sporadic armed attacks (less than
six reported per year); and
3) attacks in a small part of the area
occupied by the group, or in one
or two other locales.
5 Intermediate guerrilla activity
Has one or two of the defining traits
of large-scale activity and one or
two of the defining traits of
small-scale activity.
6 Large-scale guerrilla activity
All of the following must exist:
90