Print Email Download Reference This Send to Kindle Reddit This
submit to reddit

Alliances and the clash of civilizations

Abstract

The current thesis aims at testing the Clash of Civilizations hypothesis through the prism of alliance formation and stability theories. Although Huntington (1993, 1996) is often disputed by other scholars, it is important to consider his ideas because he emphasizes the significance of culture in international relations which is also supported by the constructivist school of thought. In addition, by focusing on alliances, the thesis escapes from the common research on the relation between conflict proliferation and different civilizations; hence aiming at opening a new niche of research. The results of the analysis favour the two hypotheses that same-civilizational alliances are more likely to form and that they are more stable, i.e. more durable. Therefore, Huntington is right when he says that culture matters in the international system and influences political decisions.

Introduction

After the end of the Cold War many scholars have tried to predict how states will interact with each other. More specifically, political scientists have been engaged in discussions about what will provoke or prevent conflicts. According to Francis Fukuyama (1992), the world has reached the end of history: democracy and market-economy have proven themselves, and Western liberalism has brought cooperation and peace among nations. Another perspective on the new order, promoted by Benjamin Barber (1995), is that despite economic globalization, states are still culturally different and have conflicts. However, these conflicts will not be of global scale but will remain within community boundaries. Perhaps the most extreme position on the world order after the end of the Cold War is taken by Samuel Huntington who developed the theory of the Clash of Civilizations. When his article "The Clash of Civilizations?" appeared in Foreign Affairs in 1993, it spurred a very controversial debate within the academic and political world; the situation was intensified after the publication of his book "The Clash of Civilizations and the Remaking of the World Order" in 1996. His main argument is that the conflicts of the future will be based on cultural differences, especially religious ones, and thus will be longer, bloodier, and more destructive. Events such as 9/11, the wars in Iraq and Afghanistan, the bombings of Madrid, London, and Bali, and the rising tension between the Western and Muslim nations seem to prove his thesis. Therefore, the world might indeed be about to experience a very conflictual and gloomy future.

After its introduction, the Clash of Civilizations has provoked numerous scholars to work on it and to test its implications. Despite its popularity, the theory has spurred predominantly negative reactions from within academia, the political sphere and the international community. One of the most influential critiques comes from Bruce Russett, John Oneal, and Michaelene Cox who, in their publication "Clash of Civilizations, or Realism and Liberalism Déjà Vu? Some Evidence" (2000), test Huntington's propositions by accounting for militarized interstate disputes (MIDs) in the period 1950-1992. Their findings demostrate that other factors than civilizations better help understand military conflicts. Giacomo Chiozza's research follows a similar line of thought in his article "Is There a Clash of Civilizations? Evidence from Patterns of International Conflict Involvement, 1946-97" (2002), which tests for additional variables which again turn out to better explain conflicts and their proliferation. Andrej Tusicisny has a particular focus on the end of the Cold War as a division point in the trend of conflict proliferation. In his article "Civilizational Conflicts: More Frequent, Longer, and Bloodier?" (2004), he fails to find a relationship between civilizational difference, the end of the Cold War, and the duration of a conflict. Going beyond the typical civilizations-conflicts paradigm are Erik Gartzke and Kristian Gleditsch who, in "Identity and Conflict: Ties that Bind and Differences that Divide" (2006), differentiate between religion, ethnicity, and language as factors in cultural conflicts and take into account secondary groups constituting part of a state's population.

As seen from the brief overview, the major focus of the research so far has been on conflicts and their relation to civilizations. The purpose of the thesis, however, is to incorporate theories about alliance formation and duration into the Clash of Civilizations hypothesis. In this way, the emphasis is shifted from conflictual to cooperative patterns thus better accounting for the development of the international system. Since military alliances demonstrate the real commitment of states (Leeds, 2003), it is worth analysing whether cultural differences play a role in the formation and consequences of alliances.

Furthermore, the majority of previous results are based on data which covers only the first few years after the end of the Cold War. To increase their sample sizes, most researchers have incorporated data from the pre-1989 years. However, as Huntington (2000) argues, such research does not test the central ideas of his hypothesis which deals with the post-Cold War era. On the other hand, having in mind the scarce data on conflicts during the 1990s, it is understandable that scholars have tested the Clash of Civilizations hypothesis on a longer timeframe. Since each year of additional data turns out to be so crucial, the current thesis uses the most recently available data in order to test Huntington's hypothesis in a more consistent manner.

Therefore, the current research question is the following: Is there a relationship between alliances and the Clash of Civilization hypothesis?. Thus, by including the formation of alliances and by expanding the timeframe of data, the thesis better analyzes Huntington's theory and better accounts for states' interaction in the international system.

The thesis starts with an introduction describing the problem posed by the Clash of Civilizations theory; discusses the importance of the issue; presents the aim of the study; and closes with the research question. The next part continues with a presentation of Samuel Huntington's theory, its meaning in terms of traditional international relations theories, and its major critiques. In addition, the most notable theories about the formation and stability of alliances are discussed. The third section of the thesis develops a working definition of an alliance and introduces the two hypotheses with their theoretical justification. The following part describes the variables and data used for the purpose of the study. It is followed by a presentation and discussion of the results of the empirical tests. The thesis closes with a conclusion focusing on the most important outcomes of the study.

Samuel Huntington and the Clash of Civilizations Hypothesis

Main Arguments

Samuel Huntington claims that after the end of the Cold War "[t]he great divisions among humankind and the dominating source of conflict will be cultural" (1993, p.22). These divisions are based not only on differences in language, religion, traditions, institutions, and history but also on the subjective relation of a person to a certain group of people (ibid). Therefore, they are fundamental and profound in nature and have a great impact on the interaction between people from different cultures. Moreover, Huntington introduces the concept of civilization as the "highest cultural grouping of people and the broadest level of cultural identity people have short of that which distinguishes humans from other species" (1993, p. 24). From that definition, there are two major implications for the development of the international system after the end of the Cold War: first, there will be more conflicts on civilizational basis, and, second, loyalty to states diminishes.

Huntington briefly describes the various stages of international conflicts after the Treaty of Westphalia: among princes, between nations, and of ideologies, but notes that these clashes concerned primarily Western societies. On the other hand, the latest development of conflicts encompasses the entire world thus bringing local unrest to the global level. He does not say that civilizations never mattered before but rather that their influence was suffocated by the conflicts of the past. Now, however, there is no other force that can diminish the influence of civilizations and, therefore, civilizational differences turn out to be significant explanatory variables of conflicts (ibid).

Huntington distinguishes between six major reasons why civilizations increasingly matter and why they foster conflicts. First, civilizational differences are basic and cannot be easily eradicated. Hence, the consequent conflicts are prolonged and especially violent. Second, current technological progress allows for greater interaction between people from all over the world. Such interaction increases cultural consciousness because people seek to identify themselves even stronger in order to stress their difference from the rest. A third reason is the process of economic development which separates people from their origins. Again, cultural identification is pursued in order to preserve one's identity. Furthermore, the West plays a dual role by trying to impose its own values on other civilizations while at the same time in fact provokes de-Westernization.. Huntington also observes that cultural differences are less easy to resolve than political and economic ones and he postulates that such irrevocable disparities feed constant confrontation. A final reason for the increased importance of culture is the intensified economic regionalism which demonstrates the influence of cultural similarities even on political and economic relations. All these factors, according to Huntington, place one civilization in opposition to another, "us" vs. "them", and as a result tension between civilizations increases. Domestic policies and state positions on global issues are inevitably affected and therefore local confrontations become global. For Huntington, states from different civilizations compete with each other through traditional means but the ultimate goal now is the security and survival of entire civilizations. He also writes that states from one and the same civilization tend to support each other and he calls this phenomenon "kin-country" syndrome. Nevertheless, Huntington does not exclude the existence of conflicts within a civilization but they tend to be less intense and devastating than the ones between civilizations (ibid).

In addition to the escalation of conflicts, Huntington briefly discusses the diminishing importance of states. Although he accepts the realists' idea that states are still important actors on the international scene, he acknowledges that they are also greatly influenced by cultural factors (ibid). Identification with a certain civilization becomes more important than loyalty to a nation-state. Therefore, historic animosity between some countries from the same civilization gradually weakens, giving rise to cooperative relations and stronger ties. As a result, foreign policies re-focus on the newly-arisen confrontation between states from different civilizations (ibid). One can argue that this shift is proof of the accommodative behavior of political leaders. Since they see that nations do not hold the loyalty of the population to such a great extent as civilizations do, decision-makers adjust their policies and international relations a way that still keeps them in power while responding to the newly developed interests of the population.

The Theory in Perspective

It is important to note that Huntington is one of the first political scientists to discuss the relevance of culture in international relations to such a great extent (Gartzke & Gleditsch, 2006). The predominant international relations view during the Cold War, realism, totally disregards the influence of culture in interstate relations. For realists, the most important actors on the international scene are states and their ultimate goal is power maximization. The international system is anarchic, self-help, without a universal authority, and constitutes a constant struggle for power, thus providing a basis for the so called "security dilemma". Additionally, in the realist view, conflicts are not based on cultural differences, but rather on the desire of states to pursue their own interests. Even neo-realism, also known as structural realism, does not change the outlook on culture since the major focus of states, according to this theory, is on security and survival (Russett, Starr, & Kinsella, 2009).

Another international relations theory, liberalism, also disregards culture. The major objective of liberalism is that cooperation among states establishes peace (ibid). Therefore, it does not discuss conflicts and Huntington's ideas are in fact irrelevant.

Huntington's hypothesis resembles constructivism theory, which states that many of the major aspects of international relations are socially constructed: they are established because of human practice and human nature and can be changed when necessary precisely because of their origin (Wendth, 1992). The theory explores the concepts of identities and institutions which are influenced by culture, reinforce its effect on political decisions and international affairs, and mutually reaffirm each other thus providing the basis for a new self-help system (ibid.). For Huntington (1993, 1996), this relationship will have negative impact on conflicts and disputes after the end of the Cold War because of the increased importance of culture as a unifying force.

Culture vs. Alliances: The Real Clash

The Debate So Far

As already mentioned, Samuel Huntington does not receive great support for his theory. Russett, Oneal, and Cox (2000) give a very critical opinion on the Clash of Civilizations hypothesis. For their analysis, they employ several variables coming from realism and liberalism and focus on conflicts between 1950 and 1992. They conclude that the Clash of Civilizations wrongly depicts the past, is very selective in its argumentation, and does not advise on future actions. There is no supporting evidence that culture influences MIDs or that it explains the fault lines of international conflicts. Quite the contrary, typical realist and liberal variables yield strong results in explaining conflictual behavior of states (ibid).

Another major critique comes from Giacomo Chiozza (2002) who also focuses on the conflictual behavior of intercivilizational dyads but uses a dataset covering the period 1946-1997. He also tests realist and liberal variables in order to evaluate whether and how the civilizational status of a dyad affects conflicts. He finds that intercivilizational dyads are not more likely to be engaged in a conflict after the end of the Cold War; furthermore, this period is not characterized by more wars and conflicts. Even more interestingly, his results show that different-civilizational dyads are less prone to conflicts than same-civilizational dyads. Again, realist and liberal variables prove themselves to be robust and to have greater explanatory power than the variables based on Huntington's theory (ibid).

Andrej Tusicisny's (2004) approach is different although he also explores conflicts. He uses the end of the Cold War as a breaking point in his analysis covering the years 1946-2001. He distinguishes between three periods after WWII with two points of transition, the 1960s and the 1980s. He finds that each period is characterized by more conflicts than the previous one but there is no particular increase in the frequency of intercivilizational conflicts after the end of the Cold War. Furthermore, there is no significant relationship between civilizational status and the duration of a conflict either before or after the end of the Cold War. He also reaches the conclusion that post-Cold War conflicts are not more intense than Cold War conflicts. What Tusicisny explicitly notes, however, is that conflicts between states in geographical proximity but from different civilizations have risen since the 1980s (ibid).

Gartzke and Gleditsch (2006) have different approach. They differentiate not between states' civilizations but between predominant and secondary groups' cultures, thus analyzing the micro rather than the macro level of cultural differences. They also find differences and similarities not in terms of one encompassing notion, such as civilization, but through various components: religion, language, and ethnicity. The focus of their research is also on conflict proliferation in the post-WWII era (1945-2000). They find that conflicts among culturally similar groups are more likely to occur than conflicts among groups of different cultural origin. However, in the specific case of religion, there is some evidence that religiously different entities engage more often in confrontation than religiously similar ones. Consistent with previous research are the results that traditional realist and liberal variables have significant explanatory power (ibid).

As seen from the previous research on the Clash of Civilizations theory, Huntington's hypothesis is not supported. Following Huntington, scholars have focused on conflicts between civilizations. Conflicts, however, constitute only one part of international politics. Other forms of state interaction such as cooperation and alliances are also of great significance; they are often used as independent variables, a fact that further highlights their importance. Furthermore, previous research has a limited post-Cold War timeframe, which casts doubt on its conclusions since Huntington (1993, 1996) focuses exactly on the years after 1990. Another problem, specifically in Russett, Oneal, and Cox's study, stems from the methodology. The way they form their "civilizations" variable and the way they include different countries in the various categories lead to self-prophetic results (e.g. they include Israel in the Western civilization which, as expected, increases the frequency of conflicts between the Western and the Islamic world). Therefore, the current thesis tries to avoid the problems from the previous research while at the same time deviating from the mainstream of research on conflict proliferation.

Alliances in International Relations

Continuing with the theories of alliances, it is important to note why states form alliances and whether the current research has incorporated culture. One of the most influential theories for the formation of alliances is Stephen Walt's balance of threat (1985). It postulates that states' survival is achieved either through internal mobilization of domestic resources or through external mobilization, in which case states form alliances. Traditionally, there are two types of alliances, balancing and bandwagoning, and their ultimate goal is defense against a potential threat. Both types of alliances are based on assumptions of preservation of the status quo, reduction of defense costs, control of and check on partner states, and increase of states' influence in the international system (ibid). There is no assumption that culture is the driving force of alliance formation.

Further research by Ashley Leeds (2003) argues that states form alliances only when they are definite in their commitment because of the high costs of engagement. In order to precisely measure the involvement of states, Leeds classifies alliances depending on the wording of the agreements and differentiates between defensive cooperation, offensive cooperation, neutrality, nonaggression, and consultation alliances. Her major conclusion is that it is not the mere existence of an alliance that influences the international system; it is the type of alliance and thus the obligations of its members that matter (ibid).

Another study is the one by Alastair Smith (1995), who claims that the relationship between alliance formation and the outbreak of war, the motivations for the alliance formation, and the reliability of alliances are interconnected and simultaneously influence foreign-policy. He focuses on the policy preferences of states: states with similar policy preferences form an alliance only if they can get the most advantage out of it. In addition, Smith incorporates the notion of perfect information between the members of the alliance and integrates the idea of friendship between states in the formation of an alliance. He concludes that changes in the behavior of states ultimately affect the international system (ibid).

In addition to alliance formation, there is research done on alliance stability and disintegration. The most influential theory on the matter is that of Stephen Walt (1997), who claims that the formation as well as the duration of alliances directly affects international security. He differentiates between rational (states' interests) and irrational (domestic politics and personal interests) reasons for alliance duration and collapse. More precisely, Walt claims that any change in the source, identity, and scope of the threat might result in disintegration of an alliance. Moreover, domestic politics also matters because it ultimately shapes the environment in which political actors cooperate and imposes strains on the way states behave in the international scene (ibid).

Niou and Ordeshook (1994) are interested in alliance behavior within an anarchic international system. Therefore, they treat an alliance as a non-cooperative and game-theoretical model which faces international conflicts. Furthermore, they describe alliances as exclusive, i.e. the benefits of an alliance are for its members only, and list several conditions for a stable alliance. Their analysis demonstrates that stability is achieved through a limited number of members; that smaller countries can form stable alliances themselves and do not necessarily need a great power; that bipolar alliance structure is not the only stable international system; and that the role of the balancer cannot be undertaken by any individual state (ibid).

As seen from the literature review on alliances, scholars generally focus on rational state behavior which supposedly brings maximum benefits and protects states' interests. The impact of domestic influence, more specifically of state's culture, on alliance formation and stability is often regarded as being irrational, hence unimportant. Nevertheless, very few of the scholars discuss in depth what states' interests are and how they are formed. Those who do, however, focus on traditional liberal and realist explanations; thus, they do not seem to account for current changes in the international system. Furthermore, meticulous classifications of alliance types do not leave space for the inclusion of culture because they are based on rational political, rather than cultural, premises. However, if Huntington is right and culture constitutes an important element of an alliance, then previous research lacks credibility and relevance. Therefore, the current thesis approaches the issue from a different perspective, including culture among the factors explaining the behavior of alliances. Having in mind the great importance of alliances in the international scene, such research proves to be valuable because it might help scholars better understand the behavior of states and alliances.

Theory and Hypotheses

Defining Alliance

Before developing a theory combining civilizations and alliances, it is important to provide a definition of an alliance. Scholars approach alliances from two perspectives: conflictual and cooperative, but in the end they all have similar underlying views. Both Leeds (2003) and Smith (1995) include the word conflict in their definitions; Niou and Ordeshook (1994) treat an alliance as collective security cooperation. All definitions acknowledge the fact that at least two states need to enter an alliance, thus reinforcing the dyadic treatment of alliance data. Leeds (2003) stresses the need of the alliance agreement to be written, which she associates with the formal character of an alliance. For Smith (1995), alliances are non-binding and voluntary because of the anarchic nature of the international system and the existence of not-honoured alliances. Niou and Ordeshook (1994) do not discuss at all the binding and formal power of an alliance while for Walt (1997) they can be both formal and informal commitments. Nevertheless, an alliance is de facto formal and binding because of the signals it sends to the international community. Having in mind that alliances are costly (Smith, 1995) states conduct very careful cost-benefit analysis before entering them. Therefore, if a state decides to defect from its alliance obligations, it hurts its reputation on the international scene and becomes an unreliable partner. Since such agreements are created for the purpose of security (Walt, 1997), one might argue that by hurting its reputation, a state also endangers its own defense, because when it needs partners against a potential threat nobody will provide help. When an agreement is written and the relations between states are well-defined, this agreement becomes de facto binding. This fact is well-captured by Leeds (2003) and Walt (1997). Furthermore, whether an alliance signals states' behavior towards conflicts or security is de facto the same because states enter collective security agreements when they are threatened by militarized disputes. Therefore, for the purpose of the current research, an alliance is defined in the following way: "an alliance is a voluntarily binding written agreement between at least two independent states which describes cooperative or neutral relations between its members in the event of a military conflict or of a potential threat of such a conflict and thus establishes a collective security system."

Hypothesis 1

According to Huntington (1993, 1996), the primary future threats for a nation will come from states from different civilizations. Therefore, following Walt (1985), the most common type of alliance will be balancing because states will try to combine their forces against the potential threat, i.e. against states from different civilizations. It is important to note that these alliances will be between states from one and the same civilization and not between civilizations. The cultural cleavages between civilizations hinder the formation of alliances among countries from different civilizations.

Another fact that favors alliance formation within the same civilization is that, as Smith (1995) claims, states within an alliance have similar policy preferences. That corresponds to what Huntington (1993, 1996) describes about the relations within a civilization. Regional integration and economic cooperation which occur between states from the same civilization (ibid) create spillover effects and spread good practices to other areas as well, including international cooperation. Thus, states which already cooperate on one level establish friendly relations; for Smith (1995) such friendship among states is in the basis of an alliance. Therefore, the establishment of an alliance helps states ensure the fulfillment of common interests which ultimately benefit all members of the alliance.

Additionally, belonging to the same civilization reinforces similarity between two countries. Walt (1985, 1997) claims that when two countries are similar, they are prone to ally. Although in his first article he refers to ideological connection (1985), later on he accepts the possibility that transnational relations such as ethnicity, religion, and culture also foster alliance formation (1997). Walt (1997), and earlier Huntington (1993, 1996), also claim that shared identity among alliance members is related to political cooperation, including defense. Thus, alignment with culturally similar states is viewed as a defense of one's culture against a potential aggressor from another civilization. If that is what matters nowadays as Huntington says, then the fight to protect cultural integrity will overcome previous conflicts.

Therefore, the first hypothesis states the following:

H1: States from one and the same civilization are more likely to form an alliance than states from different civilizations.

Hypothesis 2

Nevertheless, despite cultural differences, states from different civilizations also form alliances. That proves Walt's statement that the international system sees bandwagoning (1985); states from different civilizations are assumed to be potential threats to each other according to Huntington (1993, 1996) but still they ally. Perhaps the most eminent example is NATO which expanded during the 1990s and is nowadays an alliance of Western and Orthodox-Slavic countries (Walt, 1997). Additionally, assuming that states form alliances only with other states from their own civilization leads to a selection problem. Important intra-civilizational alliances can be omitted (such as NATO) which diverts the results of the analysis. Thus, one has to take into account both types of alliances.

However, if cleavages between two different civilizations impede proper interaction between allies from two civilizations, then it is worth examining whether such an alliance is as stable as same-civilizational one. Niou and Ordeshook (1994) claim that alliances are stable when states find it in their self-interest to preserve the alliance. Therefore, if states get the benefit they expect from the agreement, there is no reason for an alliance to be dissolved. Stability is achieved because no member has the incentive to defect and or to join another alliance (ibid). Following this line of thought, as Huntington (1993, 1996) and Walt (1997) stress the impact of shared identity and culture on states' interests and preferences, one can argue that culturally similar states benefit from same-civilizational alliances more than states in different-civilizational alliances. In other words, if states from the same civilization ally together because they perceive states from different civilizations as a possible threat, then the members of the same-civilizational alliance benefit from their agreement because they ensure their security and defense.

On the other hand, the existence of a different-civilizational alliance is jeopardized by conflictual preferences. Imagine two civilizations, A and B, and two states a and b, belonging to civilizations A and B respectively. State a has an alliance with state b although they are culturally different. However, being culturally different means that their politics also diverges, as Huntington (1993, 1996) claims. Although all civilizations other than A and B are a potential common threat for both a and b, a assumes that also other states from B threaten it while b thinks that other states from A are potential aggressors. This partial difference might create great tension between the allies because a does not want to confront other states from A and b does not want to deteriorate its relations with other states from B in order to comply with their alliance obligations. Therefore, the mere idea of potential threat coming from the partner's civilization imposes strains on the alliance between a and b.

Furthermore, it is important not to ignore the feeling of common identity between culturally similar states (Huntington, 1993). Not only have such states common political outlook, but also they cease to see themselves as separate political units (Walt, 1997). National interests, which are the basis of alliance disintegration (Leeds & Savun, 2007; Niou & Ordeshook, 1994; ibid), become of less importance and give space to a viewpoint which treats member-states' interests as one (Walt, 1997). On the other hand, different-civilizational alliances still perceive their members as separate units connected only through some similar policy issues. Alliance formation and maintenance requires costs, such as managing the alliance, lost autonomy in decision-making and increased foreign policy consultation (Leeds and Savun, 2007); these costs are higher for a different-civilizational alliance, which is an entity without a common identity. As a result, same-civilizational alliances become more stable in the long run than different-civilizational alliances.

Overall, these factors contribute to the weakening of a different-civilizational alliance. Thus, the second hypothesis states that:

H2: An alliance consisting of states from at least two different civilizations is less stable than an alliance consisting of states from the same civilization.

D

ata and Variables

The best way to test these two hypotheses is to consider all pairs of states (dyads) in the international system. In addition, multilateral alliances are also broken down into dyads. There are two main reasons for the choice of dyads as a unit of analysis. First, the entire research on alliance behavior is based on dyads; thus, the results of the thesis are comparable with previous research. Second, for Huntington's theory, which stresses the cultural cleavages between civilizations and their members, the relations between two states are crucial. When a state decides to enter a multilateral alliance, it weighs all positive and negative aspects of it including the composition of the alliance because it reflects the goals and interests of the formation. For example, if country a wants to deepen its relations with the other states from civilization A, it would avoid joining an alliance in which the majority of the members are from civilization B because such an alliance would be regarded as a threat to A and the participation of a in it as a betrayal.

The dataset used for the analysis is the Correlates of War (COW) Formal International Alliance Dataset. There, in order for an alliance to be included, it should meet several criteria (Gibler & Sarkees, 2004). First, at least two members of the alliances should be recognized by COW as international system members. Second, each alliance is classified as a defense pact, neutrality or non-aggression pact, or an entente. Third, the dates of each alliance should be identified. The dataset covers the period 1816-2000 but it also includes alliances which started before 1816 and which have not been terminated as of the year 2000 (ibid). Thus, the dataset in fact covers a slightly longer time-span.

An alternative alliance dataset is the Alliance Treaty Obligations and Provisions (ATOP) dataset which focuses on the concrete purpose of each alliance depending on the wording of the establishing documents (Leeds, 2005). After the introduction of ATOP, the COW dataset was criticized as being limited. In addition, the ATOP dataset was recently updated and now covers the span 1815-2003 while COW still covers alliances until 2000. However, ATOP includes only alliances which were formed during this period (ibid). In addition, after revising the COW dataset, more than 1000 new alliances, which are not present in ATOP, were added including some big multilateral alliances (Gibler & Sarkees, 2004). Also, it is doubtful that alliances still in force in 2000 were terminated by 2003. Even cultural clashes such as 9/11 cannot change an alliance composition to a great extent because, as Huntington (1993, 1996) claims, cultural similarities as well as cleavages are continuous. Thus, both ATOP and COW are suitable for the current research; however, since the information needed concerns a dyadic alliance per se and not its purpose, the COW dataset seems more appropriate.

The most widely used variable in the current research is alliance type, which is coded as "0" for same civilizational dyads and "1" for different civilizational dyads. States are classified in different civilizations based on Huntington's (1996) list of civilizations: Western, Orthodox, Islamic, Latin American, Sinic, Japanese, Hindu, and African. Although many researchers include the Buddhist civilization in their analysis, Huntington explicitly says that Buddhist countries do not constitute a separate civilization (ibid). A ninth category, the Rest, combines all states which cannot find their place in any of the above categories. Among these are the Buddhist countries as well as the so-called "torn countries" such as Russia, Turkey, and Mexico. In addition to Huntington's work (1996), the article published by Russett, Oneal, and Cox (2000) also groups all countries in terms of their civilization. Hence, their list is also used as a source although slightly modified. For example, Israel which is included in the Western civilization is moved now to the Rest. For states that are listed neither by Huntington (1996) not by Russett, Oneal, and Cox (2000) the CIA Factbook information on predominant religion was consulted.

The second variable, used for the second hypothesis, is alliance duration. It calculates the number of years an alliance lasts and takes the numbers as they are, without further coding them (Bennett, 1997). The reason for that is to see how long on average an alliance from the same civilization and an alliance from different civilizations last.

Results

Alliances throughout History

An overview of alliance formation demonstrates that the second half of the 20th century witnessed a tremendous increase in the number of alliances. If one looks back in time, it becomes obvious that the international system has not always been characterized with such complex inter-state relations, which necessitate the establishment of military alliances. In particular, the end of the 19th century and very beginning of the 20th century have very few alliances, during what was a relatively peaceful period. As one moves further forward in time, it becomes apparent that the more conflicts there are in the international system, the greater the number of alliances gets.

The same holds true also for same-civilizational alliances. Their formation experiences a boom in the second half of the 20th century.

Different-civilizational alliances display a different trend, however. Although they have increased throughout the years, they exhibit a cyclical pattern. Yet, the overall trend of an increase in the number of states belonging to different civilizations that decide to engage in an alliance goes against Huntington's hypothesis.

Alliance Formation

The first test aims at analyzing whether same-civilizational alliances are more frequent than different-civilizational alliances. Table 1 demonstrates that this is clearly the case. A brief look at the cross-tabs table for each year also demonstrates that throughout the entire period 1816-2000 same-civilizational alliances in each year are formed more often than different-civilizational alliances in the same year. In addition, the Chi-square test has a significant value (p<.001). Indeed, countries from one and the same civilization are more likely to establish an alliance than countries from different civilizations.

However, Huntington (2000) claims that his hypothesis pertains exclusively to the post-Cold War period. Therefore, one should also look at the results for alliance formation after 1990. Again, the number of same-civilizational alliances after the end of the Cold War is greater than the number of different-civilizational alliances.

It would be interesting to see whether there is a difference with the Cold War. Table 3 clearly proves that even during the Cold War, the number of same-civilizational alliances is higher than the number of different-civilizational alliances. The cross tabulation again yields a supporting result: during each single year same-civilizational alliances are more frequent than different-civilizational ones with a significance value of p<.001.

Alliance Stability

After the tests on the first hypothesis, the thesis proceeds with the second one. A short descriptives table reveals that the mean duration of a same-civilizational alliance is longer than the mean duration of a different-civilizational alliance.

However, the standard deviation yields results which might lead to the opposite conclusion; different-civilizational alliances might be longer, i.e. more stable, than same-civilizational ones. Therefore, to further test the significance of the results, one has to employ more profound analysis. For that purpose, exponential model is used. It is preferred because it demonstrates the relation between culture and alliance duration while at the same time is censored for 0 and thus does not allow for negative values. The analysis yields a significant result (p<.001). However, the adjusted R square has a value of only 6.3%. Furthermore, the unstandardized Beta coefficient has a negative significant value: B=-14.692, p<.001, implying that an increase in the value of the independent variable (type of alliance) is correlated with a decrease in the value of the dependent one (alliance duration). In other words, different civilizational alliances (which are coded as 1) are correlated with fewer years of duration. The combination between the exponential model and the descriptive statistics support the second hypothesis.

The analysis of the second hypothesis continues with specific attention on the post-Cold War period. Since, according to Huntington (1993, 1996), states belonging to different civilizations will face more and longer conflicts, alliances between these states are supposedly prone to disintegration more than same-civilizational ones. As can be seen from the mean values of both types of alliances, same-civilizational alliances indeed last longer than different-civilizational. However, the almost equal mean values for both types of alliances do not suggest a rapid increase in the dissolution of different-civilizational alliances.

A look at the standard deviation, however, again necessitates the use of an exponential model. Table 8 reveals that despite the significance of the results (p<.001), the explained variance is very low: adjusted R square is only 0.8%. The negative value of the unstandardized Beta coefficient (-.466), being significant as well (p<.001), supports the proposition that culturally different alliances are prone to disintegrate more quickly than culturally similar alliances. However, the value is very low which implies that culture is almost irrelevant in the duration of an alliance.

Discussion

The entire statistical analysis demonstrates that states from one and the same civilization prefer to enter into alliances with other states from their own civilization. The conclusion supports Huntington's hypothesis (1993, 1996) because it clearly shows that there is a cultural preference in the political decision. On one hand, political leaders belonging to the same civilization have closer relations because of the similar underlining culture that shapes the dialogue between them. On the other hand, since decision-makers seek to please the population of their countries and to increase their support, they tend to bond with countries which are preferred by ordinary people; these are in most cases same-civilizational countries. Thus, alliance formation is indirectly linked to common-civilization consciousness and kin-country syndrome (ibid), which are further fostered through interaction and deeper relations. In addition, states from one and the same civilization tend to have similar international policy preferences, including security (ibid). Thus, by forming alliances within the same civilization, states combine their resources and capabilities to ensure that their interests are achieved. Moreover, cultural similarity establishes special bonding between alliance partners because of shared political, social, and cultural values. Even Walt (1997) admits that solidarity among states is one of the reasons why certain states ally. As a result, same-civilizational alliances share a similar political outlook, certain interests, and promote common values. On the other hand, states from different civilizations are regarded as unreliable or even as threatening. The existing cultural cleavages are difficult to resolve and hence foster misunderstanding, mistrust, and desire to prove that one civilization is superior to others. There is a differentiation between "us" and "them" which excludes different civilizations (ibid). As a result, inevitable tension and conflicts impede the creation of alliances across civilizations.

Although different-civilizational alliances exist, they are characterized by less durability and stability than same-civilizational ones, which supports Huntington's hypothesis. This could be for various reasons. Different-civilizational alliance members might have diverging opinions on the international system status quo, which result in eventual disintegration. Another obstacle to alliance durability can be different opinions on the way the alliance should behave within the international system. Such discrepancy has serious implications because it is related not only to the bilateral relations between the member states but also to the way the alliance responds to changes in the system. If there is no coherent policy supported by its members, then the alliance sends wrong signals about its intensions and role. An even more problematic threat to the stability of the alliance is when the two civilizations to which the paired members of the alliance belong confront each other and involve the alliance (e.g. military conflicts). Then, the different civilizational origin of the alliance members turns out to be problematic because cultural similarity is stronger than mere political interests (ibid). All these factors endanger the stability of different-civilizational alliances and make them more vulnerable to disintegration. Same-civilizational alliances, on the other hand, tend to last longer because of the common interests they serve, the cultural solidarity and mutual understanding, and the more complex interaction between the member states.

Conclusion

The current thesis started from the premise that culture plays a role in the behaviour of alliances. Previous research on alliances has not incorporated the importance of culture among the factor that determine the formation and stability of alliances thus omitting the most important determinant of the structure of the international system after the end of the Cold War according to Huntington. Therefore, the hypotheses of the thesis combine two different areas of political science research that have not been discussed together before, in an attempt to establish a new trend of research. The findings support Huntington. First, states belonging to one and the same civilization are more likely to ally than states from different civilizations. Moreover, same-civilizational alliances are more stable and more durable than different-civilizational ones. The findings are all significant and imply that culture indeed affects foreign policy decisions.

However, the question arises of why nobody before has reached the same conclusion. The answer might stem from the limitations of the current research. The variables used to test the two hypotheses are based only on Huntington without including control variables as most scholars generally do. The reason for such a choice is that the thesis aimed at testing the Clash of Civilizations in its purest form. However, additional control variables might in fact contribute to the better understanding of the relation between culture and alliances. Examples of such variables are geographical proximity and the presence of great powers in an alliance. These might better illustrate the underlying reasons for the formation of same- and different-civilizational alliances. In addition, the low explained variance hints that there are other factors, not only culture, which account for alliance duration and stability. Examples range from states' interests to the institutionalization of alliances. All these are limitations of the current research and invite for further analysis and research.

Nevertheless, the importance of the findings should not be diminished. In a world divided between the West and the developing countries, between Christianity and Islamic fundamentalism, between globalization and emerging nationalism, it is important to explore all aspects of such divisions. Since culture always seems to find its niche in such oppositions, it matters to which civilization states belong. For the specific case of alliances, since they are cooperative security arrangements, the implications involve higher stakes. If indeed culture is a significant underlying factor for the formation and stability of alliances, then the future will be characterized by a multipolar, culturally-opposed world on the verge of constant conflicts and wars. In that case, state leaders should be cautious in their decisions and re-evaluate whether cultural cleavages are worth the risk.

References