Challenging the Clash:
The case for Huntington’s civilizations in General Assembly voting patterns
Adrian Fredborg
Subject: Politics Level: C
Credit: 15c
Ventilated: Autumn 2016 Supervisor:
Examiner: Pär Nyman
Department of Government, Uppsala University
1
Index
Introduction ... 2
How to test Huntington’s theory ... 3
Why the United Nations General Assembly? ... 3
Literature Review... 4
Literature Review Clash of Civilizations ... 4
Literature Review United Nations General Assembly ... 6
Sources... 9
Methodology ...10
Civilizations and country classifications ...10
Interpreting voting options ...14
Picking sessions ...15
Method of Analysis ...17
The elusive group cohesion ...17
Statistics ...17
Analysis ...18
Results ...20
Bibliography ...22
2
Introduction
When the Berlin wall fell it not only spelled the end of a hegemonic political bloc but also many of the accepted truths in the field of international relations at the time.1 The ensuing paradigmatic upheaval spawned a great many contributions, the arguably most widely
discussed and controversial of all being Samuel Huntington’s 1993 Foreign Affairs article The Clash of Civilizations?. Huntington, who hitherto was perhaps most famous for his work on democratization, fleshed out his central thesis in his 1997 book The Clash of Civilizations and the Remaking of World Order. In it he claims that “the post-Cold War world is a world of seven or eight major civilizations. Cultural commonalities and differences shape the interests, antagonisms, and associations of states. [...] Global politics has become multipolar and multicivilizational.”2
The world has thus according to Huntington gone from a bipolar state during the Cold War to a multipolar one where civilizational identity has become salient in deciding political
loyalties, strategies and outcomes in the global political arena. This is a view that has gained considerable traction in some quarters in the following years, though the article and
subsequent book was and remains controversial in academic circles. Certain critics go as far as claiming that Huntington’s work has spawned a whole subsection of political thought in the modern political discourse that takes a very dim view of Islam, seeing it as a monolithic threat to the west3. Regardless of ones take on the subsequent debate, it seems clear that
Huntington’s book was hugely influential; and its broader themes of civilizational clashes, primordialism and politically charged religion is certainly hotly debated in public discourse on both sides of the Atlantic, both in the nativist arguments of certain presidential candidates and a Europe reeling from a sudden influx of migrants. As such it seems justified to revisit the academic clash about “the Clash”.
This humble contribution to the debate sets out to untangle some of the implications of the theory and test these empirically in the context of the United Nations General Assembly voting. To answer the broader question of the validity of Huntington’s theory a somewhat narrower approach is taken, namely by asking; is there a significant difference in the civilizational group cohesion before and after the end of the Cold War?
1 Roberts, 2008, p. 343.
2Huntington, 1997, p. 29.
3 Sharify-Funk, 2013, p. 443-444.
3
How to test Huntington’s theory
The question posed above will require some further definitions and explanations to be at all meaningful in an academic context.
“They cooperate with and ally themselves with states with similar or common culture and are more often in conflict with countries of different culture. States define threats in terms of the intentions of other states, and those intentions and how they are perceived are powerfully shaped by cultural considerations. Publics and statesmen are less likely to see threats emerging from people they feel they understand and can trust because of shared language, religion, values, institutions, and culture.”4
Huntington stresses that the affinity between nations of the same civilization goes above and beyond the crude considerations of realpolitik, that cultural bonds are of utmost importance in deciding what is considered threatening and what isn’t. That states with shared cultural and historical bonds have a more similar outlook on the world and thus greater policy cooperation can seem intuitive enough. The practical effect of this would be greater political cohesiveness in a number of arenas, one of them being the United Nations General Assembly. There are a few reasons why the General Assembly recommends itself to this sort of investigation.
Why the United Nations General Assembly?
Although many scholars have regarded the United Nations General Assembly largely as a passive arena5 for state actors, or actions taken there as largely symbolic6, there are several reasons why it still is a good place to test Huntington’s predictions, and some of the apparent weaknesses of the approach might indeed work in our favour. Huntington notes that “People use politics not just to advance their interests but also to define their identity”7, and later
“culture and cultural identities, which at the broadest level are civilization identities, are shaping the patterns of cohesion, disintegration, and conflict in the post-Cold War world.”8 This leads us to the conclusion that even seemingly trivial actions of little importance can be much more important in the context of an actor’s identity, with every miniscule decision an
4 Huntington, 1997, p. 34.
5 Dixon, 1981, p. 47.
6 Voeten, 2000, p. 185.
7 Huntington, 1997, p. 21.
8 Huntington, 1997, p. 20.
4 actor is creating its identity, and the identity is key in its behaviour in a range of questions. So even if we take a rather uncharitable view of the relative importance of UNGA voting in international politics, it is nonetheless likely well-adjusted to measuring the sort of political shifts that Huntington asserts have happened since the end of the Cold War.
Literature Review
This enquiry will require some understanding of two diverse branches of research, firstly the scientific literature on Huntington’s Clash of Civilization thesis, and secondly studies
concerned with the United Nations General Assembly and its voting process. There’s a rich and diverse range of works on both fields, and worth mentioning, also a certain degree of overlap in some of the works that attempt approaches similar to mine.
Literature Review Clash of Civilizations
Huntington’s work is far from the first dealing with civilizations, indeed he develops concepts put forward by Arnold Toynbee9, and Toynbee is together with Oswald Spengler an important inspiration for Huntington’s thinking on civilizatio ns as key concepts in world politics.10 Regardless of the origins of the concept, several subsequent studies have discussed, and with varying degrees of success tried measuring the importance or very existence of civilizations.
One important contribution is Bruce Russett and John Oneal’s Triangulating Peace. This work focuses mainly on the democratic peace theory, but also devotes much effort to testing the “rival” Clash of Civilizations theory. While in many respects a very rigorous study, there are certain methodologically questionable choices made in the section dealing with
Huntington. From a statistician’s point of view, the finding that civilization is not a
statistically significant variable (when controlling for other variables) when looking into the probability for dyadic conflict, seems impressive enough. In other words, the study finds little predictive value from its civilization category in determining the likelihood of any two states being at War at a given point in time. 11 However, the study only uses data of conflicts
between 1950 and 1992, which hardly is enough to say anything definite about a theory where
9 Huntington, 1997, p. 42-42.
10 O’Hagan, 2002, p. 18.
11 Russett and Oneal, 2000, p. 264.
5 the predicted effects would only just have begun emerging at the end of or after the collected data series. Clearly, the data needs to be better adjusted to the expected timeframe of the theory.
A somewhat different take on Huntington is put forward by Bottici and Challand who argues that we should not see it as an attempt at describing reality in a scientific manner, but rather as a way of presenting a political myth. Much like Plato’s, the myth contains prescriptive
elements that serve to make it self-fulfilling, while also creating meaning in an essentially meaningless world.12
“The result of this work, which has actually intensified
after 9/11, is that this narrative has become one of the most powerful images through which people both in Western and non-Western societies perceive the world and act within it.” –Bottici p. 323.
This sort of approach of Huntington’s thesis is comparable to some of the works of Edward Said, in that it is not really interested in Huntington’s theory’s factual credentials, indeed it suggests that Huntington is not really interested in the actual facts as far as they can be determined by empirical studies: rather it is a sort of ideologically charged myth in the making. This line of enquiry, though no doubt interesting, will not be further pursued as this study takes aim at measuring the correspondence between prediction and fact, if that misses a more important philosophical point about the whole theory than so be it, but at the very least the quote illustrates the vast range of scholarly debate on the topic.
Another study that investigates Huntington’s theory from the point of view of religions and civilizations and their impact on conflict is Jonathan Fox’s Religion, Civilization, and Civil War. Fox makes the point that Huntington’s civilizational categories in most cases adhere closely to groupings of religious denominations13. This study does not suffer from the same problems pertaining to a prematurely ending data set; it does however use two different data sets, the State Failure Dataset and the Minorities at Risk Dataset. The results are somewhat ambiguous. The State Failure dataset does indicate an increase in civilizational mass killings after the Cold War while not displaying similar results in other categories, such as state failures, but not in absolute numbers, only in proportion to declining numbers of other
12 Bottici and Challand, 2006, p. 316-18.
13 Fox, 2001, p. 158-159.
6 killings.14 However while certain areas of the study could be interpreted as supporting
Huntington, overall it found rather strong evidence against several of Huntington’s
predictions, perhaps most importantly the prediction that the post-Cold War world would be filled with civilizational Wars15.
This is rather convincing as far as civilizational conflicts go, this avenue of enquiry seems unlikely to yield any enlightening difference in results. With this in mind, we once more find a motivating reason for looking more closely at the work of the United Nations General Assembly. Having in some regard failed to convince us that civilizational conflict is more prevalent, a weaker form of the prediction can be tested, that civilizational loyalty at the very least should shape the politics of the General Assembly.
Literature Review United Nations General Assembly
The academic study of the United Nations General Assembly can be viewed as a subfield of international organizations, which have been studied since the 14th century, but in a more rigorous and systematic manner since the inception of the League of Nations in the 1920s16. Research in the interwar period was close to the mainstream of international relations and focused on descriptive and normative questions in a school of thought known as idealism.17 The immediate post-war years saw idealism challenged by realism, and a gradual shift from a traditionalist focus on “examining various rules, procedures and organs of the UN”18 towards dynamics, such as bloc voting, which became a focus point in the 1950s19. Bloc voting is naturally especially interesting for our purposes, as there is some overlap and similarity with some of Huntington’s concepts. The first systematic study of bloc voting, Margaret Ball’s Bloc Voting in the General Assembly, defined voting blocs as “any group which consistently votes as a unit on all or particular kinds of issues”20 The possible grounds for bloc voting
14 Fox, 2001, p. 177.
15 Fox, 2001, p. 180.
16 Rochester, 1986, p. 779.
17 Rochester, 1986, p. 780-81.
18 Rochester, 1986, p. 784.
19 Lijphart, 1963, p. 902.
20 Ball, 1951, p. 3.
7 evaluated by Ball are regional groups as defined by the charter, geographical proximity
groups, ideological groups, common interest groups, and ones based on consultation.21 The regional groups seem to have the most clout as voting blocs in questions regarding
nominations to different offices within the organisation as there is a “sort of gentlemen’s agreement” on the equitable division of offices between the regional groups22.
However, the regional groups’ cohesion in substantive questions is somewhat limited23, and Ball finds that “When there is a high coincidence of interest or common belief among the members of a regional or other organized group, bloc voting may be anticipated.”24 That voting can be decided by both common interests and beliefs on a regional basis perhaps touches on Huntington’s civilizational thesis as a similar mechanism for cohesion, but there is definitely something of a difference between the two concepts, regional groups are official voting blocs that however approximate some of Huntington’s proposed civilizations. Another point raised in the text that further contrast the two concepts is the importance of coordination and coercion within the groups, Ball notes that the East European group only rarely break ranks25, whether this is due to common interests and identities, soviet hegemony or high levels of policy coordination is an open question. However, the necessity for further
elaborating this distinction and actually making an effort to untangle the reasons behind the cohesion is not immediately clear; indeed Lijphart makes the case that cohesion that
“it hardly seems either necessary or desirable to make a distinction between attitudes and group influences in the study of the cohesiveness of blocs. The idea of a bloc implies both of these factors, and both play a part in achieving bloc cohesion in an interdependent way.
Consequently, it is preferable to measure cohesiveness without attempting to determine the exact reasons for voting similarities as an integral part of the measuring technique.”26
From a methodological perspective Ball’s study is however of somewhat of limited use to us.
While similar voting records are employed, her criteria for viewing the voting blocs as units in any one question is based on its complete unity27 , something that was far from always the case in the UNGA of 60 members in 1950, but most certainly is not a realistic criterion in the UNGA of the 1978-2004 and its 151-191 members. So realistically, a less strict criterion will
21 Ball, 1951, p. 3.
22 Ball, 1951, p. 5.
23 Ball, 1951, p. 29.
24 Ball, 1951, p. 31.
25 Ball, 1951, p. 12.
26 Lijphart, 1963, p. 905.
27 Ball, 1951, e.g. p. 21.
8 have to be applied for the much larger groups, but one that still does not result in spurious groups being viewed as real voting groups. While Ball does include a defensible amount of data the analysis is rather non-systematic, Rochester’s assertion that that UN research in the early fifties was more interpretive than rigorously empirical seems to hold true28.
Erik Voeten’s Clashes in the Assembly attempts to test some conflicting hypotheses regarding the structure of the post-Cold War world order by investigating the UNGA voting behaviour of states. While the relative bipolarity of the Cold War world seems an uncontentious suggestion, he suggests a number of post-Cold War hypotheses that all incorporate the two concepts of realignment and dealignment. In short, either the end of the Cold War brought about voting patterns less influenced by alignment groups and more by state interest through dealignment, or the breakdown of one voting pattern was merely substituted by new voting alignment groups through realignment.29 Furthermore, to operationalize and evaluate the different theories’ claims he uses the concepts of dimensionality and overall similarity to Cold War. High dimensionality denotes a situation where alliances are less rigid and more
multipolar than previously, while the similarity criterion relates to the major lines of conflict.30 One of the theories that Voeten investigates is the so called Civilizational
Hypothesis based on Huntington’s work. It is predicted that realignment in accordance with civilizational voting blocs should appear in the voting records from the UNGA sessions. To use the previously introduced terms, we would expect to see high dimensionality combined with low overall structural similarity to the Cold War.31 However, the study, comparing roll- call votes data from UNGA sessions between 1946-88 with data from 1991-96 sessions, makes a rather simplistic two dimensional conflict analysis based on a primary east-west conflict, and a secondary south-north conflict32. The regressive analysis using civilizations as dummy variables does find some statistically significant effect for western and orthodox civilizations, and a movement of Islamic countries away from the western position, but generally rejects the hypothesis.33
One distinction that seems fruitful to make is that between culture and civilization, as they at first glance might appear similar. In Triangulating Peace Huntington is quoted as describing
28 Rochester, 1986, p. 785.
29 Voeten, 2000, p. 187.
30 Voeten, 2000, p. 186.
31 Voeten, 2000, p. 191.
32 Voeten, 2000, p. 202-203.
33 Voeten, 2000, p. 213.
9 civilizations as the “highest cultural grouping” and as the “broadest level of identity”, which form a basis for actors’ understanding of the international arena.34
Sources
The United Nations General Assembly has two types of votes, summary and recorded ones.
This enquiry relies on recorded votes since they are the only ones that identify the position of individual member states, summary votes are only recorded with number of each type of vote and are as such for all intents and purposes useless to us.35
There are a number of ways to access the voting records, chiefly among them the press releases and subsequent meeting records, earlier secondary sources that compile the data, and lastly the voting charts found in the Index to Proceedings of the General Assembly. The meeting records are quite detailed and have the advantage of including the corrections submitted after the votes, but the meticulous nature of them and the great amount of text a researcher have to go through to extract the voting data make them impractical for our purposes. There are quite a few studies that rely on compilations of UNGA data, such as ICPSR:s United Nations Roll Call Data36 or Voeten et al’s United Nations General Assembly Voting Data37. However the ICPSR data does not cover the entire period of interest for this paper, and Voeten’s data is impractical for reasons pertaining to file format, using it requires proficiency in the R programming language, something that despite my best efforts eludes me.
The Index to Proceedings of the General Assembly gives us the data in unadulterated form in simple indices that lend themselves well to being compiled in accordance with our needs. The only drawback to the indices is that all of the Cold War dates and most of the post-Cold War ones only exist in the analogue form, meaning that a rather time consuming process of digitalising the chosen data will have to be carried out in order to interact with it in any meaningful way. Despite this admittedly rather cumbersome process of transcribing the indices to Excel files, it is likely much less time consuming than going through the meeting records. To confirm this, I attempted extracting the voting records from the meeting records for the resolution A/RES/53/4, on the “Vote on Ending Embargo against Cuba”. Transcribing this single vote took the greater part of an afternoon, leading me to the conclusion that the
34 Russett and Oneal, 2000, p. 246.
35 United Nations, “General Assembly Voting”. http://research.un.org/en/docs/ga/voting
36 ICPSR, 1984.
37 Voeten, 2009.
10 paper could only reasonably rely on the voting data from the Index to Proceedings of the General Assembly.
It is worth noting that the method for input into the indices changed during the relevant period, up until 1982 the votes were recorded from the provisional Verbatim Records38, and as such also contains corrections submitted after votes. The indices after 1982 only records the votes “as is”, as they are reported to the UNBIS database through the electronic voting machine in the General Assembly, and thus contains no corrections39. To rectify this inconsistency in the data, all corrections from the 1979 and 1983 indices have been ignored and the initial vote used instead, so as to avoid any systematic bias that might have been introduced by taking corrections into account in the first two indices but not in the subsequent four. The alternative, using the corrected version for the first two indices, and consulting the verbatim records for the following four, was deemed too time consuming to be feasible. This way, any bias stemming from a discrepancy between intended and actual vote is thought to at least be relatively constant over the data set.
Methodology
Civilizations and country classifications
To properly test the theory an attempt to sort every country in the United Nations General Assembly into the different civilizations will have to be made. In order to do this, we must first explore Huntington’s own categorizations. It has been noted in previous studies that Huntington is not always very clear on what counts as a civilization, let alone who the members are40, which naturally can prove problematic. As far as is possible however, Huntington’s map of the civilizations serves as the basis for our own categorization of the civilizational groups. On page 26 and 27 in The Clash of Civilizations we find Huntington’s vision of the post-Cold War world, reproduced below:
38 United Nations Publications, 1983, p. 401., United Nations Publications, 1987, p. 335.
39 United Nations, “Index to Proceedings of the General Assembly”. http://research.un.org/en/docs/ga/voting
40 Fox, 2001, p. 171.
11 Source: Wikimedia Commons, compared for accuracy with Huntington’s map4142
While the majority of the world’s countries are readily and unambiguously categorized into one of the civilizations, there are several less than unambiguous cases. Indeed, Huntington is rather clear about the sometimes unclear nature of civilizations: “Civilizations have no clear- cut boundaries and no precise beginnings and endings [...] The cultures of peoples interact and overlap”43 Nevertheless, we will have to make these distinctions for all the members of United Nations General Assembly in order to carry out the analysis. However, it is possible that some countries are simply not members of any of the civilizations as defined by Huntington and then there is little use in categorizing them as such, instead there will be a misc. category for these cases.
The map shows what Huntington designates as the fault line between the African civilization and the Islamic one cutting through for instance Nigeria and Benin. These are what
Huntington would call cleft countries where two or more substantial groups belonging to different civilizations cleave the countries between civilizations.44 In order to correctly categorize these cases of cleft countries and countries which are either too small to appear on the map or which are unclearly marked, we need to look into Huntington’s description of the civilizations. “A civilization is thus the highest cultural grouping of people and the broadest level of cultural identity people have short of that which distinguishes humans from other
41 Wikimedia Commons, https://commons.wikimedia.org/wiki/File :Clash_of_Civ ilizations_map.png
42 Huntington, 1997, p. 26-27.
43 Huntington, 1997, p. 43.
44 Huntington, 1997, p. 138.
12 species”45A key component to Huntington’s civilizations is thus shared culture. But the
cultural component is not the only aspect, indeed the names such as Islamic and Hindu
civilizations hint at a strong religious aspect too. Religion is a rather straightforward factor to measure in cases of cleft countries in Africa, consulting the CIA World Factbook to get the relative size of the communities allow us to assign civilizations to the less clear cut cases. The same tactic however is less successful in assigning civilizational belonging to the cleft
countries Suriname and Guyana, which the map shows as being partially Hindu, partially African in their civilization. This, coupled with their regional and historical context in Latin America makes them difficult to code as belonging to either, instead prompting two more additions to the misc. category.
However, for instance in assigning civilizations to the Caribbean nations religion does not seem to be the dividing line in Huntington’s definition between the Western and Latin American Civilization. Rather, here geography and Spanish colonial history seems to be a more decisive factor, so these factors will be considered when civilization is not immediately apparent from the map. However, this region raises another methodological concern, that sometimes Huntington’s map seems to assign civilizational belonging in a manner
incongruous with the overall descriptions of the civilizations. For instance Belize is assigned to the Latin American group despite being settled by the English, English being the official language and there being a mix of Christian communities in the country.46 Such discrepancies lead to some subsequent studies reassigning the civilizational identity of some countries such as Belize.47 However, this study will not try to second-guess Huntington’s civilizational categories, as far as possible, the map will serve as the basis for the categories, and as such Belize and the Caribbean islands belong to the Latin American civilization as clearly shown on it.
It is thus only in cases where the map is unclear that the particularities of a certain country will be checked and compared with descriptions of possible civilizations, and it is only in cases where this second categorization effort is unsuccessful that the misc. category applies.
In addition to the Caribbean cases and the cleft countries of Africa, it is worth commenting on two additional cases. Firstly, the island nations of the Pacific are not visible on Huntington’s map, but they also have the problem of not being identifiable as belonging to any of the
45 Huntington, 1997, p. 43.
46 CIA World Factbook, https://www.cia.gov/library/publications/the-world-factbook/geos/bh.html
47 Russett, Oneal, and Cox, 2000, p. 606.
13 civilizations based on their descriptions found in Huntington’s book48. As such they will be assigned to the misc. group, in all likelihood Huntington did not consider the civilizational status of these nations extensively while formulating his thesis. Russett, Oneal, and Cox choose to assign some of these to these pacific islands to the Western group, but I maintain that it seems unlikely that this was Huntington’s intention, and as such it is fairer to exclude them from the analysis altogether.
Furthermore, Israel, a focus point of many resolutions in the General Assembly, lacks a clear civilizational belonging. The map is not very helpful, but nor is the text overall. Huntingto n mentions it both as distinct and a non-member of the Western civilization49, but later he describes it as not only distinct from the Islamic civilization but also implies it possibly being a civilization in its own right.50 However this potential Jewish civilization is never mentioned in any of the lists of the civilizations, nor is Israel treated as a separate civilization in other parts of the text. As such, I will assign it to the misc. category. This is a choice of
categorization that might conceivably impact the result of the study somewhat, due to Israel being the focus of many resolutions, and often voting against the majority on these51, a different categorization could for instance have skewed the cohesion level of the recipient civilization, most likely the Western Civilization. As such it is arguably a rather important choice not categorizing it as either Western or Islamic, but due to the design of the study, a choice to assign it to a civilization of its own would not affect the outcome at all compared to categorizing it as misc., as the resulting civilization at any rate would be too small to measure cohesion level in.
This basic scheme for categorizing all UNGA members will be retained in the all the different sessions with which this study is concerned. Although Huntington mentions the possibility for civilizational belonging shifting with time, he does not mention any relevant examples that occur or is successfully carried out during our timeline, nor does it seem to be the case if one checks the later dates. Indeed, it seems to be a rather arduous process that more often than not leads to countries torn between civilizations rather than a clean reorientation of civilizational
48 Huntington, 1997,p. 45-47.
49 Huntington, 1997, p. 90.
50 Huntington, 1997, p. 188.
51 Holloway, 1990, p. 280.
14 belonging.52 The list thus changes for every session investigated but the basic outline and principles are the same in all sessions.
Interpreting voting options
Any paper examining the UNGA voting records naturally run into the issue of how to interpret the different categories of outcomes, namely yes votes, no votes, abstained votes, and absentee voters. The Yes/No dichotomy is straightforward enough in that there is little ambiguity in what this means for the resolution at hand, although states could presumably vote similarly for dissimilar reasons. However, the abstained votes are an altogether different matter. Many previous studies have interpreted this as a weaker form of no, and as such they are more common than the stronger “No” vote53. In other studies, such as the Index of
Agreement the authors have gravitated towards an assumption of abstentions being as much a yes as a no, thus placing it squarely between the two poles effectively not taking sides.54 This is also the view taken by Gartzke when making an Affinity of Nations Index based on UNGA roll call data55, a classification largely motivated by the need for assigning numerical values to the different voting alternatives. Thus, although the exact classification of abstentions varies in the literature, and it is sometimes regarded as close to a neutral choice when needed for methodological reasons pertaining to the making of indices, it seems to be more common to regard it as a weaker no (at the very least when there is no need for numerical values), a way of showing disapproval without fundamentally blocking an effort56. It would thus seem most reasonable to put it into the same group as the “no’s, and only differentiate between yes and no/abstain.
The treatment of absentee votes is somewhat different in that some studies view them as similar to abstentions, while others interpret it to be caused more by a lack of opportunity to vote, such as through regime change or insufficient UN representation, rather than an active choice on the part of the absentee.57 This makes it somewhat more complicated than in the case of the abstentions, for it does not seem at all clear that all absent votes are “shirking”
52 Huntington, 1997, p. 139-157.
53 Voeten, 2003, p. 735.
54 Lijphart, 1963, p. 910.
55 Gartzke, 1998, p. 15.
56 Voeten, 2003, p. 735.
57 Voeten, 2003, p. 735.
15 their responsibility for similar reasons, or indeed to signal the same thing. A quick look into the data reveals some telling patterns but also no single readily available overarching
explanation or geographical distribution for the absentees. For instance, in the 33rd session, for the ten first recorded votes, the following countries are recorded as absent from seven or more votes: Djibouti, Libyan Arab Jamahiriya, Grenada, Benin, Cape Verde, Mozambique, South Africa, Democratic Kampuchea, Seychelles and Solomon Island. 58 While United Nations representation capacity could very well enter into the equation, something indicated by the relative absence of developed countries on the list, they come from diverse parts of the world, and it is not unlikely that South Africa (piranha state), Democratic Kampuchea (political upheaval), and Grenada for instance had different reasons for not participating fully in the voting process at this point. As such there is little reason to subsume the absent votes into any of the other groups, and instead just treat it as missing data and adjust the size of the groups in accordance with the fluctuating number of votes in the Yes and No/Abstain categories.
This solution to the problem of missing votes is naturally not entirely flawless; one could perhaps argue that there are systemic reasons for non-participation and seemingly
idiosyncratic reasons for absence is no guarantee of it being inconsequential for the results.
However, there are reasons for us not to think this will bias the results significantly: firstly, the major group with the largest percentage of absentee voters is also one of the most
homogenous when it comes to voting; the African group. Secondly, it does not chiefly seem to be a way of showing mild dissent, because absentee voters seem no more likely to vote against the majority in related votes (there are often several votes concerning different sections of the same document). Indeed, Voeten found in his data set that in 68 percent of abstentions the absentee was also absent for the next vote, which would be incongruous with the assumption that they are absent to protest a particular resolution.59 Thirdly, it seems a rather common position taken in more recent literature,6061 though as already mentioned above, by no means the only way.6263
Picking sessions
Because of methodological reasons as well as the limited time and processing power, only a smaller sample of the total number of votes held since the inception of the United Nations
58 United Nations Publications, 1979, Voting Records.
59 Voeten, 2009, p. 4.
60 Voeten, 2003, p. 735.
61 Voeten, 2000, p. 193.
62 Rai, 1972, p. 590.
63 Rieselbach, 1960, p. 302.
16 General Assembly can be included in the analysis. Thus, it is important to make the data that can be analysed count. Some concerns, such as regarding the timeline, can be gleamed from previous works (see Fox), while others are specific to this study. The timeline requires us to have sessions both before and after the end of the Cold War, and not limit the upper reach to the transitional years of the early nineties. In order to give us an equally good idea of both periods of time, we take an equal amount of sessions from the Cold War and the post-Cold War period.
The exact number of sessions is not uncomplicated to pick, but illustrates the trade-off between depth and width. With more sessions looked into, the individual session average loses in accuracy, and there is a real risk that a few salient questions with idiosyncratic
alliances shift the figures in unpredictable ways. It is for instance not uncommon that different parts of the same resolution or a series of related resolutions are voted on in separate votes.
Thus, one is inclined to limit the amount of sessions and thus gain a slightly better protection against such events due to the increase in votes analysed for each session. With this trade-off in mind, it is really only a question of how many votes one can imagine analysing. In this case, after something of a trial run with one of the indices, I settled for six sessions, and twenty votes for each, the ten first and ten last in each session. These numbers may sound unimpressive, particularly considering the astronomical numbers possible to analyse with modern computer power, but given the constraints posed by the analogue nature of the Index to the Proceedings of the General Assembly, as well as the number of countries voting in each vote, it is still a rather sizeable task.
Wanting to keep the two categories of session relatively closely clustered, four year intervals were reasonable, with a ten-year gap between the Cold War and post-Cold War. After this ratio was decided on, the specific years were arrived at after consulting the available material at the local library; some of the indices were unavailable for unknown reasons and as such the investigated sessions had to match the available material.
Period Session Year
Cold War 33rd 1978
Cold War 37th 1982
Cold War 41st 1986
Post-Cold War 51st 1996
Post-Cold War 55th 2000
17
Post-Cold War 59th 2004
The sessions above are thus the subject of the analysis. Any change in the civilizational cohesion as a result of the end of the Cold War will have had a few years to develop in this setup, so there should not be any problems stemming from trying to measure a trend at too early a point in the development.
Method of Analysis
The elusive group cohesion
The data collected in accordance with the method described above will also have to be analysed. For each of the six sessions a list of the voting records for twenty different votes, the ten first and the ten last will be compiled for each civilization. These will form the basis for the subsequent figures of group cohesion. The concept of group cohesion has perhaps been shrouded in mystery until now, but it is arrived at by quite simply taking the number of
countries in a civilization voting with the majority (in that civilization, not the overall session, although the two are more often than not the same) divided by the total numbers of votes cast by the civilization. To illustrate this, a particular vote might have twenty countries in a civilization voting yes, ten no, giving us a cohesion of twenty yes / (twenty yes+ ten no) or 20/30 = 0,6666 or rounded up to 67 % cohesion, a very low score indeed. These individual civilizational cohesion scores are then added up and divided by the possible maximum to give us a session average. The three session averages from the Cold War and the three session averages from the post-Cold War are in turn added up to give us each civilization’s Cold War and post-Cold War average.
Statistics
Having compiled Cold War and post-Cold War averages for all civilizations, we will be ready for the last part of the analysis. This part calls for the use of some rudimentary statistics. To remind ourselves, what we want to determine is if the end of the Cold War saw a statistically significant increase in the cohesiveness of Huntington’s civilizations. In order to determine this, we will have to make a hypothesis test.
18 If Huntington is right, the population’s average cohesiveness during the Cold War, P1, ought to be lower than the population’s average cohesiveness after the Cold War, P2. Or to put it differently; P1<P2- This is the alternative hypothesis, or Ha. Alternative to what one might ask? It’s the alternative to the null hypothesis, H0, which stipulates that there is no statistically significant difference between the post-Cold War and Cold War cohesiveness averages. Or to put it differently; P1=P2. This gives us;
Ha: P1<P2
H0: P1=P2
So, these are the two different hypotheses which we need to test. The Ha hypothesis is one sided, meaning that the hypothesis is only correct if P1 is statistically significantly smaller than P2, if the opposite is true, that P2 is statistically significantly smaller than P1, that lends no support to the hypothesis, rather the opposite. This would mean that the cohesiveness actually went down after the Cold War, which naturally would not be in line with Huntington’s
hypothesis.
To test the two different hypotheses, we need to use the following formula:
𝑍 = (𝑝̂1 − 𝑝̂2)
√𝑝̂1(1 − 𝑝̂1)
𝑛1 +𝑝̂2(1 − 𝑝̂2) 𝑛2
The 𝑝̂ is the approximated population average, at the Cold War (𝑝̂1), and after the Cold War (𝑝̂2). The symbols of n1 and n2 represent a given civilization’s size during and after the Cold War. These figures are averages for the three sessions, meaning that changes in the size of the different civilizations are taken into account. To actually make sense of the number derived from the equation, we must pick a significance level. For our purposes, a significance level (α) of 0,05 will do nicely. This means that in order to get a statistically significant result confirming Huntington’s hypothesis the z value will have to be higher than α = 0,05, that is z>
(-)1,96. In that case the null hypothesis will be rejected in favour of the alternative hypothesis, favouring Huntington’s hypothesis.
Analysis
Using the method described above, we first get the session cohesion for each civilization, which is nothing more than average cohesion for the civilization in the ten first and ten last
19 votes at a particular session. The cohesion for a particular vote is the majority vote divided by the total number of members voting on that particular issue. All numbers are thus percentages, where 100 % means that the entire civilization is voting for the same option, potentially reaching 50 % if the civilization is evenly split between two options. In practise 50 % is never reached, and most civilizations are consistently closer to 100 % cohesion.
Cold War Session Cohesion Post-Cold War Session Cohesion
Civilization 33rd 37th 41st 51st 55th 59th
Islamic 93 95 96 94 94 95
Latin 93 91 96 90 94 84
Western 74 82 78 91 91 93
African 98 95 92 94 99 94
The relative stability in cohesiveness of most civilizations is illustrated below. Note the increase in the green bar post-Cold War, Europe is perhaps unsurprisingly the civilization most affected by the end of the cold war.
The averages of the three session averages for cohesiveness for each period, the Cold War and the Post-Cold War, are calculated to enable the last part of the analysis. This gives us two numbers for each civilization which can be compared using hypothesis tests. Although it is tempting to draw certain conclusions based on the session or period averages, we will refrain from doing so until we have the results of the last stage of the analysis.
Cold War Post-Cold War
20
Civilization Average cohesion Civilization Average cohesion
Islamic 95% Islamic 94%
Latin American 93% Latin American 89%
Western 78% Western 92%
African 95% African 96%
Islamic: 𝑍 = (0,95− 0,94)
√0,95(1−0,95)
40 +0,94(1−0,94) 47
= 0,2046578041
Western: 𝑍 = (0,78− 0,92)
√0,78(1−0,78)
25 +0,92(1−0,92) 35
=(-)1,478454319
Latin American: 𝑍 = (0,93− 0,89)
√0,93(1−0,93)
30 +0,89(1−0,89) 32
=0,5534046704
African: 𝑍 = (0,95− 0,96)
√0,95(1−0,95)
28 +0,96(1−0,96) 29
=(-)0,1824917984
Results
Civilization z-value
Islamic 0,2046578041
Western (-)1,478454319
Latin America 0,5534046704
African (-)0,1824917984
As we can see none of the above values are higher (.) than Z= -1,96, meaning that the result is not statistically significant at our chosen significance level. We reject the alternative
hypothesis in favour of the null hypothesis, we can not assume that the overall cohesion in the United Nations General Assembly has gone up after end of the Cold War in any of the
civilizations. Getting a nonsignificant result in a hypothesis test can seem dissuading, but it
21 indicates that the approaching intercivilizational clash, as envisioned by Huntington, if at all true, has been very modest.
But the Western civilization seems rather close to having a significantly higher cohesion after the end of the Cold War, certainly that lends some credence to Huntington? In a way, and this is perhaps where the demise of the Soviet Union is most immediately apparent: the EU, which had already had increasing levels of internal cohesion before the end of the Cold War64 can reasonably be expected to have increasingly coordinated voting patterns with the new members joining the union from the Warsaw pact, regardless of any increase in actual salience of the civilizational aspect.
That supposed members of a Western civilization that were previously aligned to Moscow in their voting patterns in a most consistent manner65 realign to a new voting bloc of sorts in line with their civilization will certainly increase the cohesion of the western group. The attentive reader will perhaps recall the question of making a distinction between cohesion caused by attitude and group influence. It was concluded that Huntington’s explanation relies more heavily on the former, while still taking the latter into account. Regardless of attitude however, the end of the Cold War likely caused the new EU members to come under
increasing EU group influence, thus causing higher cohesion levels. Given this development we would however expect a much stronger increase in cohesion if the civilizational aspect was the prime motivator.
Regardless of ones view of the “close enough” status of the Western Civilization, it is at the very least quite clear that the Clash of Civilization does not adequately predict the political development of the other major civilizations, in line with the other empirically driven tests of Huntington’s hypothesis, there is scant support for the overall claim of clashing civilizations.
Indeed, one could perhaps doubt the very existence of these civilizations, especially as viable and fruitful categories for understanding the ordering principles of the international system.
Naturally there are many aspects of the proposed theoretical framework that have not been tested here, and as to their validity this study can not attest.
In a sense, there is perhaps with a modicum of relief that one sees that one cannot find support for the theory that the world is more riven with intercivilizational strife between increasingly isolationist religious and cultural groups.
64 Hosli et al, 2010, p. 25.
65 Lijphart, 1963, p. 904.
22
Bibliography
Ball, M., (1951), Bloc Voting in the General Assembly. International Organization, vol. 5, no. 1, p. 3- 31.
Bottici, C., and Challand, B., (2006), Rethinking Political Myth, the Clash of Civilizations as a Self- Fulfilling Prophecy. European Journal of Social theory, vol. 9, no. 3, p. 315-336.
CIA, The World Factbook. People and Society: Belize. https://www.cia.gov/library/publications/the- world-factbook/geos/bh.html Retrieved 2016-05-18.
Dixon, W., (1981), The Emerging Image of U.N. politics. World Politics, vol. 34, no. 1., p. 47-61.
Fox, J., (2001), Religion, Civilization, and Civil War; 1945 through the new millennium. Lexington Books, Lanham.
Gartzke, E., (1998), Kant we all just get along? Opportunity, Willingness, and the Origins of the Democratic Peace. American Journal of Political Science, vol. 42, no. 1, p 1-27.
Holloway, S., (1990), Forty Years of United Nations General Assembly Voting. Canadian Journal of Political Science, vol. 23, no. 2, 279-296.
Hosli, M., van Kampen, E., Meijerink, F., Tennis, K., (2010), Voting Cohesion in the United Nations General Assembly: the Case of the European Union. ECPR Fifth Pan-European Conference, 24-26th June, 2010, Porto.
Huntington, s., (1993), The Clash of Civilizations?. Foreign Affairs, vol. 72, no 3, p. 22-49.
Huntington, S., (1997), The Clash of Civilizations and the Remaking of World Order. Touchstone, New York.
ICPSR, (1984) United Nations Roll Call Data, 1946-1985.
http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/5512?sortBy=&q=5512&searchSourc e=revise Retrieved 2016-01-28.
Lijphart, A., (1963), The Analysis of Bloc Voting in the General Assembly: A Critique and a Proposal.
The American Political Science Review, vol. 57, no. 4, p. 902-917.
O’Hagan, J., (2002), Conceptualizing the West in International Relations Thought. From Spengler to Said. Palgrave Macmillan: Basingstoke
23 Rai, K., (1972), Foreign Policy and Voting in the UN General Assembly. International Organizations, vol. 26, no. 3, p. 589-594.
Rieselbach, L., (1960), Quantitative Techniques for Studying Voting Behavior in the UN General Assembly. International Organization, vol. 14, no. 2, p 291-306.
Roberts, A., (2008), International Relations after the Cold War. International Affairs, vol. 84, no. 2, p.
335-350.
Rochester, J. M., (1986), The Rise and Fall of International Organization as a Field of Study.
International Organization, vol. 40, no. 4, p. 777-813.
Russett, B., and Oneal, J. ( 2000), Triangulating Peace, Democracy, interdependence, international organizations. Norton, New York.
Russett, B., Oneal, J., and Cox, M., (2000), Clash of Civilizations, or Realism and Liberalism Déjà vu? Journal of Peace Research, vol. 37, no. 5, p. 583-608.
Sharify-Funk, M., 2013, Pervasive Anxiety about Islam: a Critical Reading of Contemporary “Clash”
Literature. Religions, vol. 4, no. 4, p. 443-468. http://www.mdpi.com/2077-1444/4/4/443 Retrieved 2016-08-20.
United Nations, UN Documentation: General Assembly. Dag Hammarskjöld Library Research Guide.
http://research.un.org/en/docs/ga/voting Retrieved 2016-02-01.
United Nations Publications, (1979, 1983, 1987, 1997, 2001, 2005), Index to Proceedings of the General Assembly, Dag Hammarskjöld Library, New York.
Voeten, E., (2000), Clashes in the Assembly. International Organization, vol. 54, no. 2, p. 185-215.
Voeten , E., (2003), Resisting the Lonely Superpower: Responses of States in the United Nations to U.S. Dominance. Journal of Politics, vol. 66, no. 11, p. 729-754.
Voeten, E., Strezhnev, A., Bailey, M., (2009) United Nations General Assembly Voting Data,http://hdl.handle.net/1902.1/12379, Harvard Dataverse, V12. Retrieved 2016-01-28.
Wikimedia Commons, Clash of Civilizations map.
https://commons.wikimedia.org/wiki/File:Clash_of_Civilizations_map.png Retrieved 2016-05-18.
