Culture and Negotiations between Rival States

Transcription

1 Culture and Negotiations between Rival States Patrick M. Regan Binghamton University and Russell J. Leng Middlebury College April, 2008 Prepared for Presentation at a workshop on Culture and Conflict, Binghamton University, April 11 & 12,

2 Culture and Negotiation between Rival States This paper examines the use of culture as an analytic tool for the study of diplomatic relations between rival states, specifically their readiness to use negotiation, including mediation, to resolve their disputes, and their success in doing so. It tests the general proposition that cultural similarities and differences affect how state governments interact with each other, including their management of disputes. We begin with a discussion of conceptual problems related to the meaning of the term culture generally, as well as a level of analysis issue particular to the use of culture as an analytic tool in the study of interstate relations. We suggest an approach that focuses on those cultural characteristics that citizens of a particular state share in common. Then we turn to the question of whether local cultural similarities and differences between rival states are likely to prevail over a global diplomatic culture of political realism. We hypothesize that, even in negotiations where vital interests are at stake, local cultural attributes exert influence in two ways. First, cultural similarities or differences can affect the attitudes of the parties toward each other. Second, cultural similarities or differences can affect their ability to communicate effectively with each other. We would expect states that are culturally similar to be more inclined to attempt to settle their disputes peacefully through negotiation, and to be more successful in doing so. We then describe a simple research design to test these hypotheses, and present the results of some initial findings. 2

3 Culture as an Analytic Concept During the 19 th century, and in certain circles today, the term culture has had evolutionary implications, as something that is acquired as one becomes more civilized. Some people are considered, or consider themselves, more cultured than others. Needless to say, such a view can have, and has had political implications. The global evolutionary perspective was rejected by a number of 20 th century anthropologists, such as Benedict and Mead, in favor of a focus on the plurality of cultures across societies, without moral distinctions. This too, has political implications, as evidenced by contemporary debates over moral relativism. Anthropologists traditionally have viewed culture as something that is a function of society, that is, that the cultural traits of an individual are acquired from an integrated society of which he or she is a part. Individuals acquire cultural attributes from their social environment, and those attributes provide a guide for interactions with others. Societal culture encourages common patterns of interacting and reacting to the actions of others by creating a set of shared values and beliefs to guide behavior, including interactions with others. Kevin Avruch (1998), who has written insightfully about culture and conflict resolution, argues that culture is most properly analyzed at the individual level, as the derivative of individual experience, something learned or created by individuals themselves 1 or passed on to them socially by contemporaries or ancestors (1998,:50). For Avruch, local cultures are those complex systems of meanings created, shared, and transmitted (socially inherited) by individuals in particular groups (1998:10), but an 1 We do not discuss the added complication of cultural attributes that may be self-generated. See Avruch (1998:16-17). 3

4 individual s culture may be an amalgam of influences from several groups, including one s nation, region, family, social class, ethnicity, religion, language, and occupation. This begs the question of the efficacy of examining culture at the state level. If each individual is an amalgam of multiple cultural influences, and the potency of those influences varies across individuals within the same society, how can one make valid cultural distinctions between governmental representatives based on their nationality? Cultural generalizations about individuals based on state citizenship become especially problematic today, with the increasingly heterogeneous demographic composition of so many states. Thus any discussion of national cultural attributes must distinguish between those attributes that are common to all members of the state, as opposed to stereotypical images of certain nationalities. With that caveat in mind, we propose that there are certain state-level cultural attributes, shared by virtually all members of a state, which affect the meanings attached to their relationships with other states and their interactions with those states, regardless of the individuals other cultural attributes. We recognize that the relative potency of cultural attributes observed at the state level will vary from individual to individual, but we hypothesize that these effects will be potent enough to affect how state leaders manage their disputes with rival states. In effect we assume that national leaders, at least when addressing questions of security over interstate conflict, act as if they represent an amalgamation of national cultural attributes. 4

5 How Culture May Affect Interstate Negotiation Cultural attributes observed at the national level influence interstate negotiations in two ways. First, cultural similarities or differences affect the attitudes of the parties toward each other along two related dimensions: affinity/antipathy and trust/distrust. It is reasonable to assume that states that enjoy more friendly relations with each other are more likely to attempt to resolve their differences peacefully through negotiation, and to be more successful at doing so. By the same token, trust is cited by many students of negotiation as a key variable in the readiness of states to negotiate and to do so successfully (Zartman and Berman, 1982; Pruitt, 2000; Kelman, 1997). Second, cultural similarities or differences can affect the ability of the parties to communicate effectively with each other. Effective communication is dependent on clear transmission and shared understandings of the signals that one party sends to the other during the course of a dispute. These signals are encoded in verbal or written communications, or conveyed through actions whose meaning must be interpreted by the recipient. Recipient expectations based on culturally derived rules regarding interaction, including negotiation, affect how signals are decoded or interpreted. Studies of intercultural communication have demonstrated that how signals are decoded across cultures can result in misinterpretation of the message and misperceptions of the sender s intentions (Cohen, 1997; Hofstede, 1991; Triandis, 1995; Hall, 1977, Ross, 1993). Cultural differences can cause parties to screen out information inconsistent with culturally grounded assumptions, to project meanings onto the other party s words and actions, and to misinterpret the other party s intentions (Fisher, 1980). Moreover, what 5

6 Cohen (1997) has termed intercultural dissonance, may be exacerbated by distrust and antipathy between cultures. An earlier empirical study by the authors demonstrated that cultural differences do matter, even when the stakes are high (Regan and Leng, 2003). We generated data on negotiations in a sample of 294 militarized interstate disputes (MIDs) occurring between 1945 and We found that when the participant states were culturally similar along four dimensions region, social system, democratic political culture, and religion they were more inclined to seek negotiated settlements to their disputes and to be successful in doing so. We found less consistent results in a comparable study of the relationship between cultural similarities and differences and outcomes of mediation in MIDs, although we did find a strong positive correlation between shared democratic cultures and the success of mediation. (Leng and Regan, 2003). What those studies did not test, however, was the effect of culture on lower level disputes, that is, those disputes that did not become militarized. Presumably, the effects of cultural similarities and differences on the willingness of the disputants to seek negotiated settlements, and their success in doing so, would be greater in those cases. By including negotiations between rival states in lower level disputes, that is, those nonmilitarized disputes occurring within the intervals between MIDs, along with negotiations in MIDs, we can examine the validity of the hypothesis that the effects of cultural differences dissipate when the stakes are high enough for the threat of war to be present. 2 Militarized Interstate Disputes (MIDs) are disputes between states in which at least one side threatens, displays, or uses military force (Jones, Bremer, and Singer, 1996). 6

7 Based on these arguments, we hypothesize that cultural similarities or differences influence the readiness of rival states to seek negotiated solutions to their disputes, and to do so successfully. We elaborate on this more fully below. Counter Arguments These arguments are by no means universally accepted by students of international negotiation. There are two prominent counter arguments, both of which are based on the notion of a global culture that overrides local national cultural similarities or differences. The first counter argument is related to our earlier point about the different sources of cultural influences on individuals, including one s occupation. Some scholars and practitioners argue that there is a global culture of diplomacy, which is universally understood by the elites who carryout interstate negotiations (Zartman, 1993; Nicolson, 1963 ). This is an argument that carried greater weight in the 19 th and early 20 th centuries, when most interstate diplomacy was conducted by trained diplomatists in Western states, many of whom had attended the same European universities. It is more problematic today, with the greater heterogeneity of the membership of the interstate system, as well as among those carrying out diplomacy, including individuals with far less, if any, experience in diplomatic practice. Anyone reading personal memoirs of American negotiations with representatives of Asian cultures, for example, would be hard pressed to argue that cultural differences played no significant role in negotiations. (Kissinger, 1979, 1982; Talbot, 2006). The second counter-argument is also based on the notion of a global diplomatic culture, but from a different perspective, that of political realism, particularly in the 7

8 classical tradition of practical realism, espoused by authors like Morgenthau (1978) and Kissinger (1994). It is not the niceties of diplomatic discourse that are the fruit of the global culture, but a shared understanding of the realities of power and the role that it plays in interstate relations, particularly when security interests are at stake. Viewed from a realist perspective, Henry Kissinger and North Vietnam s Le Duc Tho may have come from very different local cultures, but they spoke the same global political language, the language of Realpolitik, during their negotiations over the war in Vietnam (see Kissinger, 1979). Local cultural differences are trumped by the influence of a global political system s shared culture of Realpolitik. It is, of course, possible for interstate relations to be affected by both local cultural differences and a global culture of Realpolitik.. Kissinger and Le Duc Tho could be viewed as skilled diplomats, and practitioners of Realpolitik, whose relationship and ability to communicate could be also be affected by local cultural differences. The issue then becomes that of the relative potency of these variables. One plausible possibility is that, while local cultural differences might have an effect on negotiations where vital security interests are not at stake, they are trumped by Realpolitik when they are (Cohen, 1990). Specifically, the effects of cultural differences that one might observe in negotiations in disputes between rival states may become muted when disputes become militarized. Conversely, it can be argued that cultural differences can exacerbate the difficulties in resolving security issues. Henderson (1998) argues that strategic issues, such as territory, may become more difficult to settle because of the added effects of cultural antipathies and distrust related to cultural differences (Henderson, 1998). 8

9 Cultural Variables To investigate the question of the relative impact of Realpolitik pressures along with the broader question of the influence of cultural similarities and differences on negotiations between states engaged in enduring rivalries, it is necessary to specify which cultural variables are likely to exert the most potent effects. In the interest of parsimony, we have limited the putatively most significant cultural influences shared in common among citizens of a state to four variables. Two of these variables, religion and political culture influence the readiness of the parties to negotiate, as well as outcomes when they do negotiate. The other two, language and social culture, affect the quality of communication between the disputants when negotiations take place. We are interested in the relationship between the cultural attributes of those pairs. For example, in our examination of the effects of language, we are interested in whether the parties share the same language, as opposed to what particular languages they speak. We now turn to a brief description of each of the four culture variables, along with an explanation of its hypothesized effect on interstate negotiation. We borrow from Pruitt s (2005) readiness theory to identify two influences on the willingness of each party to negotiate with the other: motivation and optimism. 3 Pruitt s conceptualization of each party s motivation combines a realist conceptualization of a calculation of the likelihood of prevailing in the dispute and the costs and risks of continuing the conflict, with the degree of optimism regarding the likelihood of reaching an acceptable settlement with the other party. 3 We say borrow because we depart from Pruitt s (2005) approach in several respects, most notably, in our focus on the dyad, as opposed to the separate readiness of the two participants. Pruitt also includes another influences on optimism: the status of the negotiators that is, their capacity to reach an agreement that will be accepted by their governments. 9

10 Optimism is dependent on the tractability of the issues at stake, but also the party s perception of the other disputant, specifically its trust in the other as a negotiating partner (Pruitt, 2005; Zartman and Berman, 1982). States that distrust the other party are more likely to attempt coercive means of pursuing their interests and, if they do negotiate, are more likely to do so only after demonstrating their resolve through a show of strength (Leng, 2000). Conversely, those states that are optimistic regarding their ability to achieve a peaceful settlement with the other party are more likely to negotiate, and to do so earlier in the dispute. Confidence is likely to be affected by the party s degree of affinitiy/enmity, as well as trust/distrust of the other. As K. J. Holsti (1996:146) argues, realist assertions to the contrary, how states interact with each other is affected significantly by affinities or enmities based on moral judgments. We argue that optimism and confidence are influenced by similarities or differences in religion and/or the role of religion in politics, and political culture. Religion. Given the conflicts raging today, there is little doubt that religion plays a significant role in the relationships between states, as well as between sub-national and transnational groups. The most prominent exposition of this thesis is Huntington s Clash of Civilizatons? (1996). Huntington views a civilization as the largest possible cultural unit, and his primary cultural marker is religion. Aggregate data empirical studies, however, have failed to support Huntington s thesis, either as a predictor of interstate war (Henderson and Tucker, 2000) or civil wars (Gurr, 1994). 4 Part of the explanation for the failure to find empirical support for Huntington s hypothesis may lie in his attempt to 4 In an earlier study, Henderson did find a positive association between religious differences and war, although not according to Huntington s categorization. 10

11 map the globe according to competing civilizations based primarily on religion. Many countries are home to a variety of religious and non-religious groups, and many countries in which one can identify a majority religion are not in conflict with others over religious differences. We posit that if this cultural variable adversely affects relations between states, it does so in more limited circumstances, that is, between theocratic and secular states, and between theocratic states with competing religions or religious sects. 5 We make the theocratic-secular distinction on two grounds. First, theocratic states, unlike secular states, include their religion as an essential part of their identity. Identity issues, as Huntington points out, are not open to compromise or change. They are indivisible. Second, within theocratic states, there is no separation between religion and politics. Religious doctrine and practice are embedded in the fabric of the state s political culture. Thus, for the leaders of theocratic states, religious similarities or differences will affect not only their degree or affinity or antipathy toward other states, but also their approach to the management of interstate disputes. We articulate these arguments as a series of hypotheses: Hypothesis 1a. Disputants are more likely to use negotiation if they are both secular states; Hypothesis 1b: Disputants are more likely to use negotiation if they are theocratic states that share the same religion and religious sect; 5 Huntington s (1996) thesis, in fact, focuses on a subset of civilizational clashes in which religion plays a major role, including clashes between fundamentalist religious cultures, most notably Islam, and the secular West, although he extends the thesis to other potential civilizational clashes as well. 11

12 Hypothesis 2a: Disputants are more likely to resort to early negotiations if they are both secular states; Hypothesis 2b: Disputants are more likely to resort to early negotiations if they are theocratic states that share the same religion and religious sect; Hypothesis 3a. When disputants use negotiation, they are more likely to succeed in moderating or settling the dispute if they are both secular states or if they are theocratic states that share the same religion and religious sect. Hypothesis 3b. When disputants use negotiation, they are more likely to succeed in moderating or settling the dispute if they are both theocratic states that share the same religion and religious sect. Political Culture. Different political systems create different political cultures, including different ways of responding to and attempting to resolve disputes. The most frequently cited statistical finding regarding the relationship between political systems and interstate conflict has been the democratic peace, that is, the virtual absence of war between democracies during the modern interstate system (Small and Singer, 1976; Bueno de Mesquita and Lalman, 1990; Russett, 1993; Maoz and, 1993; Russett, Oneal and Cox, 2000). It is possible to make a broad cultural argument for the democratic peace based on political identity, specifically a shared self-image among democratic or liberal states (Brewin, 1991). The leaders of democratic states, it is argued, share an affinity for other members of the democratic family, which encourages them to settle their differences 12

13 through peaceful means. Moreover, they recognize that they share cultural norms of nonviolent behavior. Specifically, the normative/cultural argument for the democratic peace hypothesizes that disputes between democracies are more amenable to peaceful settlement because each government trusts the other s inclination to follow peaceful norms of dispute management through negotiation and compromise (Maoz and Russett, 1993; Dixon, 1994). That expectation, it is argued, does not obtain in disputes between authoritarian states, which have a political culture in which force plays a prominent role, or in disputes between democracies and authoritarian states because democracies, recognizing the difference in political cultures, do not trust authoritarian states to follow democratic norms of dispute settlement. A second explanatory processes of the democratic peace focuses on the structural impediments that derive from institutional constraints on the executive. The structural hypothesis is that shared decision-making powers, including institutional checks and balances, an active opposition party, and a free press constrain the government from capricious, sudden, and/or surprise attacks on its neighbors (Buena de Mesquita and Lalman, 1992; Morgan and Campbell, 1991). Presumably, leaders of disputing democratic states know that, in bargaining with each other, they enjoy the security of being able to seek a peaceful settlement through negotiation without the risk of a sudden, unexpected attack from the other party. However, the most extensive comparative test (Maoz and Russett, 1993) provides evidence that the democratic cultural norm of non-violent dispute resolution is a more robust predictor of the avoidance of dispute escalation than structural constraints. We 13

14 assume that cultural affinities among democracies is at least as important as their structural cousins, and as such focus our attention on these cultural attributes. Hypotheisis 4a: When the disputants use negotiation, they are more likely to succeed in moderating or settling the dispute if they are both democracies; Hypothesis 4b. When the disputants use negotiation, they are more likely to succeed in moderating or settling the dispute if their political systems share non-violent norms of dispute settlement. Language. Language is cited frequently as one of the primary attributes of cultural identity. Sharing, or not sharing, a common language can affect the attitudes of state representatives toward each other. During the Second World War, Winston Churchill, in his efforts to ingratiate himself with Franklin Roosevelt, continually stressed the natural affinity between the two great Anglophone nations (Kennedy, 1999:465). During the Cold War era, France s President Charles De Gaulle traveled to Quebec and fanned the flames of separatism by emphasizing the special relationship between France and the francophone majority in Quebec. The other side of close affinity that may grow out of a shared language, is the separation of self and other that is generated by language differences. Language differences can affect the quality of communication between state representatives, not only because of linguistic differences per se, but because of language s role as an expression of the cultural attributes of the negotiating parties. Cohen (1997) argues that language is the medium through which shared meanings are 14

15 transmitted throughout society, and, unlike most symbolic representations of culture, it is a dynamic medium that reflects social change, as well as retaining a country s historical identity. Language is a repository of [a] shared common sense directing a country s conduct of negotiation. It is the link between culturally embedded meaning and practice (Cohen, 1997:13). When disputants do not share a common first language, problems in understanding can occur even if the parties are communicating in a common second language, such as English. Hypothesis 6. When the disputants use negotiation, they are more likely to succeed in moderating or settling the dispute, if they share the same first language. Individualistic versus Collectivist Societies. Our second communication variable, a distinction between individualistic and collectivist societies, is drawn from research on intercultural communication (Triandis, 1995; Hofstede, 1991). The individualist/collectivist distinction represents the extent to which a culture emphasizes individual or group identity. A paralinguistic distinction is drawn between individualists and collectivists. Individualists are outcome-oriented negotiators who direct their attention to tangible outcomes or goals. Collectivists are process-oriented negotiators who emphasize the management of relationship-related and reputational interests, such as pride, dignity, trust, and respect (Ting-Toomey, 1999). Individualism/collectivism has been show to covary significantly with another prominent paralinguistic distinction in intercultural communication research: Hall s 15

16 (1977, 1959) high-context/low-context categorization. Individuals from low-context societies presume that the information of interest lies in the encoded message itself; highcontext negotiators are more attuned to contextual cues, such as the interpersonal relationship between negotiators and their nonverbal expressions (Cohen, 1997). Because it has a stronger empirical foundation, we have chosen individualism/collectivism as our indicator of cultural paralinguistic similarities and differences, but its covariation with the high-context/low-context distinction should tap the effects of both cultural indicators. We hypothesize that differences between state representatives on these dimensions will impair the quality of communication in interstate negotiations. Hypothesis 7: When the disputants use negotiation, they are more likely to succeed in moderating or settling the dispute, if they are both from either individualistic or collectivist societies. In sum, we propose, that all other things being equal, the first two cultural predictors, religion and political culture, will affect the readiness of the parties to use negotiation in attempt to settle their disputes, but that all four predictors: religion, political culture, language, and societal culture, will affect the outcome of any negotiations that do take place. Realist Variables. We noted above that the most compelling counter-argument to our general proposition comes from political realism, which suggests that a global 16

17 political culture, based on Realpolitik, will trump local cultural effects, particularly when vital security interests are at stake. Working from a classical realist perspective, the key variables of interest are (1) the gravity of the issues at stake, and (2) the comparative military capabilities of the rival states. These are the two most prominent variables in realist discussions of interstate interactions (Morgenthau, 1978), and dispute bargaining (Snyder and Diesing, 1977). Moreover, in a study of 40 militarized crises, Leng (1993) found that the combined effects of tangible interests at stake and optimism, on the part of both parties regarding the costs and risks of war, was positively associated with more coercive bargaining strategies. When issues are of vital importance to both sides, disputants are less likely to be optimistic with regard to the probability of finding common ground for a peaceful settlement (Pruitt, 2005), and hence less ready to pursue negotiations. Also, from a realist perspective, they are more likely to accept the costs and risks of the use of force to defend or pursue their interests. Thus, when vital interests, such as territory, are at stake, we would expect the disputants to be less willing to negotiate, and less successful in achieving a settlement when they do negotiate. The relationship between comparative capabilities and the willingness of the parties to seek a negotiated settlement is more complex. From a realist perspective, states should be more willing to use force in an attempt to achieve their objectives unilaterally when they are optimistic about their chances, that is, when they view the costs and risks as relatively low (Blainey, 1988). But, for both sides to be optimistic about their chances in a military conflict, at least one party has to misperceive the comparative military capabilities of the two sides, either through bounded rationality, or the withholding of 17

18 private information by one at least one side. 6 As Fearon (1995) has demonstrated, with complete information, there is no rational reason for the parties to choose war over negotiation. If they are evenly matched, the costs and risks of a military conflict should be too high for both states; if they are unevenly matched, the weaker should yield rather than accept the additional costs of a lost war. Thus, without a detailed historical analysis of the perceptions of the leaders in each of the disputes, we cannot predict readiness to negotiate on the basis of observable comparative military capabilities. But, when the parties do negotiate, classical realism would predict that, if negotiations did not fail, they would lead to a compromise settlement, or end in a stalemate when the parties are evenly matched in military capabilities, and that the weaker party would yield when there is an uneven balance in military capabilities. Hypothesis 8a. Disputants will be less likely to negotiate their disputes when they are geographically continguous; Hypothesis 8b. Negotiations are more likely to fail, or end in a stalemate, when they are geographically continguous; Research Design and Procedures Cases and Sources. To test these hypotheses, we have chosen a sample of 16 pairs of states engaged in enduring rivalries in the period between 1946 and Within each rivalry, we make a distinction between negotiations that occur within the context of 6 Leng s (1993) study estimated the optimism of state leaders through a close historical analysis of each of the cases in the study. 18

19 ongoing militarized disputes and those negotiations that occur during the more peaceful intervals occurring between disputes. This distinction allows for a more direct test of whether Realpolitik trumps cultural differences when the stakes are greater. The sample of rivalries appears in Table 1. TABLE 1: SAMPLE OF RIVALRIES Rivalry Dates Russia-Norway Afghanistan Pakistan China India India Pakistan Ethiopia Somalia Greece-Turkey 197?-2000 Iraq -Israel Syria -Israel Lebanon - Israel Jordan - Israel Saudi Arabia - Israel North Korea - South Korea Cuba - United States Ecuador United States Iran-Iraq Egypt-Israel The 16 rivalries include 430 negotiations, and 229 MIDs. These negotiations take place both within the MIDs and during the intervals between subsequent MIDs and provide the cases to test our hypotheses regarding the role played by culture in negotiations between rival states. Although our sample is overly populated with rivalries involving Israel and its neighboring state, we also include rivals involving four of the major powers, India and Pakistan as regional powers, and two Latin American rivalries involving the United States. We focus specifically on bilateral disputes. Enduring rivalries are defined operationally as bilateral relationships between states in which there are at least six militarized disputes occurring within 20 years 19

20 (Bennett, 1998). Militarized disputes are disputes in which at least one of the parties threatens, displays, or uses military force (Jones, Bremer, and Singer, 1996). These cases allow for a stringent test of the validity of the proposition that culture matters, even in disputes where vital interests are at stake. To obtain information on negotiations with our sample of enduring rivalries, we identified all bilateral disputes, which occurred between 1945 and 2000 involving these pair of rivals. From this list we generated data on negotiations occurring within them. Over the past four years we coded data on negotiations within these rivalries that incorporate security issues. Within the temporal confines of a MID we record information on negotiations pertaining to the issues related to the MID. During inter- MID intervals we record information on negotiations pertaining to national security issues. For each negotiation, we record the beginning and ending dates of negotiation, days on which the parties are actively engaged in negotiation, the affiliation and ranks of negotiators and any third party mediators, changes in negotiators or mediators, the issues under negotiation, and negotiation outcomes, as well as the cultural variables described above. The data are drawn from accounts in the international press, diplomatic histories, and memoirs of participants. These data on negotiations within inter-mid intervals are interleaved with data describing militarized disputes and any negotiations to allow for a temporal tracing of conflict and attempts to manage that conflict through negotiation. Data on the militarized disputes are taken from the Correlates of War project. Dispute attributes identified include: the beginning and ending dates of each dispute, the participants, magnitude of militarized behavior. 20

21 Out unit of observation in the data is the event-period. Our events are temporally bracketed, with each period stopping and starting based on the character of the event. For example, all rivalries begin with a militarized dispute that records the first event in the series. The end date of the MID provides the starting date for the inter-mid interval period, which ends with the onset of the next MID. If there are negotiations either within a MID or during the intervals, the temporal duration of each negotiation is recorded as a separate event-period. In effect we have a temporally defined event-based unit of observation that records negotiations within and between militarized disputes. The total number of inter-mid interval or MID event-periods in our data is 743, including 430 periods involving negotiations and 313 periods without negotiations. Outcome Variables. We are interested in whether cultural similarities or differences affect: (1) the willingness of the parties to seek negotiated solutions to their disputes, and (2) the success of negotiations that do take place. Initially we operationalize willingness to negotiate in terms of whether or not negotiations occur in the period and the time from the end of the MID to the start of negotiations. Negotiation outcomes are categorized according to an ascending order of success as: (1) full settlement of the dispute, (2) partial settlement of some issues, (3) no agreement on political issues, but moderation of the severity of the dispute (e.g., cease-fire), and (4) failed negotiations. Culture Variables. We record cultural attributes in terms of religion, political orientation, language, and social orientation. The first two culture variables, religion and political culture, may vary over the course of the rivalry. Afghanistan, or example, changes from a communist, secular state to a state under religious law under the Taliban; 21

22 Pakistan s political regime shifts from democratic to authoritarian on a number of occasions. For the religion variable, we distinguish between those pairs in which (a) the states are both secular, or are theocratic but share the same religion and religious sect, and (b) those pairs in which one state is classified as secular and the other theocratic, or both states are theocratic, but with different dominant religions or religious sects. We classify a state as theocratic when religious doctrine or practices are included in the state s legal system. The religion categorization is obtained by consulting documentary sources, as well as country experts. We employ two political culture indicators. The first indicator is a commonly used distinction between pairs of states that share democratic political systems, and pairs in which one state is democratic and the other is authoritarian, or both states are authoritarian. The democratic-authoritarian distinction is drawn from the Hewitt, Wilkenfield, and Gurr (2008) POLITY IV data on state attributes ( The POLITY IV data scores range from -10 (most autocratic) to +10 (most democratic). For each pair of rival states, we use the lowest of the two scores, the "weak link" specification, to determine the democracy score for the dyad. If the lowest democracy score is +6 or above, we code the dyad as sharing a democratic political culture. We make simple dichotomous distinctions based whether the rival states would be classified as sharing the same, or a different: language or social culture. We code the parties as the same on the language variable when the elites share fluency in a language, and also use that language in common discourse. For example, we code India and 22

23 Pakistan are sharing a common language, English, even though the traditional languages of the two states are Hindi and Urdu, because English is commonly spoken by elites in those countries. On the other hand, if the United States and Cuba were negotiating with each other, and the Cuban negotiator was fluent in English, we would code the languages as of the states as different. The distinction between individualistic and collectivist cultures is drawn largely from a comparative survey of 61 states by Hofstede (1991), who classified states according to one or the other of these two attributes. Hofstede did not include the then communist states in his study; consequently, we have made our own judgments in those few cases not fitting into his survey. Realist Variables. The two realist variables are continguity and comparative military capabilities. As a proxy for relative capabilities we record whether one of the participants to the rivalry was a major power, as defined by the Correlates of War project. There is some consensus that territorial disputes represent the most serious issues in interstate disputes, with the extant empirical research indicating a positive correlation between territorial issues and more contentious behavior (Vasquez, 1993). To tap this effect, we distinguish between those disputes involving contiguous actors and those that do not. Analysis and Findings Our results are necessarily preliminary. We begin by examining the distribution of our data with regard to our specific outcome variables. We recognize that the world is more complex than bivariate results betray particularly when culture is the focus but bivariate distributions can permit broad stroked inferences. Tables 1-8 present the 23

24 distribution of negotiation attempts by various cultural attributes. We hypothesized that negotiation patterns would differ across sects within rivalries between theocratic states, and our initial results lend tentative support to this inference. In table 1 we see that if a pair of countries in a rivalry are theocracies of they same sect they are equally like to negotiate or not. However this general pattern masks evidence that within the context of a militarized dispute theocratic states are less likely to engage in negotiations. Since the data are broken into periods involving militarized disputes and other less conflictual periods, one inference is that theocratic states actively involved in militarized hostilities with another theocratic state are not likely to enter negotiations, but during less conflictual times these same two states are more likely to find a way to use diplomacy to address their security concerns (Tables 1 & 2) Table 1 Joint Theocracy & Sect and Negotiations, complete sample theocratic, same sect negotiate 0 1 Total Total

25 Table 2 Joint Theocracy & Sect and Negotiations, MIDs Only theocratic, same sect negotiate 0 1 Total Total Common languages between rivals point to a considerably greater propensity to adopt negotiations as a form of conflict management than other pair of rivals, and this result appears to hold albeit less robustly within the context of a militarized dispute. Table 3 Common Language and Negotiations, complete sample common language negotiate 0 1 Total Total

26 Table 4 Common Language and Negotiations, MIDs Only common language negotiate 0 1 Total Total If pairs of rivals are of a similar social culture they are in general more likely to enter negotiations over security issues, but as with other cultural traits, within the context of a militarized disputes the tendency to negotiate goes away. Interestingly, when both parties to a rivalry have democratic political cultures they are more likely to engage in negotiations over security issues, except during periods of active military hostilities. In many ways this runs counter to the general notion that democracies are more attuned to and more likely to adopt negotiations as a way to avoid war. Table 5 Same Social Culture and Negotiations, complete sample same social negotiate 0 1 Total Total

27 Table 6 Same Social Culture and Negotiations, MIDs Only same social negotiate 0 1 Total Total Table 7 Joint Democratic Culture and Negotiations, complete sample both democratic negotiate 0 1 Total Total

28 Table 8 Joint Democratic Culture and Negotiations, MIDs Only both democratic negotiate 0 1 Total Total Table 9 presents a multivariate logistic regression of cultural variables on the existence of negotiations within and between MIDs. The basic question being asked is whether political and social cultural variables predict the use of negotiations when controlling for their simultaneous impact of other variables. Our results suggest that cultural variables influence considerably the willingness of the rivals to engage in negotiations, but only during the periods between militarized disputes (Model 1). However, only the level of hostilities is a significant predictor of negotiation within a MID, and in this case the more violent the dispute the less likely it is that we will observe negotiations. Cultural variables just don t seem to matter within the context of violent military hostilities. 28

29 Table 9 Logit Regression on Use of Negotiations* Model 1 Model 2 negotiate Coef. Std. Err. P> z Coef. Std. Err. P> z contiguous both democratic one democratic same social common language both theocratic both secular highest host constant *Model 1 is during Intervals Only; Model 2 is during MIDs Only Bold font p<.05; Italic font p<.10 Numb Obs Wald Chi Sq Prob > Chi Sq Log Likelihood During the intervals between MIDs, however, similar social culture leads to a greater propensity to use negotiations as does a dyad comprised of joint theocratic states. Jointly autocratic dyads involved in a rivalry are considerably less likely to engage in negotiations relative to mixed dyads, and jointly democratic rivals show no consistent pattern of negotiating during the inter-mid periods. Somewhat surprisingly neither common languages are not associated with an increased use of negotiations during the intervals between MIDs, and two secular states are more likely to use negotiation in the inter-mid period, but our level of confidence in this statistical result is weak (.10 level). Our results of bivariate distributions of the data suggest that the outcome of negotiations is not related to religious or social culture in any remarkable way. Moreover, political culture at least from the perspective of jointly democratic dyads is associated with outcomes short of success. For example, in the full population of cases jointly democratic rivals are much less likely to achieve a full settlement than other rivals. Of the 16 full settlements only one (6%) is a result of negotiations between jointly 29

30 democratic rivals, while 25% of the full settlements occur between jointly autocratic rivals. Table 10 Joint Democratic Culture and Negotiation Outcomes, complete sample both democratic outcome 0 1 Total Total

31 Table 11 Joint Democratic Culture and Negotiation Outcomes, MIDs Only both democratic outcome 0 1 Total Total We examine the role of political, social, and religious culture in predicting the outcome of negotiations by using a multinomial logit estimator with a multivariate model. Our outcome variable has four categories reflecting the degree of settlement that results from negotiations. While the results are not immediately striking, there are some interesting relationships that generally confirm our hypotheses. Of the four categories, the moderation of the dispute, often conceived of as a ceasefire, is the baseline outcome and the affect of our predictor variables on the other outcomes are judged relative to this baseline. For example, within the context of a militarized dispute, jointly theocratic rivals are more likely to have failed negotiations and jointly democratic rivals are considerably less likely to have failed negotiations than either would be to have negotiations reach a 31

32 ceasefire. But negotiations during intervals between MIDs appear to be unaffected by these cultural attributes. Table 12 Multinomial Logit on Outcome of Negotiations, MIDs Only Multinomial logistic regression Number of obs = 147 Log pseudolikelihood = Pseudo R2 = outcome Coef. Std. Err. P> z FULL SETTLE contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular highesthos~y _cons PARTIAL contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular highesthos~y _cons FAIL contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular highesthos~y _cons

33 Table 13 Multinomial Logit On Outcome of Negotiations, Intervals Only Number of obs = 280 Log pseudolikelihood = Pseudo R2 = (Std. Err. adjusted for 15 clusters in dyadcode) outcome Coef. Std. Err. P> z contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular _cons contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular _cons contiguous bothdemocr~c bothautocr~c samesocial commonlang~e boththeocr~c bothsecular _cons For clarity in interpretation we translated the estimated outcome into predicted probabilities, given a fixed set of conditions on our explanatory variables (Table 14 and 15). While we could provide predicted outcomes for each explanatory variable we use 33