族群战争的爆发 : 一个广义理论 (The Onset of Ethnic War: A General Theory) 唐世平 ( 复旦大学国际关系与公共事务学院 ) 北京,2015 04 29
已经发表 即将发表和正在进行的工作 Shiping Tang, 2011. The security dilemma and ethnic conflict: toward a dynamic and integrative theory of ethnic conflict, Review of International Studies 51 (3) 王凯 唐世平,2013. 安全困境与族群冲突 : 一个因素 + 机制的分析框架, 国际政治科学 2013 年第 3 期 Shiping Tang, The Onset of Ethnic War: A General Theory, Sociological Theory (forthcoming). Shiping Tang and Hui Li, Location, Location, and Location: The Ethno geography of Oil and the Onset of Ethnic War, under review. 熊易寒 唐世平, 石油的族群地理分布与族群冲突的升级, R&R Hui Li and Shiping Tang, Different Oil, Different Conflict, in progress. Ethnopolitics, 1945 2010: A New Dataset (1945 2010 全球族群政治的数据集构建 )(PI: 唐世平, 参与者 : 王凯等 )
三步曲和目前的工作 Tang 2011: 指出这一远大目标的可能性, 并且展现这一个远大目标的可行性 ( 非系统性证据支持 ) 王凯 唐世平 (2013): 通过比较前捷克斯洛伐克的分裂和前南斯拉夫的分裂这两个典型案例 (pathway cases), 展现这一个远大目标的可行性 ( 系统性的证据支持, 非定量 ) Shiping Tang, The Onset of Ethnic War: A General Theory, Sociological Theory (forthcoming). Morespecific empiricalinvestigations: investigations: 构架一个全球性的 油气资源和族群冲突 的数据集 ( 基本完成 ) 在此基础上, 验证基于广义理论上的一个小理论 : 石油的族群地理分布才是真正决定石油是否导致族群战争的爆发以及导致族群战争的升级的关键 构建一个全球性的 族群政治指数 这个指数可以用来作自变量, 也可以用来作因变量 更重要的是, 这个指数有一定的预测能力 : 预测某个国家和地区爆发大规模族群冲突的可能性 接下来 : 进一步探讨族群冲突的爆发和延续
Ethnic Conflict: State of the Field I Overall assessment: the field has become extremely fragmented, lacking an integrated theory or analytical framework. More specifically: Most quantitative studies are only marginally theoretical, or even atheoretical (e.g., Collier and Hoeffler 2004). Most quantitative i studies are merely correlational:absence l of explanatory causal mechanisms Most quantitative studies assume explanatory factors to be independent from each other. Indeed, not only have existing studies of ethnic conflict often failed to integrate factors into a integrated theory, but also have tended to pit some factors or mechanisms against others (e.g., Gagnon 1994 95; vs. Arfi 1998) Most studies, especially quantitative ones, have failed to differentiate initialfactors and intervening regulatory factors from political processes that work together to drive ethnic conflict. Yet, processes such as manipulations of ethnic identity and hatred by elites have often been one of the crucial processes that drive ethnic conflict.
Ethnic Conflict: State of the Field II Ei Existing i studies, especially ill quantitative i ones, have tended dd to overemphasize tangible (i.e., material) factors, including tangible interests (e.g., oil, gas, diamond, territory, measureable ethnic power relationship), as the drivers of ethnic conflict. As such, they have implicitly adopted a largely materialistic instrumentalism approach toward ethnicity and ethnic conflict. Such an approach, however, is utterly misleading, as Connor (1984[1994]) pointed out long ago. Existing studies of ethnic conflict have failed to integrate factors into an integrated theory (Sambanis 2003; 2004; Horowitz 2008). Indeed, most studies tend to pit their favored factors or mechanisms against others. Qualitative studies, sometimes focusing on a single case or several cases, tend to make sweeping generalizations that cannot be supported.
Toward A More Integrated (General) Theory I conjecture that t a key cause behind this state t of the field has been that we simply have yet to find appropriate theoretical tools that can bring all the factors and processes under the same roof. I thus advance a possible solution to this challenge. Critically building upon earlier attempts of synthesizing (e.g., Horowitz 1985; 2001; Fearon and Latin 2000; Petersen 2002; Kaufman 2001; Lebow 2008), I seek to take several key steps toward a general theory of ethnic conflict, with two analytical innovations. First, I start with mechanisms rather thanfactors as almost all existing studies have done. More specifically, I deploy two meta mechanisms that drive intergroup conflict, namely, the security dilemma/spiral and intergroup intragroupintragroup interaction. I then go back to comb through the many factors that have been identified to contribute to ethnic conflict by asking this question: can a factor be reasonably linked with ethnic conflict through the two meta mechanisms? By so doing, I shall be able to link most of the relevant factors with ethnic conflict, both onset and duration. Second, I explicitly identify potential interactions of the various factors and mechanisms. The resulting framework is a more integrated theory of ethnic conflict.
族群间的安全困境 : 既有研究的问题 安全困境 (security dilemma) 是国际政治中的一个核心概念和核心理论 但是, 长期以来, 它的意义和潜在的解释能力没有被认识到 Posen(1993) 年首先将 安全困境 作为一个分析工具引入族群冲突研究领域, 但是, 长期以来, 研究族群冲突的学者对这个分析工具的使用仍然存在几个显著的问题 特别是 : 将安全困境等同于冲突螺旋 (e.g., Posen 1993; Kaufman 2000; Roe 2004) 或者将安全困境错误地简化成 承诺问题 (Fearon 1995; Lake & Rothchild 1999; Walter 2001; Toft 2003) 仅仅将安全困境作为一个前提条件, 一个自变量 这是错误的静态的 理解 主要解释变量通常是单一的 比如, 某些研究只探讨 族群的分布 对族群冲突的影响 而另一些研究只讨论恐惧的作用 ; 在一系列的著述中,Tang (2009;2010;2011) 纠正了这些错误 更重要的是,Tang (2011) 强调, 安全困境机制 ( 理论 ) 不仅能够整合 无政府状态下的相互恐惧, 而且可以将资源 地理 历史 国家制度安排 政府政策以及政治精英的角色等因素, 通过四个心理学的桥梁 ( 利益 恐惧 尊严 仇恨 ) 融入进一个整合的 机制 + 因素 解释框架 这样的 安全困境 / 机制 + 因素 分析框架将扭转族群冲 突研究领域 碎片化 的趋势, 进而整合所有和族群冲突有关的因素, 形成一个有机的整体, 从而构建一个广义的族群冲突理论
迈向一个族群冲突的广义理论 Tang (2011) 强调, 安全困境机制 ( 理论 ) 不仅能够整合无政府状态下的相互恐惧, 而且可以将资源 地理 历史 国家制度 安排 政府政策以及政治精英的角色等因素, 通过四个心理学的桥梁 ( 利益 恐惧 尊严 仇恨 ) 融入进一个整合的 机制 + 因素 解释框架 这样的 安全困境 / 机制 + 因素 分析框架将扭转族群冲突研究领域 碎片化 的趋势, 从而构建一个广义 的族群冲突理论 这样的一个理论可以整合所有和族群冲突有关的因素, 形成一个有机的整体 基本研究设计 : 以 安全困境 的机制为主轴, 而不是以因素为出发点 先对因素进行分类和分层 ( 同时进行 ), 探讨某一个因素如何影响 ( 利益 恐惧 尊严 仇恨中的 ) 哪一个 motive, 然后通过影响 ( 能力 利益 决心 ), 进而推动族群冲突 在这个过程中, 用安全困境机制把这些因素 ( 通过心理学的桥梁 ) 连动起来 王凯 唐世平 (2013) 安全困境与族群冲突 : 一个因素 + 机制的分析框架 可以是是这个广义理论之前的一个 pilot study
Dynamic Interactions of the Drivers and Ethnic Mobilization i by Ethnic Elites 利益 不满 恐惧 仇恨这四种心理驱动力存在相互增强的效果, 因为这四者都暗含了 期望改变 的心态 且这四个驱动力可以相互加强 因此, 任意一种驱动力的出现或增强, 都会导致其它业已存在的驱动力的增强 则族群间的安全困境越容易被恶化 在每一个族群内部, 都会有一部分精英试图通过动员族群, 来获取政治权力和地位 而每一个族群内部的族群动员, 都会被另一个族群的精英和民众看成是这个族群 图谋不轨 的证据和迹象, 从而加剧两个族群之间的 安全困境 或者是 不安的螺旋 (spiral of insecurity) 因此, 如果我们能够把文献中已经发现的因素或者还未被发现的因素, 通过 安全困境 和 族群动员 联系起来, 我们就向一个族群冲突的广义理论迈出了决定性的一步
安全困境 : 整合心理因素 心理驱动力 (psychological ldi drivers) 和物理驱动力 (physical ldi drivers) 相互作用, 共同调节安全困境, 从而决定安全困境和冲突螺旋是被恶化还是被遏制 心理因素可以影响利益的判断 决心和对手段的偏好 (interest, resolve, and intentions as preferences over means) Hatred: 集体仇恨 能很快推动冲突升级, 任何暴力行动都可以挂上 集体复仇 的正当理由 Resentment(honor, prestige, underpinned by prejudice & discrimination/e.g., EPR ): 群体不满 反映了群体对社会资源的分配方式或结果感觉到不公平, 并且希望对此进行改变 两种 不满 : 对中央政权 对其他族群 Fear (fear of annihilation): 如果族群之间的 恐惧 长期得不到缓解或者被进一步增强, 族群为克服 恐惧 而进行的实力竞争加剧, 族群间的安全困境更加容易恶化成冲突螺旋 但是, 恐惧本身难以导致冲突 (Tang 2009; 2010; 2011)
安全困境 : 整合物质因素 Capability (opportunity):allies, terrain, resources (oil/gas, diamond) 族群需要武装能力\ 财富 \ 地域聚集才能进行冲突 Interest (physical):size of territory, terrain, control of resources/money, drug production/trafficking 因此, 物理资源 ( 特别是比较集中且很容易套现的 commodity 性资源, 比如 oil/gas, diamond) 有多重影响 : 既影响 利益 又影响 能力, 从而影响 决心 值得注意的是,Interest (psychological/symbolic) personal ambition, desire to revenge, desire to correct wrongs, g, nation statehood etc can be understood to be symbolic interests.
Toward a General Theory: Two Building Blocks Kaufman (2001) and Petersen (2002) made important progresses toward a more integrated theory. Even these two works, however, suffer from serious shortcomings. We critically build upon them and other works. Security Dilemma/Spiral Model (SD/SM) The security dilemma (SD) is a key concept and dynamic theory of conflict in international/intergroup/interpersonal relations. SD/SM is a potent tool for integrating material and psychological factors that because SD/SM explicitly i l contends that physical and psychological factors can act as regulators of SD/SM and that final outcomes depend on the specific mixtures of these regulators. Inter Intra Group Interactions Intra group interaction between elite and mass is a key to the understanding of ethnic conflict and this intra group interaction is constantly under the influence of the dynamics of inter group interaction At the epistemological level, inter intragroup interaction can subsume the rational choice approach and a social constructionist /symbolic interaction approach.
Processes: Ethnic Mobilization and Elite Mass The idea that elite manipulation is one of the keys to intergroup conflict can be traced back to Simmel (1955[1908]) and Coser (1956). Afterall all, ethnic conflict would be extremely difficult, ifnot impossible, without some kind of organizational support, and to build this organizational support, elites and mass must work with each other. The heart of ethnic mobilization is to make group members willing to sacrifice for a group and to inflict violence and atrocities upon the other group: ethnic mobilization bl is to mobilize bl politics largely, l if not exclusively, alone ethnic lines (Mann 2005; Collins 2012). Intra group tactics (e.g., deploy mass media to paint conciliatory policies, especially those proposed by opposing elites within one s own group, as dangerous, if not traitorous) and Inter intra group tactics [e.g., Deploy mass media to construct and fan ethnic fear or insecurity (i.e., our group is in danger of being enslaved or annihilated) by painting the other group as inherently hostile and imminently threatening via mass media.]
Table I. From Underlying Factors to Master Drivers Master Drivers Specific Underlying Factors Emotion: fear, honor, hatred, and anger Tangible Interest/Greed Opportunity/feasibility (internal and external constrains) Fear (insecurity): Collapse of central authority; hijacking of the central authority by one group; withdrawal of colonial power leading to the fear of ethnic annihilation/subordination, especially with ethnic mixing and earlier episodes of violent il conflict; nationalizing i policies i by a dominant group; ifl influx of ethnic hi aliens; relative demographic decline. Honor (resentment/grievance): Ethnic solidarity; domination (subordination, exclusion, discrimination, repression); influx of ethnic aliens; economic inequality (real or perceived); redistribution ib i policy. Hatred: Earlier episodes of violent conflict, especially ethnic war and cleansing Anger, esp. Rage/Fury: Local brutality; natural disasters that stuck a group. Demand of equality and equal opportunity; demand of autonomy (cultural, economic, political); i l) demand d of secession/independence; d accesses to state bureaucracy and military rank-and-files; natural resources, esp. in the territory of the subordinate group; economic inequality & redistribution policy Collapse of central authority; political instability at the centre/regime transition; economic crises; a weakened state for whatever reason (e.g., by earlier episodes of violence, in-fighting within the ruling elite; defeat in an interstate war); contagion/diffusion of conflict; other political contexts (e.g., regime type, civil society; democratization/decentralization) Capability/Power Access to weaponry; military organizations i and combat experiences from earlier conflicts; absolute size of group & relative size of the group within the whole population; proportion of young men [within the subordinate group]; Support by external allies, either state or diasporas; distance between the rebelling region and the central government; terrain of the subordinate group; natural resources within ihi the subordinate group; price of primary commodity goods; Total GDP (of the state, central government); overall state capacity weakened by earlier conflicts.
Table II. The Power of an Interactive Approach: An Illustration Majority domination/ minority in subordination: Minority domination/ majority in subordination Outcomes Can be stable or unstable, Highly unstable. The subordinated without other separatist/secessionist movement possible majority group will seek outright factors independence, if not outright taking over of the country. Discovery of oil Very likely to be a powder-keg. Extremely likely to be a powder-keg. and other One the one hand, the subordinate minority One the one hand, the subordinate substantial natural resource within the core territory of the group will seek some kind of re-distribution of resources. It may seek external support. On the other hand, the dominant majority majority group will seek outright independence, and it can easily draw outside support. subordinate group group will seek control of the resources. On the other hand, the dominating minority group will also seek control of the resources. Discovery of oil Implications for onset: Stabilizing, if not Implications for onset: extremely risky, and other actually strengthening the status quo because the dominant (minority) group now has substantial natural it provides more resources for the majority more resource for repression whereas resource within the group to dominate and fight, if necessary. the subordinate (majority) group has core territory of the dominant group Implications for duration: If war breaks out, war tends to be bloody but short (the majority group has more resources to fight). more reason to resent and rebel. Implications for duration: If war breaks out, war is likely to be bloody and long partly because the teminority otygoup group can gain external support by selling booty futures.
实证 : 石油的族群地理区位与族群冲突 石油的族群 地理位置或者分布 (the ethno geography of oil, especially its location) 才是决定石油是否导致或者加剧族群冲突的关键要素 具体地说, 在一个国家内部, 只要少数族群聚集区域有一定数量的石油资源, 那么这个少数族群就容易和中央政府发生矛盾 而如果这个少数族群此前就和控制中央政府的多数族群有过流血冲突因而仇恨的话, 那么这个少数族群和多数族群之间就有可能发生族群冲突 而如果在少数族群聚集区域中的石油是在冲突过程中被发现的话, 石油的发现将加剧既有的族群冲突 而如果少数族群聚集区域没有石油 ( 这时候, 多数族群聚集区是否有石油也不重要 ), 或者少数族群和多数族群是混杂居住的, 则石油对一个国家内部的族群冲突没有影响
图 1: 石油的族群地理分布与族群冲突的爆发或升级 : 驱动力与机制 中央政府 ( 多数族群主导 ) 少数族群 在少数族群核心领地发现油田 ( 经济冲击 ) 内部殖民化 : 国家加强对少数族群地区的掠夺, 中央集权化, 增加新的控制手段或强化现有的控制手段, 汲取石油资源的收入但少数族群只能分享极少的财政份额, 鲜有环境保护 双方的利益升级 ( 贪婪 ) 少数族群与多数族群的收入差距拉大, 外族群移民进入, 环境污染恶化 少数族群的相对剥夺感加剧 ; 族群边界变得显著 ( 族群政治动员 ); 怨恨升级 历史积怨 仇恨与愤怒加剧 ; 双方均利用石油收入扩军备战 族群冲突爆发或升级
四个正面案例和一个负面案例 四个正面案例 : 苏丹 南苏丹 印尼 亚齐 伊拉克 库尔德 尼日利亚 尼日尔河三角洲 之所以选择这四个案例, 不仅因为它们都盛产石油天然气, 而且长期以来饱受族群冲突之苦 ; 更重要的是, 它们分布在非洲 中东和南亚, 国家规模 族群结构 人口规模和地理环境存在较大差别, 构成了最大差异案例 (most different cases) ) 如果在差异如此之大的案例中, 我们都可以发现上述机制的存在, 那么说明我们的理论是强有力的 结论 : 油田区位假说得到上述案例的支持 : 假定一个国家业已存在族群冲突, 少数族群地区发现大型油田会加剧冲突的烈度 ; 假定一个国家不存在族群冲突, 但存在一定的仇恨, 少数族群地区发现大型油田可能促使冲突爆发 推论 : 石油探明储量增长会延长冲突的持续时间 ( 南苏丹 尼日尔河三角洲 库尔德 ), 而石油资源的逐渐枯竭会使冲突趋于缓和 ( 亚齐 ) 负面案例 : 加蓬 尽管族群林立, 石油丰富, 但是, 因为没有少数族群的聚集, 石油不容易和族群政治联系在一起, 因此, 迄今为止, 族群和平相处
The Ethno Geography of Oil and Ethnic Conflict: Some Quantitative Evidences: Key Theses: Discovery of oil and other substantial resources within a subordinate group is very likely to be a powder keg for ethnic conflict onset and may prolong existing conflict. The reason is that oil and other substantial resource within the subordinate group impact at least four of the five immediate drivers: fear (due to internal colonialization), anger/resentment, t tangible interest, t capability. In contrast, discovery of oil and other substantial resource within the dominating group has little impact on ethnic conflict onset. And if war does break out, war tends to be bloody but short (the majority group has more resources to fight).
Fig. 1: Oil/Gas and Ethnic Conflict: Descriptive
Table 1: The Ethno-geographical Location of Oil and the onset of ethnic conflict, group-level, with USGS dataset (Dependent variable is onset of ethnic conflict experienced by a group between 1945 and 2005) Model (1) Model (2) Ethonset (EPR) Ethonset (ACD) territorial governmental Oil location (USGS, group) 0.48** 0.93** 0.26 (0.24) (0.40) (0.41) Cap distance 0.15** 0.25*** 0.089 (0.069) (0.093) (0.094) Border distance -0.15** -0.24** -0.031 (0.065) (0.099) (0.074) excluded 1.28*** 1.07*** 1.40*** (0.28) (0.41) (0.36) downgraded2 1.75*** 1.20** 2.04*** (0.38) (0.55) (0.46) lrsize 0.27*** 0.10 0.59*** (0.087) (0.11) (0.12) War history 0.71*** 0.51** 1.02*** (0.16) (0.26) (0.24) GDP cap(log) -0.37*** -0.33-0.38*** (0.13) (0.22) (0.14) Population(log) -0.11-0.031-0.30** (0.12) (0.15) (0.14) Peace years -0.10 010-0.17* -0.054 0054 (0.075) (0.087) (0.13) gspline1-0.00023-0.00058 0.000079 (0.00059) (0.00065) (0.0011) gspline2-0.00016-0.00000082-0.00038 (0.00038) (0.00042) (0.00076) gspline3 0.00023* 0.00021 0.00035 (0.00014) (0.00016) (0.00031) constant -1.75-4.11** -0.19 (1.32) (1.91) (1.71) Pseudo R 2 0.11 0.13 N 19749 19749 Wald chi 2 147.9*** 442.0*** Note: model 1 is logit model and model 2 is multi-nominal logit model; Standard errors in parentheses;* p<.10, ** p<.05, *** p<.01
Table 2: The Ethno-geographical Location of Oil and the onset of ethnic conflict, country-level (Dependent variable is onset of ethnic conflict experienced by a country between 1945 and 2005) Model (1) () Model (2) () Model (3) () Model (4) () Ethonset (EPR) Ethonset (EPR) Ethonset2 (EPR) Ethonset2 (EPR) Oil location (USGS, country) 0.78*** 0.75** 0.79** 0.73** (0.29) (0.32) (0.34) (0.36) lrexclpop 0.36*** 0.25*** 0.38*** 0.30*** (0.084) (0.080) (0.097) (0.099) egipgrps 0.15*** 0.098** 0.17*** 0.13* (0.032) (0.040) (0.043) (0.067) pimppast 0.37 0.68* 0.90* 1.10** (0.40) (0.40) (0.53) (0.52) gdpcapl -0.12*** -0.10** -0.084*** -0.062* (0.035) (0.040) (0.031) (0.035) lpopl 0.15** 0.13* 0.098 0.095 (0.066) (0.068) (0.077) (0.086) ongoingwarl 0.21 0.18 (0.57) (0.59) year 0.010 0.0079 0.017*** 0.014** (0.0062) (0.0063) (0.0061) (0.0067) ethfrac 1.30*** 1.06 (0.48) (0.67) lmtnest t 019** 0.19** 0.096096 (0.092) (0.10) regchg3 0.12 0.29 (0.25) (0.26) anocl 0.30 0.39 (0.22) (0.28) oilpcl 0.0049-0.000095 (0.037) (0.038) Constant -27.22** -23.70* -40.47*** -36.29 (12.3) (12.5) (12.1) (13.2) Pseudo R2 0.13 0.14 0.12 0.14 N 7661 7566 6863 6773 Wald chi2 243.6*** 190.6*** 115.4*** 145.7*** Note: logit models; model 3 and model 4 drop ongoing wars; Standard errors in parentheses;* p<.10, ** p<.05, *** p<.01
Table 3: The Ethno-geographical Location of Oil and the onset of ethnic conflict, country-level (Dependent variable is onset of ethnic conflict experienced by a country between 1945 and 2005, differentiated into two sub-categories: ethnic infighting and ethnic rebellion) Model (1) Model (2) infighting rebellion infighting rebellion Oil location (USGS, country) 0.050 0.90*** 0.18 0.94*** (0.50) (0.34) (0.58) (0.37) lrexclpop -0.036 0.49*** -0.21 0.46*** (0.15) (0.089) (0.16) (0.089) egipgrps 0.31*** 0.040 0.30*** 0.049 (0.062) (0.041) (0.063) (0.047) pimppast 2.57** 0.043 3.05*** 0.13 (1.04) (0.40) (1.17) (0.44) ethfrac -0.18 1.38*** 1.08 1.33*** (1.18) (0.53) (1.02) (0.50) gdpcapl -0.16** -0.090** -0.15** -0.092** (0.081) (0.035) (0.076) (0.041) lpopl -0.28** 0.32*** -0.41*** 0.30*** (0.12) (0.076) (0.12) (0.086) ongoingwarl 1.09 0.065 0.93 0.065 (1.05) (0.67) (0.99) (0.68) year 0.029** 0.0034 0.024* 0.0037 (0.013) (0.0069) (0.013) (0.0072) lmtnest 0.47*** 0.093 (0.18) (0.11) regchg3 0.039 0.14 (0.57) (0.29) anocl 0.36 0.36 (0.40) (0.23) oilpcl 0.014 0.0070 (0.026) (0.047) constant -63.89** -16.55-54.22** -17.38 (26.7) (13.5) (26.2) (14.1) Pseudo R2 0.16 0.17 N 7661 7566 Wald chi2 320.3*** 330.8*** Note: multi-nominal logit models; Standard errors in parentheses;* p<.10, ** p<.05, *** p<.01
方法论的思考和贡献 我们对族群冲突的研究充分体现了我们对方法论的理解 首先, 我们的实证假说建立在好的理论之上 其次, 我们定量假说包含因素 ( 含时空 ), 特别是因素之间的相互作用 再者, 我们的定性研究的假说包含了三个部分 : 因素 ( 含时空 ) 因素相互作用 机制 而因为只有定性研究可以解决机制问题 ( 而这也是定性研究的核心优势之一 ), 我们的定性案例研究特别试图展现我们认定的机制的作用 最后, 我们的结果, 无论是定量还是定性, 都非常明晰 因此, 我们认为我们的这项研究充分体现了是定性 定量以及其他方法的结合的方法论优势
缅甸的族群冲突 ( 战争 ): 导因 I 去殖民地运动之后兴起的许多现代民族国家都试图把一个国家变成一个 大族群 ( 优势族群 ) 据有的国家 ( nationalizing state ; Brubaker 1996) 而这样的国家通常容易族群冲突 缅甸属于这一类国家的典型 ( 类似的马来西亚 印尼 越南等 ) 具体地说, 这些 nationalizing state 的主要做法是 : 在中央层面, 将国家机器转化成为优势族群的 ( 国家 ) 机器 主要部门, 特别是强力部门 ( 军队 警察 内部安全 ) 在地方层面, 让国家机器最大程度地渗透到少数族群内部 通过向少数族群地区的移民政策, 试图进一步挤压少数族群的生存领地 ( 内部殖民化 ), 同时达到控制少数族群地区的财政 ( 含自然 ) 资源等等 从政治地理上, 将原本的少数族群领地分成几份, 制造促成少数族群内部分裂的 地方小认同 从政治上分化少数族群内部, 以更好的控制少数族群 将优势族群的文化 ( 文字 语言 ) 变成官方文化, 并且强迫少数族群接受 ( 胁迫下的同化 ) 不幸的是, 所有这些举措, 都非常容易激化少数族群对多数族群 ( 控制的中央政府 ) 的不满和仇恨, 因而非常容易导致族群冲突或者加剧已有的族群冲突
缅甸的族群冲突 ( 战争 ): 导因 II 按照我们关于族群冲突爆发 ( 及持续 ) 的广义理论, 缅甸族群冲突多发且不易解决非常容易理解 缅甸的许多少数族群都有相当强烈的族群自我认同 缅甸的许多少数族群都有明显的核心聚集区 ( 或者说是他们认定的 家园 ), 且这些核心聚集区处于国家的边界地区 许多靠近边境的少数族群核心聚集区都有丰富的自然资源 ( 如, 翡翠 森林 ) 某些少数族群核心聚集区内有相对规模的毒品生产基地, 是的这些少数族群能够有实力进行战争 这些靠近边境的少数族群核心聚集区很多都是丛林 崇山峻岭的地形地貌 这高度利于通过游击战和运动战防守, 而不利于政府军的 围剿 以及移民的留驻 历史上, 这些少数族群和缅甸族都有过大规模的流血冲突, 双方积怨已深 在这样的基础上, 缅甸中央政府的 nationalizing state 的政策只会进一步激化少数族群对多数族群 ( 中央政府 ) 的不满和仇恨, 因而非常容易烽烟再起 缅甸内部的民主化有可能进一步激化族群矛盾和冲突