在对发展中国家的政治研究中，存在着“政治发展—政治稳定”的二元分析框架，政治发展的讨论主要围绕民主与专制展开，政治稳定的讨论主要围绕秩序与动荡展开。然而，对于已经实现民主化并且一定意义上实现民主巩固的拉美国家来说，介于政治稳定与政治动荡之间的政治变化，欠缺充分的学理解释。政治变化的讨论主要关注政党及领导人的上台和下台，在对民主化以来拉美国家政治变化的研究中，“向左转”、“左退右进”、“左右并进”这样带有意识形态色彩的标签被频繁使用。所谓“左转”和“右转”，共同点是“转”，都是关键性选举引发的政治变化，而左右之分越来越不足以厘清其中的差异，反而因界定左右的困难给研究造成负累。本文将关键性选举引发的政治变化作为主要研究对象，尝试跳出左右之争，探讨决定政治变化的各类因素，对民主化以来拉美国家政治变化的多样性做学理分析。基于对现有解释的综述，尤其是对跟从单纯利益逻辑和带有结构决定倾向的观点的批评，本文选择在历史制度主义的方法论指导下，建立以制度过程为核心关切的解释框架，将拉美国家的政治变化视作结构因素和能动者因素互动下的政治认同建构的结果。在研究设计上，结构因素选择以政治参与和政治吸纳能力的相对关系为指标的政治衰朽程度作为变量，能动者因素选择以政治认同建构主体，即关键性选举中的政治挑战者，建构策略的选择作为变量，交叉得到四种假设，理论上推演出政党轮替、政党重组、政治重组和政权更迭四种类型的政治变化。选择符合假设条件的典型案例进行检验，具体指对智利、哥伦比亚、玻利维亚和秘鲁四个国家由关键性选举引发的政治变化做了比较研究和综合分析。本文的主要发现是，民主化以来拉美国家政治变化的实质是多样性的政治认同的变化，与意识形态的左与右不必然存在对应关系。从过程看，关键性选举中政治挑战者的策略选择对于推动政治认同变化发挥了决定性的作用，从结果看，政治变化类型的多样性受制于政治衰朽程度。本文建立的解释框架不排斥意识形态对政治变化的影响，但是认为执着于左右的标签难以把握拉美政治的变化逻辑，故本文的创新之处在于超越左右，从中层理论的视角考察拉美政治变化多样性的动因，尝试对制度因素和人的因素在其中的互动进行比较政治学意义上的分析。

Studying politics in developing countries does not go too far from a binary framework of analysis, namely “development – stability”. Political development theories try to tell democracy from autocracy while debates over political stability mainly focus on political order or the lack of it. However, for the Latin American countries that have already achieved democratization and, to a certain extent, democratic consolidation, the political changes relatively less stable remain to be further explained from the academic point of view.The rise and fall of political parties and political leaders are the main concern in the discussion of political changes. A bunch of ideological labels are frequently applied when describing political changes in Latin America since the third wave of democratization, for example “the Left Turn”, “the Right Turn” or “the Right Striking Back”. The common ground here is to turn, which refers to political changes brought by critical elections. But the increasingly blurred boundaries between the Left and the Right has compromised its validity in accounting for the divergences. The goal of this thesis is to theoretically treat these divergences beyond ideological judgement, going through the various factors that determined political changes of “turned” nature.Critical of the tendency to lean on structural determination, this thesis choses the historical institutionalist approach that places institutional processes at the center of analysis. Critical of the purely functionalist view of politics, the main argument of this thesis sees political changes as result of the construction of political identity through the interaction between structure and agency factors. In terms of research design, two variables are presented in the model. One is the degree of political decay, measured by the relative ratio of political participation and inclusivity. And the other is the strategy adopted by political challengers in critical elections, as key agents in constructing political identities. Four types of political changes are thus hypothetically derived from this matrix, a) party alternation, b) party realignment, c) political realignment, and d) regime change. The research design is followed by a comparative study of four cases – political changes via critical elections in the four countries of Chile, Colombia, Bolivia and Peru.The main finding of this research is that diversified changes in political identity, rather than ideological affinities, define political changes in Latin America since democratization. The strategic choices of political challengers in critical elections played a decisive role in constructing changes in political identity, while the variety of political changes is subject to the degree of political decay. Hanging on to those ideological labels contributes in limited ways to the understanding of changing politics in Latin America though the explaining framework does not exclude the impact that ideologies may very well still have over political change. The highlight of this thesis is the efforts in trying to set up a mid-level theoretical framework to examine the diverse political changes in Latin America, expanding analysis of the interactions between institutions and key figures in politics.