Abstract
The recent events of the Arab Spring have re-focused the world’s attention on the power of nonviolent action, demonstrating in a dramatic fashion that unified people power movements can overthrow seemingly-entrenched dictators in a very short period of time. A sizeable body of research on nonviolent action seeks to understand how, why and where it can be effective in altering structures of oppression, inequality and marginalisation. Detailed case study analysis has been made of most of the major campaigns of nonviolent action in the twentieth century, showing that it can be a powerful method of collective political action. There is also a small but growing interest in using quantitative methods to research nonviolent action, exemplified by Karatnycky and Ackerman’s (2005) study of its role in democratic transitions and Chenoweth and Stephan’s (2011) comparison of the success rates of violent and nonviolent conflicts. However, much of the research on nonviolent action to date has been affected by selection biases in favour of large, well-publicised campaigns. There has been little attempt beyond this to define the empirical boundaries of unarmed conflict. The consequence of this is that what we know about nonviolent action is limited to large campaigns; we know little about the “nonstarters,” the campaigns that are too small, peripheral or unsuccessful to draw attention (Chenoweth & Stephan, 2011, 15). Knowledge is especially lacking on what impact extremely violent repression has on smaller opposition movements (Carter, 2009).
This thesis asks the question: How effective are nonviolent tactics in the face of severe repression? In order to answer this I adopt a novel approach to studying nonviolent action, by shifting the analytical focus from campaigns to tactics. By making the contentious tactical interaction between civilians and the state the focus of enquiry, data can be gathered from news sources, NGO reports and academic accounts in a systematic way, using theoretically-grounded definitions and without the selection biases affecting previous work. This has generated a new dataset on nonviolent tactics met with extreme government repression: the State Violence against Unarmed Protests 1989-2010 dataset. This includes 38 cases of severe government repression of civilians using nonviolent tactics, 18 of which have not been recorded in previous empirical work. Nonviolent tactics are shown to be capable of being successful even in the face of extreme violence, with 26.3% resulting in success for the civilian. The key factors associated with the outcomes of these cases are shown to be the aspirations of the opposition group, and whether the conflict is divided along ethnic lines. It is concluded that this analytical framework is an effective means of gathering empirical data on nonviolent action, and can complement existing research.