Recent research on the motives of civil war has been dominated by the juxtaposition of greed versus grievance factors. This academic debate has generated a vast amount of empirical evidence to assess the explanatory power of the grievance concept but neglected its non-empirical limitations. In an attempt to direct attention back to the theoretical and methodological issues surrounding the grievance concept, the purpose of this analysis is twofold: first, I outline a grievance-based explanation of ethnic violence which argues that political institutions that provide low chances of political representation are likely to increase the odds of ethnic violence due to the intrinsic worth of political representation, its impact on the distribution of resources and powers, and its effects on perceptions of security. This explanation then serves, second, as a reference point to highlight three non-empirical limitations of the grievance concept, namely the lack of direct measures for grievances, the failure to identify clear reference categories of relative deprivation and the questionable assumption that all ethnic groups share common values. I conclude that due to these limitations, scholars working with the grievance concept inevitably remain in the realm of conjectures without being able to test the causal mechanisms they stipulate properly.