Lack of trust has been widely used as an explanation for the failure of peace negotiations. However, we know little about how mistrust can be reduced between belligerents involved in negotiating peace. Why are some confidence-building strategies more successful than others? For theory-building purposes, this article explores how a party can send conciliatory signals to the other party that increase trust by exposing itself to three different kinds of political risks. More specifically, it compares the variables that reduced mistrust -or failed to reduce mistrust -during two peace negotiations in Sri Lanka: in 1994Lanka: in -1995Lanka: in and in 2002. Using a theoretical framework that combines social psychology and rational choice approaches, this article examines the communicative signaling process between the parties.In addition, by drawing out the implications from this argument, we offer some insight into why the peace process in Sri Lanka became politically stalemated in 2003. We also use our comparison of Sri Lanka's peace processes to develop general propositions about the dynamics that can reduce mistrust. The main proposition that remains to be tested empirically is whether obstacles to peace can be transformed into important catalysts for the reduction of mistrust.