In an effort to explain conflictual and cooperative actions in the Middle East and Northern Ireland at a time (1995-1999) when international structural factors were relatively constant, this article focuses on cognitive factors. Specifically, statements of leaders representing multiple actors in the Israeli-Palestinian and Northern Ireland relationship are examined. Texts from these leaders serve as data for the independent variable, problem representation. Hypotheses argue that the existence and centrality of actor problem representations can help explain behavior, in a manner that adds to more widely used scholarly explanations. More specifically, the article explores the relationship among leaders' "problem representations" and conflict between the entities they lead. Problem representation is measured in three dimensions using Foreign Broadcast Information Service (World News Connection) texts: (1) centrality of enemy image, (2) how inclusive/exclusive the leader is in describing the in-group and principal outgroup, and (3) the key definition of the problem (coded in terms of three strategies: justice, governance, or threat). Conflict, the dependent variable, is measured in terms of KEDS-TABARI event data on deeds (not words) in the Israeli-Palestinian and Northern Ireland conflicts. Strikingly, the analysis finds that Israeli actions are strongly related to prior and current Palestinian leaders' problem representations, much more than they are to Israeli leaders' problem representations. Similarly, Palestinian actions are found to have a clear relationship with prior and current Israeli leaders' problem representations, much more than they are with Palestinian leaders' problem representations. These results are particularly strong when the problem representation is one of the overall political leadership on the "other" side. Additionally, in-group and out-group inclusivity are the most significant predictors of actions. For Northern Ireland, the same two themes prevail, although not as strongly: there is a clear statistical relationship between each side's problem representations and the other side's actions, stronger in fact than the relationship between their own side's representations and actions. Also, in-group and out-group inclusivity produce a strong statistical relationship with conflict and cooperation. Finally, results are compared with a "tit-for-tat" hypothesis, and found to embellish that hypothesis.
