In this paper I examine what constructivist approaches to IR tell us about how states should act when confronted by atrocity crimes in the context of a politically pluralist international society. Building on the work of theorists who responded to Richard Price and Christian Reus-Smit's call to substantiate the constructivist's claim to explain “moral progress,” and to better inform normative assessments, I claim that the constructivist emphasis on historical and social contingency does not rule out ethical standpoints, suggesting instead a “pragmatic” ethic. Norms are hypotheses rather than absolute values. The task of the pragmatic constructivist is not to establish beyond doubt the normativity of a norm—the task is to test the norm for how well its “meaning-in-use” supports action that ameliorates lived social problems. Pragmatic constructivists can commit to the Responsibility to Protect (R2P) as an idea that might reconcile various communities of normative practice and ameliorate vulnerability without upsetting international order. To the extent particular practices (e.g., prevention) have proven (and continue to prove) useful in doing this, pragmatic constructivists can equate the institutionalization of those practices with normative progress. However, R2P is a “complex” norm. It recognizes that knowledge of how to respond to ongoing atrocity is context specific and cannot therefore be fixed. The task of the pragmatic constructivist here is to assess the practical judgment of those that claim to speak for the norm by weighing the consequences of acting out their prescription in the specific context of a particular crisis. I test the meanings of R2P in use during the Syria crisis from 2011 to 2012. R2P was problematic to the extent certain meanings reinforced policy ends (e.g., protection through political/criminal accountability) that were impractical and pursued at the expense of non-ideal but realizable goals (e.g., protection through peace/aid).