The Un-Realism of Realism?s Epistemological Presuppositions.

Realist IR scholars have become increasingly committed to the canons of positivist social science. Contemporary realist theorizing largely presupposes a strict separation of subject and object, emphasizing ?theory testing? and ?prediction? as touchstones of social science. For example, Stephen Walt noted with satisfaction that security studies (which ?fits comfortably within the realist paradigm?) ?adopted the norms and objectives of social science,? including the development of theories and ?the testing [of} their predictions against a scientifically selected body of evidence.? And John Mearsheimer urged scholars to ?use their theories to make predictions about the future . . . The world can be used as a laboratory to decide which theories best explain international politics.? If the world is a laboratory, it is incumbent upon realist theorists to detach themselves from the political events anticipated by their theories in the same manner that, say, a chemist avoids intervening in the experimental processes underway in his lab. In practice, however, realist IR scholars often intervene in political reality. They consult national security agencies, whisper in the ears of decision makers, write op-ed pieces, give interviews, and sign petitions in an attempt to influence policy and shape political events. For example, in recent months realists of all stripes have engaged in an overt political campaign against U.S. policy in Iraq. In their actual practices, then, realist scholars attempt to construct and re-construct political reality without realizing that the very act of their political intervention subverts the neat subject-object separation presupposed by their theories. By abandoning positivist presuppositions and drawing instead on constructivist approaches that regard purpose and analysis as ?part and parcel of a single process? (E. H. Carr), realists could greatly reduce the dissonance between their epistemological principles and their actual practices. ..PAT.-Conference Proceeding [ABSTRACT FROM AUTHOR]