American political science has long aspired to emulate both the objective research methods of the natural sciences and their practical successes in controlling their objects of study. Regrettably, the putative tension between these two ambitions is rarely discussed. This essay seeks to touch off such a discussion by illuminating a significant problem that produces tension between objective knowledge accumulation and practical control of politics, but not of nature: self-disconfirming analysis. The problem is that in some situations, successful realization of the normative implications of political analysis may create new political patterns that are no longer consistent with the law-like regularities uncovered by that analysis. I demonstrate how this problem is manifest in the work of Robert Putnam, whose career exhibits a commitment to (naturalistic) scientific rigor as well as a passion for sociopolitical change. If the agenda implied by Putnam's scientific research were to be implemented, some of the causal claims established by that research would be removed from actual operation. I argue that the failure of political science to realize its naturalistic aspirations is at least partly attributable to this problem.