Because people often say one thing and do another, social psychologists have abandoned the idea of a simple or axiomatic connection between attitude and behavior. Nearly 50 years ago, however, Donald Campbell proposed that the root of the seeming inconsistency between attitude and behavior lies in disregard of behavioral costs. According to Campbell, attitude- behavior gaps are empirical chimeras. Verbal claims and other overt behaviors regarding an attitude object all arise from one "behavioral disposition." In this article, the authors present the constituents of and evidence for a paradigm for attitude research that describes individual behavior as a function of a person's attitude level and the costs of the specific behavior involved. In the authors' version of Campbell's paradigm, they propose a formal and thus axiomatic rather than causal relationship between an attitude and its corresponding performances. The authors draw implications of their proposal for mainstream attitude theory, empirical research, and applications concerning attitudes.