There are currently a large number of models which identify self-evaluation (self-esteem) as an important source of motivation. However, these models often posit qualitatively different antecedents and consequences. The present studies focus on the questions of whether these qualitatively different behavioral systems affect the same or different mediating variables, and whether the motivation is to maximize or simply maintain a particular level of self-evaluation. In Study 1 we found that providing subjects a "self affirmation" (Steele, 1988) opportunity reduced their propensity to engage in self-evaluation maintenance behaviors (SEM; Tesser, 1988). In Studies 2 and 3 we found that making salient positive SEM scenarios reduced the propensity to engage in dissonance reduction whereas making salient a threatening SEM scenario did not. These results were interpreted as indicating that these hypothetical self-systems affect the same mediating variable and that the motive is to maintain rather than maximize self-evaluation. © 1991.