There is a paradox at the centre of a substantial amount of writing on the topic of leadership in organizations, particularly school leadership. On the one hand, there is what may be called a common-sense consensus that leadership is vitally important as a cause for setting and achieving organizational goals. Yet, on the other hand, an impressive body of empirical research concludes that the effect size of leadership on certain key organizational outcomes, such as promoting student learning, is small. We begin by discussing the first of these claims. We then discuss the second and finally discuss proposals for a resolution which is framed within the perspective of leadership theorized as a form of critical learning. In our resolution we argue that the contingencies of leadership contexts are sufficiently different to compromise the goal of producing a single leadership model. Instead, we urge that the role of school leaders in promoting learning, or other organizational goals, needs to be discerned from the leader's own theory that guides their decisions and actions, at least where that theory is developed from epistemically successful problem-solving practices. Such a stance implies that the paradox is generated by false assumptions and that both large-scale empirical studies and generalized leadership models are inappropriate tools for discerning the contribution of leaders to school outcomes.