The irritation of doubt: When is it OK to scratch your beliefs?

Building upon Peirce's notion of "the irritation of doubt", I outline the role doubts play in belief formation, drawing a distinction between two different phenomena: Peircean doubts and scary doubts. I argue that the former are an essential and useful ingredient in our belief dynamics, whereas the latter are potentially dangerous and yet unavoidable biases, even though they might occasionally be redeemed by their practical merits. This analysis is intertwined with Castelfranchi's views on belief formation and goal processing, with the aim of highlighting a common root between pragmatism and goal theory, and to provide a unifying picture of the mind as a goal-directed, coherence-seeking control system.