Guilds are social scientists? favoured historical example of institutions generating a 'social capital? of trust that benefited entire economies. This article considers this view in the light of empirical findings for early modern Europe. It draws the distinction between a ?particularized? trust in persons of known attributes and a ?generalized? trust that applies even to strangers. This is paralleled by the distinction between a ?differential? trust in institutions that enforce the rights of certain groups and a ?uniform? trust in impartial institutions that enforce the rights of all. Guilds had the potential to generate the particularized and differential trust to solve market failures relating to product quality, training, and innovation, although the empirical findings suggest that they often failed to fulfil this potential. Guilds also had the potential to abuse their trust, and the empirical findings show that they indeed manipulated their social capital of shared norms, common information, mutual sanctions, and collective political action to benefit their members at others? expense, blocking the spread of generalized and uniform trust. Counter to the assumptions of social capital theory, the example of preindustrial guilds suggests that the particularized and differential trust fostered by associative institutions do not favour but hinder the generalized and uniform trust fostered by impartial institutions.