Abstract Implicit evaluation can be defined as the automatic effect of stimuli on evaluative responses. A major advantage of this definition is that it is neutral with regard to the mental processes and representations that mediate implicit evaluation. Whereas many existing models postulate that implicit evaluation is mediated by the automatic spreading of activation along associations in memory, it is also possible to entertain the idea that implicit evaluation is due to the automatic formation or activation of propositions. In line with such a propositional model of implicit evaluation, evidence suggests that implicit evaluation (a) can be based on instructions and inferences, (b) is sensitive to information about how stimuli are related and (c) can reflect several propositions that differ only with regard to how stimuli are related. Although it might be difficult to differentiate between propositional models on the one hand and association–activation or dual process models on the other hand, merely considering the idea that implicit evaluation might be mediated by propositions offers a new perspective on existing findings and leads to novel predictions about the conditions under which implicit evaluation occurs.