This dissertation is concerned with how world knowledge (plausibility) affects reading beyond how predictable what comes next is. Previous studies have explained the effect of plausibility in terms of probability, either because plausibility is the conditional probability of a role filler given what precedes or because plausibility affects reading through predictability. This thesis proposes that plausibility does not reduce to predictability. Instead, I will argue that the effect of world knowledge on reading can be explained in terms of the semantic similarity between the actual input and other possible words that could have been used. I first show that predictability (as estimated through a cloze task) does not explain plausibility (as estimated via a plausibility judgment task) when a more precise measure of predictability is used and materials have a wider range of predictability scores. Second, I show that plausibility affects reading time later than predictability, suggesting that the mechanism underlying the effect of plausibility on reading differs from the one underlying the effect of predictability. Third, I show that inferred information about how participants in events are socially related can also affect plausibility and reading. Finally, I show that plausibility can be partly explained as how close a participant role filler is to the medoid of the semantic cluster that the input belongs to. Overall, the studies reported in this dissertation show that plausibility is not reducible to predictability and that the plausibility of event descriptions corresponds, in part, to how semantically related what is read is to other event participants in the same ad-hoc semantic domain. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml.]