Linguistic expressions of modality and evidentiality have been the object of much active and exciting research for many years, but continue to pose challenges in all areas of semantic theory. Key aspects of their behaviour at the syntax–semantics interface, such as interactions with quantifiers and other operators or the calculation of modal implicatures triggered in embedding contexts, are not yet fully understood. Likewise, both the semantic interpretation and the pragmatic role of nondeclarative sentences in modal constructions have only recently begun to attract significant attention, partly because the standard formal apparatus requires substantial augmentations in order to be applicable in that domain. But even one of the most fundamental and least controversial aspects of that formal apparatus, the use of possible worlds in the model theory, is subject to criticism and controversy. The articles in this collection address open issues in each of these three areas. The first paper, Yurie Hara’s ‘Evidentiality of Discourse Items and Because-clauses’, addresses the problem of implicature computation involved in Japanese, focusing on contrastive wa-marking and sooda, a marker of ‘hearsay’ evidentiality. The use of contrastive wa-marking presupposes a contextually given stronger alternative, e.g. Mary and John came in (1b). (All glosses in this section are Hara’s. ‘Con’ glosses contrastive wa.) When this presupposition is satisfied, the sentence conventionally implicates that the speaker does not know that the stronger alternative is true, or knows that it is false.