The presupposition of even
I present a new observation with regard to the felicity of using even: There is no apparent focus/QUD congruence for even-sentences. For example, Even Mary came cannot be used to answer a question like who came or who was unlikely to come. Instead, the felicitous use of Even Mary came is to address issues like how successful the exhibition was, how enthusiastic people were, how urgent the matter was, etc. Thus I propose that the use of even is QUD-sensitive, always with regard to a contextually salient degree question. Even brings a degree-based presupposition of additivity, not an entity-based one (see also Greenberg 2018 for a similar view). An even-sentence presupposes that its prejacent is associated with a degree value, a benchmark value higher than the usual contextual threshold, resolving a degree question with an increasingly positive answer. E.g., under a relevant scenario about how popular a certain talk was, Even Mary came is roughly interpreted as (the talk was so popular) that Mary came. Under the current analysis, the entity-based additivity and likelihood-based scalarity of even, which are considered presuppositions under the traditional view, are now considered implicatures.