Relevance and Questions

Notice that this analysis – just like Groenendijk & Stokhof’s one – predicts wh-questions to have a mention-all reading only. To account for mention-some readings it is then normally assumed that wh-questions are, in fact, ambiguous between those two readings of questions, and that the appropriate reading is then selected by pragmatic criteria (as in van Rooy, to appear). In this note, however, I want to show that we don’t need to assume that wh-questions are ambiguous between mention-all and mention-some readings, and that, in fact, both can be derived from a single underspecified formula with only one contextually given parameter. I will show that by slightly changing (2) and (3) we are done. If we want to account with one formula for both mention-all and mention-some readings of wh-questions, it is obvious that we have to give up rules like (1) and (2), because the identity used in these formulas guarantees that we will end up with an equivalence relation, which we don’t want in case of mention-some readings. Instead of (1) and (2), I will propose the following rule: