LAUSR

This paper has three main goals. First, to motivate a puzzle about how ignorance-expressing language like maybe and if interact: they (surprisingly) iterate and when they do they exhibit scopelessness. Second, to argue that there is an ambiguity in our theoretical toolbox and that resolving that opens the door to a solution to the puzzle. And third, to explore the reach of that solution (it turns out to do work in unexpected places). Along the way, the paper highlights a number of pleasing properties of two elegant semantic theories (data semantics and update semantics), explores some meta-theoretic properties of dynamic notions of meaning, dips its toe into some hazardous waters (epistemic contradictions and presupposition projection), and offers characterization theorems for the space of meanings an indicative conditional can have. 1 Ignorance and information I am ignorant about a great many things. And so are you. Our language, equipped as it is with modals and conditionals, is well suited to express some of this. (1) a. Maybe the picnic is a success. b. The weather’s gotta be fine by now. c. If the weather held, the picnic is going as planned. This is good: by sharing our ignorance we can winnow away at it. Modal claims like these are equally quantificational claims: that there is a possibility compatible with the relevant information in which the picnic is a success, that all of the possibilities compatible with the relevant information are fine weather possibilities, that none of the weather-holding possibilities compatible with the relevant information are also plan-disrupted picnic-wise possibilities. But when we exchange information about our ignorance, we also