Preferring to Learn: An Attitudinal Approach to Polar Questions

The semantics and pragmatics of polar questions remain undertheorized. Standard proposition-set accounts (Hamblin, 1976; Groenendijk and Stokhof, 1984), which take question meanings to be identifiable with the (appropriately defined) set of their answers, have been developed mostly for wh-questions; when applied to polar questions, they predict massive synonymy, and therefore struggle to account for the differential acceptability of polar questions across contexts. For instance, imagine that your interlocutor sees you writing with your left hand. They ask: (1) a. Are you left-handed?b. #Are you right-handed? c. Are you not right-handed? d. #Are you not left-handed? On the basis of this and other evidence, a number of authors have recognized the need to distinguish the radical proposition in some way (Krifka, 2001; Romero and Han, 2004; Roelofsen and Farkas, 2015) and identified epistemic and evidential conditions on polar question felicity (Büring and Gunlogson, 2000; Sudo, 2013). But no account derives the felicity of these questions in various contexts from their systematically different meanings. This dissertation presents an analysis of polar questions on which they express the speaker’s conditional preference for doxastic update with the radical proposition. This preference is structured by an ordering source consisting of the speaker’s goals and the necessary conditions to achieve those goals. This combination provides an account for fine-grained polar question felicity across contexts, as well as the “functional heterogeneity” (Kaufmann, 2012; Condoravdi and Lauer, 2012) of polar questions, which sees them able to perform various speech acts like offering, requesting, conversation-starting, and more (Bolinger, 1978). In the second part of the dissertation, this analysis of polar questions is applied to what I term neg-biased polar questions, more commonly called ‘high-negation questions’ (Ladd, 1981), such as the following: (2) a. Aren’t you right-handed?b. Aren’t you not left-handed? The major puzzle is how the “high negation” of these questions gives rise to an inference of speaker bias without contributing propositional negation of the radical (Goodhue, 2018). I argue that negation scopes over the entire attitudinal meaning of the question, expressing the speaker’s lack of a preference for doxastic update. These two components conspire to guarantee that the speaker has a prior belief in the radical proposition. Here I also discuss negative polarity items in polar questions, which have been a sticking point in the literature on neg-biased questions especially. I argue that a long-assumed “inner/outer” ambiguity in neg-biased questions is an illusion owing to the independent ability of NPIs to appear in polar questions, and in particular the pragmatic effect of choosing an NPI contributing domain-widening (Kadmon and Landman, 1993). Zooming out to consider broader consequences, this analysis is one on which polar questions primarily serve as a strategy for belief management. It also implicitly problematizes the concept of polar questions as a cross-linguistically stable phenomenon; languages may—and do—have different strategies for expressing speaker interest of various kinds in the truth of a proposition, or the answer to a question, of which the canonical English polar question is just one