Polar question bias

A polar question is sensitive to two bodies of information bearing on its prejacent p: the contextual evidence available in the discourse, as Büring and Gunlogson put it, and the prior belief of the speaker, the original bias of Ladd and of Romero and Han. Each supports p, supports ¬p, or neither, so its state is a sign, 1, -1 or 0, and reading it relative to ¬p negates it. Under a prior the state is the relevance sign of the body of information for p in Merin's decision-theoretic semantics, DTS.relevanceSign, which swapping p for ¬p negates.

A bias value in Sudo's scheme is the set of states a question is compatible with: [+s] (BiasValue.plus), its complement [−s] (BiasValue.minus) and [neutral]. A BiasProfile pairs an evidential and an epistemic value, and its felicity is the set of situations, pairs of the two states (Situation), that it admits. The forms and readings of polar questions are Negation.Position and Option Polarity.

References

• Büring & Gunlogson 2000, Aren’t Positive and Negative Polar Questions the Same? formalized
• Ladd 1981, A First Look at the Semantics and Pragmatics of Negative Questions and Tag Questions
• Romero & Han 2004, On Negative Yes/No Questions formalized
• Sudo 2013, Biased polar questions in English and Japanese