Bhatt & Dayal (2020): the polar question particle kya: #
Hindi-Urdu polar kya: is a polar question particle (PQP) — a class
distinct from clause-typing Q-morphemes like Japanese ka: it occurs only in
polar questions (not in wh-questions), sits in a projection above CP the
paper labels ForceP, and embeds only in the quasi-subordinated configuration
of [DG09] (complements of rogatives), not in ordinary
subordination. Semantically kya: presupposes that its sister question has a
singleton alternative set (eq. 23) and is otherwise the identity; a polar
question denotes the singleton {p} (eq. 22b). The same singleton
restriction models the (bias-introducing) Mandarin particle nandao in
[Xu12a] (cross-linguistic picture in [Xu17]), which the paper draws on.
The singleton-alternative presupposition (eq. 23) #
⟦kya:⟧ = λQ⟨st⟩t : ∃p ∈ Q[∀q[q ∈ Q → q = p]].Q — defined only when the
sister question Q has a singleton alternative set, and then the identity
on Q. The presupposition is Question.IsSingleton; the felicitous sister
is the subtype SingletonQuestion W. In the paper a polar question denotes
a singleton (eq. 22b: ⟦did John leave⟧ = {John left}) — the substrate's
one-cell ofSet p, not its two-cell Hamblin polar.
Felicitous case: a polar question in the paper's sense is a singleton.
Defined case: on a felicitous sister, kya: is the identity.
The headline restriction (ex. 4): a wh-question with two distinct answer cells is non-singleton, so kya: rejects it.
A two-cell Hamblin polar likewise fails the presupposition.