What moves where? A typological-syntactic approach to multiple wh-questions

This paper presents a new syntactic analysis for multiple wh-constructions.
Adopting Richards (1997), I assume that there are two types of languages concerning
wh-movement: such which A-move their wh-words (A-languages) and such which
A’-move them (A’-languages). I expand this account by assuming that in both types,
wh-movement targets the CP. This is done via A’-movement, as well as via
A-movement. Building on recent work on cross-clausal A-dependencies (CCA;
mainly Wurmbrand 2018), I adopt the idea of [A]-features inside CP. Based on Rizzi
(1997), I propose a split CP domain whose different parts (ForceP, FocusP,
TopicP) can either have A’- or A-quality. Wh-movement targets ForceP and FocusP,
CCA-elements move to TopicP. The CP heads are ordered in an implicational
hierarchical way, and their feature make-up entails the properties of the embedded
and higher-ordered heads. Within this ordering, there is a threshold where
A’-qualities shift to A-qualities. I assume that certain CP-heads are able to contain
A-properties and by that, the CP domain contains A’- as well as A-properties. The
option of having A-quality is restricted by embedding options: An A-CP-part
cannot embed an A’-CP-part within the same domain (CP). I claim that languages
pattern into three types, depending how high in the CP the A’/A-shift is located.
This assumption predicts that there is a correlation between A-wh-movement and
CCA-phenomena. This is indeed the case and will be summarised as a novel
typological generalisation: “Whenever a language A-moves its wh-words, it allows CCA
(but not the other way around).” This generalisation reflects that A-wh-movement
entails CCA, as expected by my analysis of the CP domain. My account ties
together CCA-phenomena and A-wh-movement in a syntactically novel way and
might shed new light on the universal composition of the CP-domain.