This paper addresses the two main theoretical questions concerning clausal ellipsis: (i) whether remnants move out of the E(-llipsis) site, or remain in-situ with their surroundings elided, and (ii) what is the E-site of clausal ellipsis. In this paper, I provide novel evidence involving contrastive fragments in Thai that favours the movement-plus-ellipsis view of clausal ellipsis (see Merchant 2004) over competing in-situ analyses cf. Abe (2015), Ott & Struckmeier (2018) and cleft-based source analyses (Barros et al. 2014). The empirical evidence presented here shows that fragments exhibit properties more compatible with ex-situ focus than clefting and in-situ structures. Furthermore, I demonstrate that fragments are derived from focus fronting and deletion, and that constraints on fragments closely track those on overt A′-movement to FocP: whenever ex-situ focus is blocked, fragments cannot be licensed. To this end, I demonstrate that E residing on the targeted head of focus or the trace of focus undergoes E extension, which is responsible for Foc′ ellipsis.
Woraprat Manowang (Fri,) studied this question.