Extra deletion in fragment answers and its implications

Abstract

In this paper, I examine a phenomenon in Korean involving fragment answers and consider its implications. Taking as a point of departure the generalization that case markers on ellipsis remnants in fragment answers can be omitted only in string-final position, I argue that in these situations, PF deletion extends into the ellipsis remnant, deleting parts of it, such as a case marker, a postposition or, sometimes, even the head noun, up to recoverability and under adjacency to a string of elements that are deleted in PF for independent reasons. This parasitic deletion process, which I term “extra deletion,” sheds light on the nature of PF deletion, which I argue operates on strings of elements, similarly to the way that syntactic operations target constituents. The crucial idea is that, although it is mostly syntax that determines what is to be deleted (and, thus, elements that undergo ellipsis are usually syntactic constituents), PF deletion also has its own guidelines when it applies—namely that elements that are elided should form an unbroken, continuous string. What is interesting is that in contexts of extra deletion, the string of deleted elements is extended beyond what is initially marked for deletion by syntax, an important consequence of which is that PF deletion can ignore syntactic constituents. Furthermore, I make the novel observation that there exists a significant parallelism between fragment answers and right node raising, which has not been noted in the literature due to the sheer differences in their structure, surface form, and use. I argue that postulating extra deletion allows us to capture the parallelism straightforwardly, which in turn provides an additional argument for the PF deletion analysis of the latter construction, for which there have been several alternatives.