Case and Postposition Stranding in Multiple Fragments: Why is the Final Fragment (not) Special?

Abstract

This paper examines the variability of Dependent Marker (DM) drop in multiple fragments. Building on Park (2013), we first observe that only the right-most fragment can optionally drop in multiple fragment answers, which is hard to be captured by the main approaches to fragments. Assuming that multiple fragments are derived by ellipsis we argue that stranding DM (i.e., case-marker and postposition) is possible under ellipsis, but that DM-stranding movement across an intervener outside of ellipsis site leads to a PF-crash, therefore capturing the fact that only the right-most fragment can strand its DM. The analysis can also capture a puzzling fact that Korean allows postposition stranding although it is not a P-stranding language under regular movement (Merchant 2001). Finally, it is shown that the analysis can be extended to a different type of fragments, dubbed as focus doubling construction.