Remarks on Ha (2012): Back to a Multi-Dominance Approach to the Transitive Free Relative

Abstract

To account for the dual function of the nucleus in the so-called transparent free relative construction (TFRC), various theories have been entertained in the literature: One of the most representative among them is a multi-dominance based approach (MDA, van Riemsdijk 1998, 2000, 2001, 2006, Citko 2011, among others). Ha (2012) claims that the MDA is not tenable, as it cannot deal with polarity licensing and it also bears a linearization problem due to the dual root property. He instead proposes a sideward movement analysis (SMA). His criticism on the MDA is mostly legitimate but it will be shown that it can be avoided, if the dual root property is ultimately eliminated, as in Citko’s (2011) approach. Furthermore it will be shown that the SMA faces a new kind of linearization problem, a new kind of polarity licensing problem, and a cyclicity problem. Grasping that these problems arise due to a presumed derivation in which the nucleus ‘completely leaves and c-commands into’ the transitive free relative (TFR), this paper suggests a new version of MDA: It postulates a null head which takes the nucleus as complement and the TFR as specifier (or possibly as adjunct). The postulated null head is assimilated to & in an endocentric analysis of &P in the sense that it is category-neutral and its categorial status comes from its complement (i.e., the nucleus). The newly proposed structure also meets the right edge condition, just as the right node raising construction involving an &P (Wilder 1999, 2008).