Plural Dependent Elements in the RNR Context: Some Remarks on Ha's (2006a, b, 2007, 2008) Ellipsis Account

Abstract

Ha (2006a, b, 2007, 2008) advocates a semantically constrained ellipsis approach to the right node raising construction (RNR): RNR results from an ellipsis operation when conjuncts in coordination entail each other modulo focus-closure. He further claims that his ellipsis approach can accommodate Chung's (2004) observation concerning the privileged status of plural dependent elements (PDEs) in the RNR context (as opposed to the non-elliptical context), by manipulating their syntactic positions: PDEs are base-generated in the latter conjunct only (in the case of dummy plural marker -tul) or outside of the coordination (in the case of distributive adverb kakkak 'respectively'). This paper shows that Ha's account of the PDE distribution in the RNR context is not adequate. Crucially, it will be shown that his account makes an erroneous prediction that PDEs are to behave alike whether a coordinate structure is RNRed or not. This problem persists as far as ellipsis is a PF process, while PDEs are licensed in syntax.