Speaker: Zoë Belk, Ad Neeleman and Joy Philip (UCL)
Title: What divides, and what unites, right-node raising
Date: Thursday 18 March
Venue: Skype (contact us to get access to the meeting)
Time: 15.15 – 16.30 hrs
We argue, following Barros and Vicente (2011), that right-node raising is either the result of ellipsis or of multidominance. The analysis is supported by four considerations. (i) Right-node raising has properties indicative of ellipsis, as well as properties indicative of multidominance. (ii) These properties can be combined, but only in limited ways. A pivot created through ellipsis may contain a right-peripheral secondary pivot created through multidominance, but not vice versa. Hence, a linear asymmetry in mixed patterns is predicted that is indeed present in the data. (iii) The two derivations are not in free variation, due to a restriction that multirooted multidominance structures can only be closed under coordination. Therefore, right-node raising in non-coordinate contexts displays properties indicative of ellipsis, but not properties indicative of multidominance. (iv) Multidominance gives rise to a difficulty in linearization that can be solved through a pruning operation. We show that the same operation delivers right-node-raising-as-ellipsis. Thus, right-node raising is either multidominance plus pruning, or pruning only. This explains why the two derivations are subject to the same word order restrictions.