ON THE LEFT EDGE IN UG: A REPLY TO MCCLOSKEY Richard S. Kayne 1. Introduction In a recent article in this journal, McCloskey (1999) has suggested, on the basis of data from Irish, that our theory of UG should countenance a class of rightward movements applying in the derivation of PF representations. As he notes, this might be compatible with the antisymmetry proposal of Kayne (1994) if that proposal were taken to hold only of ‘‘narrow’’ syntax. Whether McCloskey’s suggestion would lead to a minor or to a major decrease in the restrictiveness of our characterization of UG is not easy to say. 1 The idea that there is a class of PF movements that resembles to a significant extent the class of familiar syntactic movements, but differs from them in having in principle no effect on interpretation, is rather like the idea that there is a class of LF movements that resembles familiar syntactic movements, while differing from them in having in principle no effect on pronunciation. In Kayne 1998, I argued that many analyses based on LF (covert) movement can and should be replaced by analyses using only overt movement. A parallel conclusion for PF would be that UG should not contain two classes of otherwise similar movement operations distinguished in principle by having or not having an effect on interpretation. If all syntactic movement operations 2 are uniform in feeding both phonology and interpretation, then McCloskey’s suggestion that there are (rightward) PF movements in Irish could not be correct. 3 Syntax 3:1, April 2000, 44–51 ß Blackwell Publishers Ltd, 2000. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA 1 As he himself notes, ‘‘. . . at present we have no theory whatsoever of what the properties of such [PF] movements might be . . .’’ (p. 207). 2 This formulation leaves open the possibility of a true phonological counterpart of movement such as (some version of) metathesis. 3 Similarly for stylistic movement of the sort alluded to by Chomsky (1998, fn. 45); for relevant discussion, see Kayne and Pollock 1999 (especially fn. 1). Chomsky (1998, fn. 69) suggests that head adjunction might be part of the PF component; if this were restricted to head adjunction, there would be no relevance to McCloskey’s proposal, which appears to involve only phrasal phenomena. For head adjunction to systematically be part of PF, English -n’t would have to be taken to originate above the final landing site of the subject in (ii), given: (i) *Why did anybody help us? (ii) Why didn’t anybody help us? (iii) *I know why anybody didn’t help us. That there is no phrasal PF movement is supported by McCloskey’s own footnote 6, citing Taraldsen (1981) (and Gue ´ron and May 1984) on the fact that relative clause extraposition affects Condition C of the binding theory—see also Kayne (1994, 122–123). This is relevant because relative clause extraposition (see McCloskey’s (38) and (39)) is one of the movement operations that McCloskey’s general discussion leads him to classify as phonological.