You are using an old browser with security vulnerabilities and can not use the features of this website.
Relatively recent literature (e.g. Massam 2001, Öztürk 2005) re-analyzes instances of some phenomena which were analyzed as “genuine” incorporation via head movement as a different phenomenon altogether, dubbed “Pseudo-Noun Incorporation” (PNI), which essentially views the “apparently” incorporated element as an NP which is simply aligned immediately next to the verb (before the verb in head-final languages and after the verb in head-initial ones). No head movement is involved. The “pseudo-incorporated” material is phrasal, rather than an N, as it would be in “genuine” noun incorporation (NI).
In this talk, I address some facts from Turkish which have not been widely discussed in the literature. These are facts that are somewhat reminiscent of Holmberg’s Effect, i.e. the interaction between verb movement and object shift in some Germanic languages, where the latter is contingent upon the former (cf., among others, Holmberg’s work, after whom this relationship is named), as well as of phenomena described under the heading of Specificity Effect (cf. work by Diesing and others), whereby subextraction from non-specific phrases is possible, while such subextraction from specific phrases is not.
I argue that, in addition to being stipulative, the “Specificity Condition” on subextraction does not cover all relevant facts of subextraction in Turkish, when interacting with the specificity of the hosts of the extracted subconstitutents. Instead, I argue that what makes the successful instances of subextraction possible is head-movement of the noun heading the non-specific potential host, adjoining to the verb. Thus, what’s at stake here is a form of the Government Transparency Corollary (Baker 1988), possibly recast in more recent terms as Phase Extension (cf. den Dikken 2007 a., b.).
Such an explanation would not be available, if the relevant Turkish facts were instances of PNI, i.e. were best characterized as a simple concatenation of a bare NP and a verb, as claimed in Öztürk (2005). I then turn to apparent problems for the head-movement based analysis of these Turkish facts, pointed out by Öztürk, and show, based on adjacency requirements (and their loose nature) of (other) complex verbs in Turkish, that the problems are indeed only apparent. Finally, I discuss the difference between this type of (Pseudo-)Noun Incorporation and the instances originally discussed in Baker (1988), where NI results in changes with respect to the valency of the verb, while the instances addressed in this study do not trigger such changes.