Pseudo noun incorporation in Hocąk

•Baker's (2014) analysis of PNI cannot account for the data from Hocak (Siouan).•Contra Baker 2014, there is no surface adjacency requirement for Hocak PNI.•Hocak also does not display V-to-T movement.•The Hocak data are compatible with Massam (2001).

This paper provides evidence from Hocąk (Siouan) that Baker's (2014) analysis cannot account for all cases of pseudo noun incorporation (PNI). Baker (2014) argues that PNI is best analyzed as complex predicate formation derived via head movement of the object NP's head noun into the verb. Baker contends that the resulting structure cannot be linearized at PF if any material intervenes between the two copies of the noun, and thus that this movement must be string-vacuous. Here, I present data from Hocąk PNI that are problematic for Baker's theory. First, I show that internal arguments that are not base-generated in a position linearly adjacent to the verb can nonetheless be incorporated. Second, NP-internal modifiers can surface between the incorporated noun and the verb, which is also not compatible with Baker's theory. Instead, this paper demonstrates that Massam's (2001) analysis of PNI in Niuean (Oceanic) is able to account for the Hocąk data. Massam proposes that PNIed objects are NPs that are base-generated in object position. Thus, while Baker's head movement theory may be the right analysis for PNI in other languages, I argue that the data from Hocąk demonstrate that it cannot be the only possible mechanism.