The precise motivation for affixation has not been entirely settled. Noyer (1992) and Hankamer & Mikkelsen (2018) argue that the identity of an affix is recognized in syntax-free contexts or in postsyntactic environments. On the other hand, Richards (2010, 2016) proposes a way of identifying affixes by looking into their metrical dependencies initially detected in narrow syntax. Here, I argue alongside Richards (2016) that these suprasegmental features are visible in syntax and that they trigger XP-movements (see also Branan 2018). According to Contiguity Theory (Richards 2016), overt movements triggered by syntactic features such as [uwh] and Extended Projection Principle (EPP) in the Minimalist Program (Chomsky 1993, 1995) are reanalyzed as operations sensitive to the interaction between syntax and prosody. To be more specific, narrow syntax looks at certain phonological information that works in favor of the initial shape of prosody. This suggests that syntactic movement is sensitive to prosodic contiguity prior to spell-out. Richards (2016) discusses some key motivations for movement. They include Probe-Goal Contiguity, Affix Support, and Untethering. Adopting some of the basic assumptions proposed in Match Theory (Selkirk 2009, 2011), Contiguity Theory looks into phonological motivations for wh-movement and EPP. In this paper, I argue that Swahili demands additional explanation as to how prosodic requirements are satisfied. In detail, I present an analysis accounting for the wh-in-situ phenomenon as well as the presence of EPP in Swahili. With regards to the presence of EPP, I propose that Swahili tense affixes require metrical boundaries on both left and right of their peripheries. The metrical boundary on the right is satisfied by the phonological content inside vP. The metrical boundary on its left is satisfied by an XP targeting [Spec,TP] which gives rise to the desired EPP-effect.
Preposition stranding versus pied-piping: Negative Shift of prepositional complements in dialects of Faroese
