Richard Faure (Tours–CeTHiS)

Finite complement clauses (FCCs) are famous for resisting placement in argumental (A‑)positions: When they are objects, they tend to extrapose (Emonds 1976), when they are subjects, to topicalize (Koster 1978), an observation that is robust crosslinguistically. This behavior was attributed to their being case-less (Stowell 1981) and/or to their semantic type, for instance, <e,t> (instead of e, as expected of arguments, Moulton 2009, for other versions of this stance, see Elliott 2020, Bondarenko 2022, Moltmann 2024). What was little noted, though, is that they are reluctant to occupy most discourse (Ā-)positions as well. In this talk, I argue that this can only be captured if we go back to FCCs’ denoting entities (Vendler 1967, Zucchi 1993) and draw different conclusions from their lacking case within the framework of Phase Theory (Chomsky 2000). I contend that FCCs are prototypical arguments, whose placement restrictions are attributable to their being unmovable in-situ arguments and being licensed vicariously (Gunkel and Hartmann 2020, Faure 2021, Kiemtoré 2022). I then address apparent counter-examples to this inertness of FCCs and show that movements of FCCs do not exhibit the trappings of narrow syntactic movement. They are crosslinguistically amenable to two subcases: When they are anaphoric and when they are contrasted. Following Horvath 2010, I claim that marked discourse factors such as anaphora and contrast are treated late and therefore that their effects escape the constraints of Narrow Syntax.