Where There is Fire There is Smoke : Local Modelling of Successive-Cyclic Movement

Abstract

The goal of this thesis is twofold: The first aim is to diagnose for phase heads on the basis of semantic, morphological and syntactic reflexes of movement, and to propose a phase model resting upon these empirical observations. This work examines most known reflexes of successive-cyclic movement and develops a phase model resting upon the empirical evidence.The result is that all core heads on the spine of the derivation - C, I, v and V - are phase heads. The main outcome is thus that each phrase is a phase. Phases are therefore not defined on semantic grounds, but by purely syntactic criteria: A phase comprises the smallest possible operation space that permits a successful derivation.The second aim is to propose a uniform analysis for morphological and syntactic reflexes of successive-cyclic movement. The basic idea of the analysis is that morphological and syntactic path effects are due to the deletion of morpho-syntactic features of probing heads as a consequence of movement to the edge. Impoverishment thus happens in the syntax. Probe impoverishment has an effect on the inflectional markers that are post-syntactically inserted into the probe: If morphosyntactic features are deleted before vocabulary insertion takes place, then a marker M may not fit anymore into the relevant context. In this case a less specific marker than M is inserted. I formulate this generalisation as follows: When a language shows different exponents in movement and non-movement contexts, then the marker appearing in the context of movement is less specific than the marker appearing in non-movement contexts (= retreat to the General case, emergence of the unmarked).I apply the new analysis not only on morphological, but also on syntactic path effects such as verb inversion. The decisive advantage of the new approach is that it offers a uniform analysis for all morphological and syntactic path effects which is yielded by a minimal change in the modelling of movement to the edge. The analysis correctly derives the surprising characteristic of long movement in languages like Chamorro that higher verbs do not register the argument status of the passing wh-element, but the respective argument status of the clause from which the wh-element is extracted. In addition, the analysis derives the effects of (seeming) extraction restrictions correctly without actually invoking These restrictions. Furthermore, syncretisms in the verbal morphology of Chamorro and Irish can now be treated as occurrences of one and the same underspecified marker. Finally, the new analysis delivers a new argument in favour of finite control as movement, and resolves two puzzles in the syntax in Kikuyu (subject-object extraction asymmetry, location of wh-subject below complementiser) by deriving that wh-subjects in this language can stay in situ while at the same time extraction morphology shows up.