Speaker: Magdalena Lohninger (MIT & University of Vienna)
Title: Is composite [Ā/A] probing extended A-movement?
Time: Thursday, 21 November, 12:30pm – 2pm
Location: 32-D461
Abstract: In the last ten years of syntactic research, composite [A’/A] probes have been employed to account for a variety of unrelated phenomena: i) topicalization, focalization, wh-extraction, relativization with A-properties (van Urk 2015, Ostrove 2018, Scott 2021, Chen 2023), ii) passivization/raising with A’-properties (Wurmbrand 2019, Colley & Privoznov 2020, Lohninger et al. 2022, Chen 2023), iii) A’-extraction restrictions to the closest DP (Erlewine 2018, Branan & Erlewine 2020, Coon et al 2021, Branan 2022), iv) φ-agreement sensitive to information-structure (Mursell 2021, Bárány 2023). In this talk, I focus on the phenomena i)-ii) and present a comparative investigation of Dinka extraction (van Urk 2015), Mandarin Chinese BEI passives and low foci/topics (Chen 2023), Khanty passives (Colley & Privoznov 2020), and Balinese and Malagasy promotion to pivot (Erlewine, Levin & van Urk 2019, Lohninger & Katochoritis to appear). I explore whether these constructions exhibit random mixtures of A’- and A-properties or systematic distributions thereof and show that they follow a highly predictable pattern: whilst composite A’/A constructions resemble A-movement in most respects (creation of new antecedents for anaphor/variable binding, lack of reconstruction for principle C, feeding case, agreement and subsequent A-movement), they differ from classical A-chains in three aspects: i) the ability to skip intervening DPs, ii) obligatory information-structural or clause-typing effects and iii) the landing-site at a phase edge.
Based on this observation, I suggest that even though composite [A’/A] probes search for a goal conjointly, the movement chain they invoke is A-movement. I assume that standardly, [A’] probes (like [top], [foc], [rel] or [wh]) occur on phasal heads (C and v), whilst [A] probes ([D] or [φ]) appear lower (e.g. T or V). In some (rare?) cases, [A] features can end up on phasal heads together with [A’] (due to different reasons, e.g. [φ] under-inheritance, syntactic language change, formation of non-canonical passive heads, etc.), creating a composite A’/A construction. Following van Urk & Richards 2015, I assume that if [A] and [A’] occur on the same head, they search together for a fully fitting goal (for economy reasons), whereby the search can skip intervening DPs and prefers a goal that is information-structurally marked (a top, wh, foc or rel element). After [A’/A] found a goal conjointly, I propose that it is then the [A] probe alone which invokes movement, based on the idea that A-operations are timed before A’-operations (Abels 2007). If [A] and [A’] compete as movement triggers on the same head, the [A] probe gets precedence, hence [A’/A]-induced movement is an A-chain. Thereby, an [A’/A] phase head can be understood as a way of extending the time window for A-movement to the phase-edge (as opposed to the usual case where A-operations are completed before the phase head is reached).
Based on this analysis, the distribution of A’- and A-properties in composite constructions becomes predictable, which counteracts the seemingly anything-goes impression [A’/A] probes might sometimes create.