Schedule 74. StuTS

In this presentation, I propose a novel approach to deriving reflexives in transitive constructions that relies on the emergent cooperation between v and Voice in binding-as-agreement (BAA). I characterize anaphoricity with a reflexive-Voice that acts as a probe searching its c-command domain. The unvalued φ-features on a reflexive internal argument (IA) fail to satisfy (in the terminology of Deal, 2015) the Voice-head. The two form a composite probe (as in Paparounas & Akkuş, 2023) due to the failed agreement. Once the next element is merged, Voice finds the external argument (EA) in its specifier, which values both elements.

Departing from works such as Rooryck & Vanden Wyngaerd (2011) and Diercks et al. (2020), this analysis does not stipulate that antecedents and anaphors directly agree with one another as probes and goals. Moreover, I propose the conceptual revision that the IA in reflexive constructions does not remain in-situ. Rather, it undergoes object shift as a result of obligatory EPP on the phase-head v — motivated by observations of argument extraction in ergative/absolutive languages (e.g., Mandar: Brodkin, 2022; Chuj: Brodkin & Royer, 2021). Crucially, adopting this proposal allows the IA to be directly accessible to the Voice-probe when it begins search, and the parsimonious strong Phase Impenetrability Condition (PIC; see Chomsky, 2000; Gallego, 2020) can be maintained. Voice has no need to access the already spelled-out phase, since the EPP feature on v feeds movement of the IA to the phase edge. From a broader perspective, this novel combination proves especially relevant for future research on the distributional pattern of Condition A as a window into agreement relations.

This talk will be streamed online and shown in the indicated room.

Especially with the advent of minimalist approaches to syntactic derivation, a growing pursuit since Chomsky (2000; 2001) concerns laying out the mechanisms of the Agree operation, the primary tool for feature-based dependencies (see Deal, 2023). Debates over the years have included search-space directionality, locale (syntax versus PF), and halting conditions. The particular line of research explored here involves the reducibility of binding to Agree (binding-as-agreement — henceforth: BAA), one of multiple ways in which researchers have attempted to decompose Conditions A–C (see Chomsky, 1981).

Two particularly intuitive principles guide a BAA approach. The fundamental restriction of precedence in anaphor distribution is described by Haspelmath (2023) as "antecedent-reflexive asymmetry" (2023: 37) — a postulate which entails Condition A. Moreover, the cross-linguistic empirical pattern emerges that pronominal forms and antecedents tend to match in φ-features (see Collins & Postal, 2012; the Pronominal Agreement Condition as in Angelopoulos et al., 2023).

Based on antecedent-reflexive asymmetry and the Pronominal Agreement Condition, downward agreement serves as a natural explanation at play for local anaphors. Rooryck & Vanden Wyngaerd (2011; R&VW) is a notable example of a BAA approach that seeks this. However, their account argues for a direct Agree relation between co-referent items — an account that falls out of place from a framework that delineate functional heads as probes, not lexical items such as anaphors.

Taking on a more fine-grained structure of the verbal system (i.e., where Voice introduces external arguments, e.g., Harley, 2013), anaphors adjoining to the edge of vP would not c-command antecedents. As a result, I depart from R&VW’s view that anaphors probe. Instead, I build on the analysis of Turkish by Paparounas & Akkuş (2023) in having Voice mediate this relation. The reflexive (IA) still moves to the phase edge, but serves as as goal rather than a probe. Agreement with Voice initially fails, since the anaphor carries unvalued features. Once the pre-valued EA is merged, Voice and the IA — as a composite probe — successfully Agree with the new specifier (as in cyclic Agree). T probes as expected, and the predicate verb can subsequently undergo head-movement to converge on linear order.

The IA undergoes movement to [Spec,vP] for a different reason than R&VW originally proposed. While this movement, referred to as the “Principle for the Anaphoric Properties of Agreement (PAPA)” by Diercks et al. (2020: 361) still occurs, it is instead motivated by the EPP-feature on phase heads that triggers movement, such as with object shift in erg/abs voice alternations (e.g., Mandar: Brodkin, 2022; see Legate, 2014 on Achnese). As a consequence of adopting the novel pairing of object shift and BAA on a cross-linguistic scale, the information in the transferred vP phase need not be accessible to Agree. The parsimonious strong PIC can be freely maintained at this step via the conceptual unification between multiple argument systems.

Info

Day: 2023-10-28
Start time: 09:30
Duration: 00:25
Room: NIG Raum 1
Track: Theoretical Linguistics
Language: en

Links:

• iCalendar

Files

• Signed Extended Abstract

Feedback

Click here to let us know how you liked this event.