That-trace effects are yet to be explained away: challenges for prosody-based accounts
Luis Miguel Toquero-PĂ©rez
July 2021
 

**[NEW TITLE: that-trace effects: haplology is the answer.]** That-trace effects in English have been treated as violations of prosodic wellformedness conditions, i.e. Kandybowicz (2006a) and Sato & Dobashi (2016): the C head and the trace cannot be parsed in the same prosodic phrase. Satisfying this condition is claimed to allow for amelioration effects induced by sentential adverbs, narrow focus on the embedded verb, and pronominal resumption. However, I argue that a [C t] prosodic parse cannot be the reason for that-trace effects since such a prosodic phrase is not empirically supported by several data points, and in fact should not be formed in the first place as it violates prosodic well-formedness conditions. Adopting some insights from the Cyclic Linearization model (Fox & Pesetsky 2005b;a; Davis 2020b), I offer an explanation for the effects and their amelioration in terms of an interface condition: the Two-Copy filter. The filter bans two strictly identical intermediate copies from occurring in the same spell-out domain as it would lead to a lethal ambiguity problem during linearization. The effects are thus a subtype of the Obligatory Contour Principle effects. Once the identity is avoided by pre-emptive rules (Radford 1977; Nevins 2012; Neeleman & Koot 2017) early during the derivation (i.e. adverb intervention), or repair mechanisms at different PF stages, the effects fail to obtain.
Format: [ pdf ]
Reference: lingbuzz/005548
(please use that when you cite this article)
Published in: Submitted
keywords: that-trace effects; syntax-pf interface; syntactic haplology; identity avoidance; relative clauses; narrow focus; resumption;
previous versions: v3 [November 2020]
v2 [November 2020]
v1 [November 2020]
Downloaded:1088 times

 

[ edit this article | back to article list ]