Past Participle Agreement Revisited
Nicholas Longenbaugh
September 2018

Past participle agreement (PPA) in the Romance languages is well known for its overt correlation with movement, a fact that helped give rise to the Spec-Head model of Agreement (Kayne 1989; Pollock 1989; Chomsky 1991, 1993). In the decades following the original studies of PPA, however, there has emerged extensive evidence that agreement is not generally confined to a Spec-Head configuration, but rather may operate at a distance, without movement of the agreement target into the specifier of the agreement trigger. Granting this, the appealingly simple analysis of PPA in terms of Spec-Head agreement breaks down; to a first approximation, the challenge is this: if there is an agreement licensing head in the vP, and if agreement is, generally speaking, licensed at a distance, without movement, why is PPA conditioned on movement of the internal argument, at least in some cases? That said, however, PPA is considerably more complicated than the familiar Spec-Head data suggest. First, there are languages where PPA is obligatory, and hence not coupled with movement at all (see Loporcaro 2016 for a full list). Let’s call these PPA in situ languages. Second, there are languages that follow the more familiar Spec-Head pattern documented by Kayne. Let’s call these movement-based PPA languages. Even in such languages, however, there are exceptions to the familiar Spec-Head pattern of PPA, where agreement obtains in the absence of movement. We can therefore establish the following desideratum for a theory of PPA. First, it must address the fundamental question of why PPA is coupled to movement in a variety of languages and constructions, granting that agreement is intrinsically licensed at a distance. Second, it must explain why PPA is sometimes decoupled from movement, generally in PPA in situ languages and in some specific passive and unaccusative clauses in movement-based PPA languages. In this paper, I develop a theory that meets these desideratum, with the following basic logic. PPA with in situ objects in movement-based PPA languages vis blocked because these languages do not make dependent-case marked DPs accessible to agreement (Bobaljik 2008; Preminger 2014). The addition of movement into the derivation licenses the otherwise impossible option of agreeing before the external argument is merged, avoiding the case-blocking effect. This captures the correlation between movement and PPA. PPA in situ languages, in contrast, differ in that they make dependent case accessible to agreement, so that PPA obtains in all clauses, irrespective of movement. The exceptional PPA without movement in movement-based PPA languages is shown to follow from additional, independently motivated aspects of the theory.
Format: [ pdf ]
Reference: lingbuzz/004195
(please use that when you cite this article)
Published in: Ms, MIT
keywords: agreement, movement, participle, romance, economy, case, scandinavian, syntax
Downloaded:73 times


[ edit this article | back to article list ]