Search

We reply to Erlewine and Kotek’s (2018) claim that the phenomenon of covariation under focus ( Tanglewood sentences; Kratzer 1991 ) is subject to syntactic islands and that it should therefore be handled by a focus movement theory (contra Kratzer’s view). We present novel data that are at odds with Erlewine and Kotek’s conclusions and demonstrate the necessity of an island-insensitive mechanism to capture focus covariation. We revisit Erlewine and Kotek’s main arguments against such a system and show that they are systematically confounded. Moreover, removing the confounds cancels the force of the arguments, corroborating the central point of this article.

Modern generative linguistic theory furnishes a variety of general principles that appear to be at work in the grammar of all the world’s languages. One of the most basic and uncontroversial of these principles is that Agree/Move operates according to the constraint Attract Closest, which dictates that the closest suitable goal must be the target for the relevant operation (Rizzi 1990; Chomsky 1995, 2000; Richards 1998). The Polynesian language Niuean (Tongic subgroup, predicate initial word order, ergative-absolutive case system) presents a well known challenge to the universality of {Attract Closest}. The challenge manifests in a variety of distinct constructions in Niuean, but the best known case involves an operation first documented by Seiter (1980), which he terms “raising.” Specifically, Niuean raising appears to license an A-type dependency between the subject position of the matrix clause and the object position of an immediately embedded clause. This is illustrated in (1), where the semantic object of the embedded subjunctive clause, Sione, appears as the syntactic subject of the matrix predicate maeke. (1) To maeke a Sione$_{1}$ [ke lagomatai he ekekafo $t_{1}$]. fut possible abs Sione sbj help erg doctor ‘It’s possible the doctor can help Sione.’ (lit.: Sione is possible that the doctor help [him]) Granting that the filler-gap dependency in (1) is A-type, this is both a clear violation of {Attract closest} (Rizzi 1992; Chomsky 1995; Richards 1998) and a typological anomaly. Our aim in this paper is to argue that such apparent violations of {Attract Closest} are only that. Specifically, we show first that the challenge inherent in Seiter’s raising construction is pervasive throughout the language: in general, objects are accessible to syntactic operations even if the intervening clause-mate subject is also a licit target. In other words, Niuean clause-mate subjects and objects are equally accessible to syntactic operations. Then, we argue that this typologically uncommon equal-accessibility follows from the convergence of several otherwise independently attested operations: (i) a configurational system of case licensing, with a $v$P as the case computation domain; (ii) obligatory object shift to Spec($v$P); (iii) an EPP on T triggering V/VP-raising rather than DP externalization. The resulting basic clause structure is then as below, so that Niuean adheres to standard locality constraints. (2)

This chapter summarizes major results in the domain of experimental approaches to ergativity, focusing on three major topics. First, it discusses studies that explore the competition between accusative and ergative alignment, where researchers have attempted to derive the typological preference for accusative alignment from processing- and learnability based constraints. Next, it examines studies concerning the interrelated issues of long-distance dependencies and agreement. The unique dissociation of case and argument-hood in ergative languages has afforded researchers new means of testing conclusions regarding the privileged grammatical status of subject, the relative import and function of case and agreement in the grammar, and the origins of constraints on extraction in ergative languages and beyond. Given that linguists have only recently begun to conduct experimental research on ergative languages, we conclude by suggesting areas for future research where ergativity might provide genuine insights rather than just replicate existing studies of accusative languages.