Movement and alternatives don’t mix: Evidence from Japanese

Abstract

Certain quantificational elements (“interveners”) have long been known to disrupt the interpretation of wh-in-situ (Hoji 1985 and many others), but the correct description of the set of interveners and the nature of intervention effects have been the subject of continued debate. In Erlewine and Kotek (2017), we offer a new generalization concerning the nature of intervener-hood in Japanese: A quantifier acts as an intervener if and only if it is scope-rigid. We argue that this generalization is explained by — and in turn supports — Kotek’s (2017) account of intervention effects as reflecting a logical incompatibility between Predicate Abstraction and the computation of Rooth-Hamblin alternatives. In this paper we provide additional evidence in support of the above generalization, and test several of its predictions.