On the unified analysis of three types of relative clause construction in Japanese, and on the "salient reading" of the internally headed type. A reply to Erlewine & Gould (2016)

In this reply, we argue against two proposals made by Erlewine & Gould (2016), and in favor of the following two theses: [i] Japanese internally headed relative constructions disallow an interpretation that is crucially based on contextual salience, and allow only an interpretation that reflects the content of their (explicit or implicit) internal head(s). [ii] None of Erlewine & Gould's attempts to provide independent support for their unified analysis of internally headed, externally headed and doubly headed relative constructions of Japanese is tenable. Therefore, these writers failed to provide convincing justification for abandoning the approach developed by Grosu & Landman (2012) and Landman (2016) with respect to the analysis of the internally headed relatives of Japanese, an approach which moreover yielded explicit accounts of two types of data for which Erlewine & Gould have no obvious analysis.