The Diversity of Wh-in-Situ
In short, participle agreement in Passamaquoddy bears a striking re semblance to Chamorro ‘‘wh-agreement’’; relative tense in Coptic bears a striking resemblance to Chamorro ‘‘operator-C agreement.’’ These resemblances can be explained if all four patterns are instances of the wh-agreement phenomenon, and wh-agreement is the morphological re flex of Agree holding between a probe that bears the feature [X] and a goal that bears the feature [wh]. When both probe and goal also bear an operator feature, what emerges is the Coptic/‘‘operator-C agreement’’ pattern; when the probe bears no other relevant features, what emerges is the Passamaquoddy/‘‘wh-agreement’’ pattern. Crucial to this account is the assumption that some instances of wh-in-situ are only apparent, in that they arise as a consequence of wh-movement in overt syntax, followed by pronunciation of the lowest as opposed to the highest copy. (For discussion of another type of case in which wh-in-situ is only apparent, see Simpson and Bhattacharya 2003.) In this section, we situate our conclusions in the context of recent theoretical discussions of the diversity of wh-in-situ.
Our arguments that instances of wh-in-situ in Passamaquoddy and Coptic are derived via syntactic wh-movement before the derivation is shipped to the phonology do not entail that every instance of wh-in-situ in natural language involves a hidden movement configuration. In fact, our proposal is fully compatible with the idea that there is no unitary wh-in-situ phenomenon: some types of wh-in-situ may be derived by movement in overt syntax, while other types may involve covert movement after Spell-Out (Nissenbaum 2000; Richards 2001), and still others may participate in no movement whatsoever (Reinhart 1997). The possibility that there are distinct types of wh-in-situ was first raised by Cole and Hermon (1994). The claim has been investigated in depth by Cheng and Rooryck (2002), who demonstrate the diversity of wh-in-situ not only across languages but also within individual languages (such as European Portuguese and Mandarin). Our analysis is compatible with the spirit of Cheng and Rooryck’s proposal, though not with the details, since we document a type of wh-in-situ that they do not recognize.
The claim that there is no uniform analysis of wh-in-situ makes a pre diction when combined with our assumption that wh-agreement tracks the history of wh-movement before Spell-Out. The prediction is this. One might expect to find languages that di¤er from Coptic and Passama quoddy in that wh-agreement flags overt wh-movement but, crucially, not wh-in-situ. But such languages should provide other evidence besides the absence of this special inflection that wh-in-situ does not involve movement in overt syntax. Palauan seems to be just such a language.
