The Weekly Newsletter of MIT Linguistics

Colloquium 11/7 - Klaus Abels

Speaker: Klaus Abels (UCL)
Title: Guess what else!
Date: Friday, November 7th
Time: 3:30-5:00p
Place: 32-141

Ross’s seminal paper on sluicing, that is, elliptical wh-questions of the type in (1), contains two generalizations that have driven analyses of sluicing in radically different directions.

(1) Somebody just left. – Guess who!

On the one hand, Ross observes that, at least in languages where this is directly observable, the wh-phrase in the elliptical question must bear the same case as its perceived correlate in the antecedent sentence, as in the German example in (2).

(2) Er hat jemandem geholfen, aber er verrät nicht {wem | *wen | *wer}.
he has someone.dat helped but he divulges not who.dat | who.acc | who.nom}
‘He helped someone but he won’t divulge who.’

On the other hand, sluicing ameliorates island constraints, as seen in the contrast between the acceptable (3) and the ungrammatical full version (4).

(3) They want to hire someone who speaks a Balkan language, but I don’t know which.
(4) *They want to hire someone who speaks a Balkan language, but I don’t know which Balkan language they want to hire someone who speaks.

The case matching effect in (2) is often taken as a straightforward argument for the presence of syntactic structure at the ellipsis site which is (nearly) identical to the syntactic structure of the antecedent. The island amelioration effect seen in (3) suggests the exact opposite.

In the first part of this talk, I will report on joint work with Gary Thoms. In this work, we use contrast sluices in languages with resumptive pronouns as a diagnostic tool. Contrast sluices are examples like (5), where the correlate in the antecedent clause is definite and the sluice asks about the identity of a different relevant entity.

(5) He gave the car to his son and guess what else!

The cross-linguistic distribution of island repair in contrast sluices strongly suggests that sluicing does not literally repair island effects. It also strongly suggests that ellipsis identity for sluicing in general cannot be understood as strict syntactic identity.

This conclusion calls for a careful evaluation of the case-matching effect, a task that will be taken up in the second part of the talk. Finally, a possible way forward will be suggested based on Fox and Katzir’s structural theory of focus alternatives.