Antecedent-Correlate Harmony effects in sluicing

Abstract:

Sluicing is the ellipsis of a Wh-question leaving a Wh-phrase overt (1). In sluicing, the Wh- phrase (the Wh-correlate, adopting the terminology in Chung et al. 1995, in bold) typically corresponds to an indefinite in the antecedent (the inner antecedent, underlined), in some intuitive sense.

(1) Jack ate something, but I don’t know what Jack ate.

Dayal and Schwarzschild (2010) (henceforth, D&S) note that sluicing is sensitive to the Antecedent-Correlate Harmony (ACH) generalization (2), citing the data in (3a-3d).

(2) Antecedent-Correlate Harmony (ACH)
Wh-correlates & inner antecedent agree on the presence/absence of a contentful head noun. (D&S, pg. 100)

(3) a. * Joan was eating a doughnut. Fred didn’t know what.
b. Joan was eating a doughnut. Fred didn’t know which doughnut.
c. * Joan was eating something. Fred didn’t know which doughnut.
d. Joan was eating something. Fred didn’t know what.

ACH appears, at first, to be a syntactic generalization, though D&S take ACH to follow from a Merchant (2001)-style licensing condition for sluicing (contra Fox and Lasnik 2003, Chung et al. 1995); ACH violations are shown to result in the lack of informational equivalence required for sluicing to go through.
In this talk, I present new data, showing licit ACH violations with what exactly (4a-4d):

(4) a. Sally bought some food, but I don’t know what (exactly).
b. She owns a pet, but I can’t recall what (exactly).
c. Sally fired someone, but which employee isn’t clear.
d. Obviously, someone won the game, but I can’t recall which contestant.

I discuss questions the new data raise for D&S, and propose an account based on the new data. I argue that ACH effects stem from the interaction of the constraint in (5) (Barker 2012), with Heim (1987)’s semantics for what, and show that Merchant (2001)-style informational equivalence overgenerates.

(5) No Answers
The antecedent for a sluice cannot be an answer to the sluiced question.

I show that ACH is sensitive to the semantics of inner antecedent/Wh-correlate pairs. Specifically, certain inner antecedent head nouns set up the discourse in such a way as to ensure violations of (5) (unacceptable ACH violations), whereas others do not (acceptable ACH violations). The proposed account does not derive the deviance of (3c), which follows instead from independently needed constraints on deaccenting and ellipsis (cf. Romero 1998, Rooth 1992). Finally, I argue that (5) stems from the theory of ellipsis licensing proposed in Anderbois (2011), where sluicing is taken to be anaphoric to issues.

References:
Anderbois, S. (2011). Issues and Alternatives. PhD thesis, UC Santa Cruz.
Barker, C. (2012). Scopability and sluicing. in press.
Chung, S., Ladusaw,W., and McCloskey, J. (1995). Sluicing and logical form. Natural Language Semantics, 3:239–282.
Dayal, V. and Schwarzschild, R. (2010). Definite inner antecedents and wh-correlates in sluicing. In Staroverov, P., Braver, A., Altshuler, D., Fasola, C., and Murray, S., editors, Rutgers Working Papers
in Linguistics, volume 3, pages 92–114, New Brunswick, NJ. LGSA.
Fox, D. and Lasnik, H. (2003). Successive cyclic movement and island repair: the difference between sluicing and VP-ellipsis. Linguistic Inquiry, 34:143–154.
Heim, I. (1987). Where does the definiteness restriction apply? In Reuland, E. J. and ter Meulen, A. G. B., editors, The representation of (in)definiteness. MIT Press, Cambridge, Massachusetts.
Merchant, J. (2001). The Syntax of Silence. PhD thesis, UC Santa Cruz.
Romero, M. (1998). Focus and Reconstruction Effects in Wh-phrases. PhD thesis, University of Massachusetts Amherst.
Rooth, M. (1992). Ellipsis redundancy and reduction redundancy. Paper presented at Prague workshop on Context, Prague.

ALL WELCOME - PLEASE DO COME.