Particles and Deletion

Much current theorizing analyzes ellipsis as a PF operation silencing designated syntactic domains from which focal constituents are extracted prior to deletion.¹ In many cases, this requires exceptional evacuation movements that are not observed—and often altogether illicit—in corresponding nonelliptical forms. We introduce a novel set of data involving modal particles in German that militates strongly against obligatory movement of ellipsis remnants, suggesting that deletion instead applies to independently generated surface forms, in a manner analogous to deaccenting. The discussion will focus on clausal ellipsis only, ignoring other, crosslinguistically less common types of incompleteness such as VP-ellipsis and pseudogapping.

1 Move and Delete?

Ever since Merchant's (2004) groundbreaking work on short answers, much of the theoretical literature has accepted a conception of ellipsis as a bona fide syntactic phenomenon, involving both movement (in narrow syntax) and deletion (at PF). To illustrate, consider B's short answer in (1), an instance of clausal ellipsis. Merchant argues that a fragment of this kind is derived by Ā-movement of the remnant to a left-peripheral focus position (2a) and subsequent deletion of TP (2b), paralleling the derivation of sluicing (Ross 1969, Merchant 2001).

(1)

A: Who did Mary talk to yesterday?

B: John.

(2)

a. [_FP John F_E [_tp she talked to t yesterday]]→.

b. [_FP John F_E [_tp she talked to t yesterday]]

Merchant proposes that deletion is triggered by an E(llipsis)-feature borne by a licensing head (F in (2), or C in sluicing). Remnants move to the specifier of this head, thereby escaping the ellipsis site (TP, on this approach). This move-and-delete approach (henceforth, MDA) is the predominant theoretical framework for the study of ellipsis in the current generative literature, often adopted without argument (see, e.g., Brunetti 2003, Heck and Müller 2003, Wang and Wu 2006, Toosarvandani 2008, Nakao 2009, Boone 2014, Ortega-Santos, Yoshida, and Nakao 2014, Sailor and Thoms 2014, Weir 2014, Yoshida, Nakao, and Ortega-Santos 2015).

In support of the MDA, Merchant (2004) presents a number of constraints on fragments that appear to derive from constraints on movement. However, it remains to be seen whether these restrictions can receive alternative, more principled explanations—for example, on the basis of pragmatic factors (see Weir 2014, Ott and Struckmeier 2017). Be that as it may, it is clear that the MDA introduces a systematic asymmetry between elliptical and nonelliptical forms, in that the derivation of the former necessitates movements not required for the latter. As a result, the MDA requires movements that are otherwise illegitimate even in simple cases like (1B), since English does not normally front answer foci (see Brunetti 2003 on Italian). Thus, the natural nonelliptical variant of the fragment in (1) in the same context is (3a), not (3b) (here and below, capitals indicate stress, italics deaccenting).

(3)

a. She talked to JOHN yesterday.

b. #JOHN she talked to yesterday.

The assumption that (1B) derives from (3b), rather than from the independently available (3a), has significant implications for acquisition, burdening the learner in effect with acquiring a special syntax for elliptical constructions. The problem is even more severe for verb particles, bare-quantifier fragments like everyone, negative polarity items like any book, nonextractable PPs, and other categories that resist Ā-movement but can nevertheless surface as ellipsis remnants (see Valmala 2007, Weir 2014, and Dagnac to appear for a battery of pertinent examples).

The problem of exceptional movement is further amplified by cases involving multiple remnants, which necessitate otherwise illicit multiple fronting. For instance, multiple sluicing in English (4) requires multiple wh-fronting; why-stripping (5B) requires exceptional fronting of the second remnant; and swiping (6) requires inversion of the fronted wh-phrase and its associated preposition (Merchant 2002).

(4) Some student gave a gift to some teacher, but I don't know which student to which teacher.

(5)

A: Peter kissed Mary last week.

B: Why Mary?

(6) John went to the movies, but I don't know who with.

Given that the MDA insists on a systematically exceptional syntax for elliptical forms, it faces the challenge of having to...