A Note on the Movement Analysis of Gapping

1 Complex Gaps and Remnant Movement

Gapping had traditionally been classified as a subcase of ellipsis until Johnson (1996, 2009) proposed an alternative analysis in terms of across-the-board (ATB) verb movement. His analysis of (1) is given in (2), and it rests on the following assumptions: (a) coordination happens at the vP level; (b) as a consequence, the subject of the first conjunct raises to Spec,TP, whereas the subject of the second conjunct remains in Spec,vP;¹ (c) the verb undergoes ATB movement out of vP to a position that Johnson identifies as PredP.

(1) Randy drank scotch and Amy [___] rum.

(2) [_TP Randy

Assumptions (a) and (b) are also found in Coppock 2001 and Lin 2002, and they are meant to capture the fact that negation and modals that appear to be embedded in the first conjunct actually take scope outside coordination (Siegel 1984).²

(3) We can't eat caviar and him [____] beans.
               [¬◊(p∧q)]
(i.e., it is not possible that simultaneously we eat caviar and he eats beans)

Johnson's innovation, and the focus of this squib, is assumption (c), namely, the idea that the gap arises as a consequence of ATB verb movement, rather than ellipsis. This idea offers an elegant explanation of some otherwise puzzling properties of gapping: the facts that, unlike VP-ellipsis, gapping may only occur in coordinate structures (4) and cannot be embedded (5). The illicit environments bleed the application of either ATB extraction or verb movement (see Johnson 1996, 2009 for details).

(4)

1. a. Sandy plays the guitar,{and/or/*because/*after/*if/*better than} Betsy [___] the harmonica.

2. b. Sandy plays the guitar {and/or/because/after/if/better than} Betsy did/does [___] (too).

(5)

1. a. *Amanda went to Santa Cruz, and Bill thinks that Claire [___ ] to Monterrey.

2. b. Amanda went to Santa Cruz, and Bill thinks that Claire did [___] too.

Nonetheless, a movement analysis of gapping also has its own complications. This squib focuses on a problem already identified by Johnson (1996, 2009), namely, cases where the gapping site contains more elements than just a verb. Johnson refers to such cases as complex gaps and gives examples where the additional material is a direct object (6a), a small clause subject (6b), or a small clause complement (6c). To this list, we can add complements to control verbs (6d). These examples clearly cannot be derived via ATB verb movement if verb movement is head movement. To solve this problem, Johnson proposes that English verb movement is actually remnant predicate movement (see Kayne 1998, Baltin 2002), so that the verb may pied-pipe some VP-internal constituents. As an illustration, (7) gives the derivation for (6d). For simplicity, I omit any structure above the vP level and the asymmetric extraction of the first-conjunct subject.

(6)

1. a. Phil read things quickly, and Mike [___] thoroughly.

2. b. Some found Mittie clever with pictures, and others
   [___] good with children.

3. c. I made Sal fond of it on Tuesday, and [___] Holly
   [___] on Wednesday.
   [= . . . and I made Holly fond of it on Wednesday]

4. d. Randy wants to write a novel, and Amy [___] a play.

(7)

A prediction of this analysis (which Johnson does not discuss) is that, if the gapped strings in (6) have undergone movement, then they should also be movable in other environments. This, however, is incorrect: as (8) shows, the same strings that can be gapped cannot be topicalized.³ This contrast is unexpected.

(8)

1. a. *Read things, Mike (did) quickly.

2. b. *Find Mittie, some (did) clever with pictures.

3. c. *Make fond of it, I (did) Sal on Tuesday.

4. d. *Want to write, Randy (did) a novel.

Importantly, (8) cannot be explained by appealing to a generalized ban on remnant topicalization in English, regardless of what its underlying cause could be: Huang (1993) shows that the fronted predicate in (9a) is a remnant predicate containing the vP-internal trace of the subject. The same holds for (9b).

(9)

1. a. [t_i read a book], Amy_i certainly will.

2. b. ?[Believed t_i...