Poster: Left-Peripheral Interactions in English Imperatives

January 4, 2013
LSA Annual Meeting, Boston, Massachusetts

Abstract

Many syntactic accounts of imperatives involve the CP layer (e.g. Han 2000, Zanuttini 2008), but they often adopt a unitary CP or ad hoc imperative-specific projections. I present evidence that independently derived facts about the English left periphery not only accommodate imperatives, but predict interactions with information-structural movement, negation, and Wh-extraction in imperatives.

Following Haegeman (2004), I divide the English CP domain into three projections: 

(1) Subordination/Force/Topic > Focus > Finiteness

As in Rizzi (1997), Force determines clause type; no dedicated imperative projection is required. The highest head, i.e. the complementizer, bears features [±Sub, Force{dec/int/imp}, ±Topic] for the three categories it subsumes.

Clause-typing: English clause-typing cannot require verbal movement to Force, contra Koopman’s (2007) model for Dutch. In English, the verb does not raise as high as Force, as it must follow both contrastive topics (in FocusP) and imperative subjects (in FinP), whose relative order is also fixed (3).

(2) These stocks buy ___ immediately; those avoid at all costs!

(3) *Everyone these stocks buy immediately!

(4) **Buy these stocks everyone immediately!

Topics: In English, the availability of topicalization covaries with other features of the complementizer.  ØMATRIX.DEC, ØIMP , that, and ØSUBORD.DEC do not equally license topics. Only ØMATRIX.DEC can bear a [+Topic] feature and move a (non-contrastive) topic to its Spec.

(5) The book ØMATRIX.DEC, John bought ___.

(6) *The book ØIMP buy ___!

(7) *John said the book that he bought ___.

(8) *John said the book ØSUBORD.DEC he bought ___.

Contrastive topics may be marked [+Focus] and moved to FocusP; they are also prosodically marked (9). 

(9) John said [ØSUB.DEC  [the book ØFOCUS [he bought ___]]], not the magazine.

Focalization is possible even in imperatives (see 2 above), since Focus is distinct from Force.

Negation: Subjects cannot appear to the left of negation in imperatives. If subjects are hosted in Finiteness and polarity is associated with Focus (Zanuttini 1997), this follows directly from the fact that Focus > Finiteness.

(10) *You don’t do that! (high subject)

(11)    You, don’t do that! (vocative)

(12)    Don’t you do that! (low subject)

Extraction: It is impossible to Wh-extract from an imperative. English has true embedded imperatives, as proven by binding data (Crnic and Trinh 2009). Any extracted item must pass through the Sub/Force/Topic projection, but when headed by ØIMP, it is incompatible with Wh-elements (14–15). Contrast an infinitival embedded command, which lacks a complementizer, and thus allows extraction (16).

(13) Johni said [ØIMP send hisi mother to the store].

(14) *Where did Johni say [ØIMP send hisi mother __]?

(15) *Who did John say [send __ to the store]?

(16)   Where did John say [to send his mother __]?

Beyond English: Other languages license topics in imperatives. Korean does so (17); it also distinguishes Subordination (Bhatt and Yoon 1992) and Force (Zanuttini et al. to appear), suggesting that Subordination, Force, and Topic are all distinct projections.

(17) Chayk un    ilke-ra!
       book   top read-imp
       “Books, read!"