It has often been noted that whatever process is responsible for VP ellipsis can ignore certain aspects of lexical identity and morphosyntactic form. Consider the examples below, similar to ones cited by Sag and Hankamer (1984: 331–333), McCawley (1988: 764), Kitagawa (1991: 528), Fiengo and May (1994: 218), and Kempson et al. (1999: 230). In these examples, the elided VP is understood as (a), not as though it were word-for-word identical to the antecedent VP.

(1)	Jill:	Have you answered my message?
	Jack:	No, I haven't __.
		(a) [answered your message]
(2)	Jane:	I really appreciate you.
	Jill:	You do __?
		(a) [appreciate me]
(3)	Mike:	I thought you'd be elected.
	Jack:	Yes, I believe you did __.
		(a) [think I'd be elected]

Kitagawa (1991: 528) observes that VP ellipsis in examples of this sort allows only a strict reading: the pronoun in the elided VP and the corresponding pronoun in the antecedent VP refer to the same individual (e.g. Jill in (2)). He and others have concluded from this that the ellipsis process, whatever it is, is sensitive to a pronoun's referential index but indifferent to its person, number, gender, and other syntactic features. I think it is intuitively clear why this is so. Linguists and philosophers alike have acknowledged that pronouns have a special linguistic status (see e.g. Benveniste 1966, Jakobson 1971, and Kaplan 1989): they are expressions whose reference depends on the context of use. Because a pronoun's reference can shift from utterance to utterance, its syntactic features do a poor job of identifying a fixed referent in any way that persists across utterances. (In current linguistic theory, the abstract device that serves this function is the referential index). The point is obvious for person features. It is also valid for number and gender features, given that number and gender impose presuppositions on the referent's possible identity that may not remain (as) informative as the context changes. Consequently, the ellipsis process can afford to ignore them all.

Assuming something like this, what accounts for the following?

(4)	Jack:	I don't want to be divorced from you.
	Jill:	Well, I do __!
		(a) [want to be divorced from you]
(5)	Jill:	For instance, I would be reluctant to criticize you in public
	Jane:	I wouldn't be __.
		(a) [reluctant to criticize you in public]
(6)	Jack:	You pushed me first!
	Mike:	No, you did __!
		(a) [push me first]

In the reading of these examples that interests me (shown in (a)), the elided VP is understood as word-for-word identical to the antecedent VP. Notice that in this reading the pronouns in the elided and the antecedent VP's refer to different individuals (e.g. in (4), Jill is saying that she wants to be divorced from Jack). If one assumes, with Sag (1976), Fiengo and May (1994), and others, that the ellipsis process demands identity of reference or identity of referential dependency, then the possibility of a sloppy reading in (4–6) seems mysterious. The puzzle is that neither kind of relation seems to link the pronoun in elided VP to the corresponding pronoun in the antecedent VP. But the interpretation is nonetheless allowed.

The solution I want to suggest to this puzzle draws on a further piece of traditional linguistic wisdom concerning pronouns: I and you are opposed in that you picks out the participant in the speech event who is not the speaker. Likewise, I picks out the participant who is not the hearer. (For a particularly explicit statement, see Benveniste 1966: 232). To put the point differently, I and you have a relational use: these pronouns stand in the "salient other" relation, and that relation can form the basis for a referential dependency (cf. Heim, Lasnik, and May on other). I claim that this referential dependency can be accessed by the process responsible for VP ellipsis. When that happens, the result is the sloppy reading of (4–6).

Not all speakers I have consulted are comfortable with sloppy readings involving the "salient other" relation. Some accept the sloppy reading more readily if the larger context involves conflict or—as Jim McCloskey observes—negotiation, presumably because both kinds of contexts serve to bring the "salient other" relation to the fore. Compare (4) and (7) with the more harmonious (8).

(7)	Jane:	I'll negotiate with you.
	Jill:	Okay, I will __, too.
		(a) [negotiate with you]
(8)	Jack:	I love you.
	Jill:	I do __, too.
		(a) ??[love you]

In (8), Jill's response seems to be more a statement of self-love than the completion of a mutual declaration of love.

On the other hand, if the context is structured so as to highlight the "salient other" relation, it is possible for the sloppy reading to emerge even when VP ellipsis involves pronouns (more generally, DP's) other than I or you. To see this, consider first

(9)	Jane:	Jack would be reluctant to criticize you in public.
	Mike:	Jill wouldn't be __.
		(a) [reluctant to criticize me]
		(b) [reluctant to criticize you]

Without further information, the strict reading of the elided VP in (9a) seems necessary or highly preferable. But as pointed out by Bill Ladusaw, if one knows that Jack and Mike are partners and Jill and Jane are partners, then the sloppy reading in (9b) is easy to accept. Similarly, in (10), an example which contains neither I nor you but does involve a situation of conflict, the most natural interpretation of the elided VP is the sloppy reading that depends on the "salient other" relation:

(10)	At the same time as Jilli's lawyer claimed that Jackj had been unfaithful to heri, Jackj's lawyer insisted that Jilli had __.
	(a) [been unfaithful to himj].

References

Benveniste, Emile. 1966. 'Structure des relations de personne dans le verbe'. In Problèmes de linguistique générale, 225–236. Paris: Gallimard.

Fiengo, Robert and Robert May. 1994. Indices and identity. Cambridge, MA: MIT Press.

Heim, Irene, Howard Lasnik, and Robert May. 1991. 'Reciprocity and plurality.' Linguistic Inquiry 22: 63–101.

Jakobson, Roman. 1971. 'Shifters, Verbal Categories, and the Russian Verb.' In Selected Writings, Vol. 2: Word and Language, 130–147. The Hague: Mouton.

Kaplan, David. 1989. Demonstratives: An essay on the semantics, logic, metaphysics and epistemology of demonstratives and other indexicals. In Themes from Kaplan, ed. by Joseph Almog, John Perry, and Howard Wettstein, with the assistance of Ingrid Deiwiks and Edward N. Zalta, pp. 481–563. New York and Oxford: Oxford University Press.

Kempson, Ruth, Wilfried Meyer-Viol, and Dov Gabbay. 1999. 'VP ellipsis: Toward a dynamic, structural account.' In Fragments: Studies in ellipsis and gapping, ed. by Shalom Lappin and Elabbas Benmamoun, 227–289. New York and Oxford: Oxford University Press.

Kitagawa, Yoshihisa. 1991. 'Copying identity.' Natural Language and Linguistic Theory 9: 497–536.

McCawley, James D. 1988. The syntactic phenomena of English. Vol. 2. The University of Chicago Press.

Sag, Ivan A. 1976. Deletion and Logical Form. Ph.D. dissertation, MIT, Cambridge, MA.

Sag, Ivan A. and Jorge Hankamer. 1984. 'Toward a theory of anaphoric processing.' Linguistics and Philosophy 7: 325–345.

Endnote

^* Happy Birthday, Jorge! And special thanks to Chris Kennedy, Bill Ladusaw, and Jim McCloskey for comments. [Back]