No abstract

No abstract

This paper presents a relative theory of tense which accounts for tense behaviors in English in both matrix and embedded contexts. Crucial to this proposal is novel data which shows that the temporal orientation of finite embedded clauses is constrained by the attitude verbs which embed them. While such constraints have been observed in modal auxiliaries and non-finite-embedding attitude verbs, finite-embedding-attitudes present an important new perspective because they show that whatever constraints are introduced by the modal must be able to coexist with a tensed prejacent. This proposal is compositional, and a formal analysis is given, by which modals, including attitude verbs, may constrain the temporal possibilities of an embedded clause, without themselves introducing a reference time argument. The constraining effect of attitude verbs also eliminates the need to stipulate a general Upper Limit Constraint, and allows for a means to capture correspondences between modal flavor and temporal orientation. The result is an elegant accounting of embedded tense with empirical coverage superior to previous proposals.

No abstract

1 The Problem Discourses like those pointed out by Lewis (1979), which involve modals whose domains appear to shift contextually, have not yet been given a formal treatment in the literature.(1) a. Yaron: This must be a pen. I'm looking right at it. b. Itamar: Not so; you could be the victim of a deceiving demon.A similar example not addressed by Lewis, involving teleological rather than epistemic modality, can be seen here:(2) a. Alice: I want to go outside, but I don't want to get wet. b. Bryan: You have to wear a raincoat. c. Alice: No I don't. I could cover every inch of my skin in duct tape. d. Bryan: I suppose that's technically true, but you get my drift.Lewis's basic take on these cases is that they involve a modal being used with a particular domain of admissibility; this domain is then expanded via accommodation to render the original claim false. For descriptive purposes, I will refer to the kind of domain shifting in (2c) as Lewis-shifting; correspondingly, I will refer to the kind of domain restriction that Bryan is employing in (2b) as Lewis-restriction. L-shifting has some interesting properties which set it apart from other kinds of contextual variability. First is pedantry. In (2c) Alice comes across as pedantic -we can imagine that she could have accepted Bryan's utterance and it would have been perfectly felicitous; compare to (3). However, her conversational move is not uncooperative in the strongest sense, i.e., it does not crash the discourse; compare to (4).(3)Bryan: You have to wear a raincoat. Alice: Yeah, that seems right.(4) Bryan: You have to wear a raincoat. Alice: # What? How would a condom help me in this situation?The second property of L-shifting is a kind of faultless disagreement. Bryan is not just factually mistaken, as indicated by his unapologetic response in (2d). This is important because it means that Bryan's original utterance in (2b) is in some sense false, but in some sense true. Compare to a genuine case of factual wrongness, in a context where there is an umbrella available, in (5). Bryan cannot use the same unapologetic concession; he must fully concede the wrongness of his previous assertion.(5) Bryan: You have to wear a raincoat. Alice: I don't have to. I could just use an umbrella. Bryan: #Well, I guess you're technically right. But you see my point.Bryan': Oh! You're right. I was wrong; wearing a raincoat is not the only option.