This Monday, we will have Ash Asudeh presenting his current research (in collaboration with Gianluca Giorgolo) on “Interpretive Intentions”. Please see the abstract below.
In this paper we discuss some issues that have to do with reference and modality. Our discussion is motivated by the following examples:
(1) Dr. Octopus punched Spider-Man but he didn’t punch Spider-Man.
[Same referring expression, non-modal context]
(2) Reza doesn’t believe that Jesus is Jesus.
[Same referring expression, modal context]
(3) Mary Jane loves Peter Parker, but she doesn’t love Spider-Man.
[Different referring expression, non-modal context]
(4) Reza doesn’t believe that Hesperus is Phosphorus.
[Different referring expression, modal context]
The fourth example is just an instance of Frege’s famous opacity puzzle, but we see that we can actually construe this as a more general problem. In one dimension, we either have a purely assertive context, as in (1) and (3), or we have some linguistic marker that changes the interpretive perspective of the expression (‘believe’, in this case), as in (2) and (4). In the other dimension, we see different ways to refer to the same entity: either we use the same expression twice, as in (1) and (2), or we use a different but co-extensional expression the second time, as in (3) and (4). The interplay of these two seemingly unrelated parameters generates some challenging readings for the expressions we want to discuss. Examples such as (4) are quite familiar from the literature on intensionality. We will instead focus on the case of sentences like (2). The interesting fact about such examples is that beside the logically implausible reading that attributes to Reza the non-belief in a tautology, we also have another reading in which Reza is said to not believe that the entity he calls Jesus has has the same properties as the entity that the speaker or some other individual calls Jesus (e.g., being a deity). We present a compositional analysis that in a uniform way generates all the readings associated with all four examples above
The intuition behind our analysis is that in these examples we observe that some default interpretation (often the speaker’s) is pitted against some alternative interpretation. We formalize this intuition by letting certain linguistic expression manipulate interpretation functions, so as to mix together different lexica. This allows us to provide a general account of (1)-(4) that treats ambiguities created by intensional contexts non-scopally, rather than relying on the typical de re/de dicto scopal account, which would fail to fully generalize to this data set and which has certain specific problems that we will discuss.
Our analysis follows a recent suggestion in the formal semantics literature that tries to extend the expressive power of the meaning language in a principled and controlled way with the use of (strong) monads. Monads are mathematical structures that have found application in the theory of programming language semantics to model ‘impure’ computations. They encapsulate the notion of mapping a space of values (for instance the values of a simply-typed theory constructed over the basic e and t types) to a space populated by more complex values. The fundamental property of monadic mappings is that they allow us to freely combine standard simple values with complex ones. This allows us to consider more complex types of meaning only when truly necessary, avoiding the notorious problem of generalizing our lexical entries to the worst case.”
We will meet at the usual time and place, 2:30 to 4pm in VSSM 5220.