One way creatures of fiction seem to differ from real things is in their essential properties. While you and I might not have done many of the things we did do, Anna Karenina could not, surely, have been other than a lover of Vronsky. Is that right? Not straightforwardly: while it is true that “Necessarily, someone who was not a lover of Vronsky would not be Anna” it is also true that “Someone who was necessarily a lover of Vronsky would (...) not be Anna”. I use a framework developed by Stalnaker to explain this, and to shed light on the semantics of fictional names. (shrink)

One way creatures of fiction seem to differ from real things is in their essential properties. While you and I might not have done many of the things we did do, Anna Karenina could not, surely, have been other than a lover of Vronsky. Is that right? Not straightforwardly: while it is true that “Necessarily, someone who was not a lover of Vronsky would not be Anna” it is also true that “Someone who was necessarily a lover of Vronsky would (...) not be Anna”. I use a framework developed by Stalnaker to explain this, and to shed light on the semantics of fictional names. (shrink)

I defend the claim that it is possible for thoughts to be context-sensitive. Assuming that a thought is a sentence of Mentalese and content is a function from indices to truth-values, then a thought, T, is context-sensitive IFF at least one of the following three conditions are met: T exhibits character-underdeterminacy, where T is character underdetermined iff a component of T makes an explicit reference to the context to establish content. T exhibits type-underdeterminacy, where T is type underdetermined iff there (...) are tokens of T that have distinct truth-values. T is token-underdetermined, where T is token underdetermined iff for some possible states of affairs its truth-value is indeterminate. Additionally, there must be a mechanism by which the content of T can be determined by appeal to the context. Otherwise T is not sensitive to the context. The most significant objection to the possibility of context-sensitive thoughts comes from a regress argument. The concern is that any process of removing context-sensitivity from a thought will token an additional thought. If that further thought is context-sensitive, then another thought will have to be tokened. This process will only terminate when a non-context-sensitive thought is arrived at. Yet this seems to show that context-sensitivity in thought, if possible, can only be a contingent, detachable feature of thoughts. To counter this argument, I will present three versions of context-sensitive thought, mirroring each of -. These views make use of a range of relationships between thoughts and their contents or extensions that are affected by the context, but I will argue they do not require the generation of an additional thought. This provides a range of views on which thoughts can be treated as essentially context-sensitive. (shrink)