the boundaries of language

I effectively ended up taking last week off from posting here, without meaning to. I flew Austin to Glasgow on the 30th of May, and I've spent the last week or so getting over jetlag and slight illness, and attending the Arche audit in St Andrews.

To some extent, the group of people that were salient in Arche when I used to be around has fragmented (albeit in a very healthy way). We had people at the audit who had travelled from Austin, New York, Western Ontario, Spain, Israel and Leeds, and who hadn't all been in a room together for at least a year. The upshot was the Audit felt more like something of a reunion than anything else, and so the atmosphere in the seminars was terrific. It felt like a great example of how fruitful philosophical discussion can take place without becoming a competitive sport.

Mark Sainsbury kicked off the talks by discussing his new work on fictional discourse. In his last book Reference Without Referents (which really deserves to be much more widely read and discussed), Mark tentatively suggested that fictional names can be handled within his theory of reference by the same mechanisms that other seemingly empty proper names can be. In short, Mark adopts a (negative) free logic which allows that names can be meaningful - given axioms within a broadly Davidsonian truth-theoretic semantic framework - even if they have no bearer. The meaning of sentences with empty names as constituents can be determined compositionally but, as the 'negative' in negative free logic suggests, all such sentences will be false; Mark's suggestion in the book was that this account can be extended to fictional discourse, and our disposition to hear sentences like 'Sherlock Holmes is a detective' as true is ascribed to our tendency to conflate truth with fidelity to the stories or to fail to recognise an implicit (and, crucially, intensional) 'according to the fiction' operator prefixing the sentence.

Even in RWR Mark was explicit that this couldn't be the complete story, and so he's been trying to see what further progress can be made. He starts from the thought that the RWR story is ok for sentences like 'Sherlock Holmes is a detective' which are within the fiction so to speak - evaluable for fidelity to the story rather than genuine truth - but cannot be the right account of 'Sherlock Holmes is famous', intended as a claim about the fame of the fictional character created by Conan-Doyle and thus evaluable for genuine truth. The reason is that we accept the latter sentence as true simpliciter, but since 'Sherlock Holmes' doesn't refer to anything, negative free logic incorrectly rules that the sentence is false. In a radical departure from RWR, Mark's new suggestion is that in such occurances fictional names refer to abstract objects. The view that fictional names refer to such entities is hardly novel, but Mark twist is that he doesn't want to extend it to sentences made within the fiction like 'Sherlock Holmes is a detective', since although such an account delivers the same assignment of truth-value as negative free logic (false, since abstract objects can't be detectives), he thinks it puts any reasonable account of the creative and imaginative processes behind fiction beyond our grasp. This is really only the background to Mark's new work, which largely concerns making sense of a range of phenomena now he's adopted this dualistic account of discourse featuring fiction names; anaphora, cross-realm sentences (such as 'Sainsbury admires Sherlock Holmes for being a detective), and other hideously tricky cases. The discussion brought out just how tricky the issues are here, but it's fascinating stuff and I'm looking forward spending some time this coming year working on it.

We also had interesting discussions on the Caesar Problem, Zalta and Linksy's version of neo-logicism, modal realism (Andreas blogs about this talk here), vagueness and omniscience, and Williamson's pessimism regarding the prospects of a reductive analysis of knowledge. The audit closed (for those of us sadly unable to get up for Carrie Jenkins' talk at 9am on the last day) with Kit Fine talking about Frege's Puzzle and related issues in the philosophy of language. In essence, Fine's tactic is to show how Millianism, the thesis that the meaning of a proper name is exhausted by its referent, can avoid committment to the intersubstitutivity of coreferential proper name salva significatione by recognising that there may be semantic relations between proper names flanking the identity sign which are not necessarily known by speakers competent with each of the names taken singularly (that's probably not quite right, but it gives the flavour of the proposal, and Fine has a new monograph on this material coming out soon). Discussion centred on the proposal's compatibility with a plausible principle of compositionality, though Mark and I were also worried about how to extend the solution to Frege's puzzle to failures of intersubstitutability outside identity contexts (these kind of worries are nicely discussed by Wo).

I had some great conversations with various people, though a particular highlight was getting to hear Fine talk about how his views on vagueness have shifted in the 30 years since 'Vagueness, Truth and Logic' over dinner. All in all, the Audit was a great experience once again.

Labels: Conferences, Mark Sainsbury, Philosophy of Language