Thursday, October 15, 2009

Kaplan and the shotgun

This is a non-serious post connected to Dan LdS's talk on Wednesday (which I suppose took place...). To be more precise, it's conencted to Egan's paper that Dan was considering. I haven't re-read the paper, but I remember I wasn't convinced that the Kaplanian framework has serious problems with (at least some of) the examples Egan is giving. Also, it struck me as a bad thinig that Egan doesn't think it necessary to sharply differentiate his view from the multiple-utterances view. If I remember correctly, Egan's preferred view is that an utterance produced at a given context of utterance expresses a multiplicity of propositions, which proposition is expressed being determined by the context of utterance together with the context of assessment (don't remember whether he actually uses this latter term, but let's stick to it for the moment). On the other hand, the multiple-utterances view has it that each time different assessors are presented with a sentence produced in a context of utterance (different from the respective contexts of assesment), that counts as a different utterance of the sentence at issue. [Let me here note that even the shotgun metaphor that Egan uses is closer to the multiple-utterances view than to his preferred one: the shotgun is the producing of the sounds, the bullets are the different utterances and the wounded people are the different propositions expressed. (Yes, this is Romanian mafia speaking. Shhhh...)] Now, the thing is I'm not sure Kaplan has problems with the multiple-utterances view. In defining truth in a context, Kaplan speaks of occurences of sentences. Could these be equated with utterances? If not, is the relation between occurences and utterances such that an occurence of a sentence can only result in one utterance being produced? My "I'll be back in 5 minutes" post on my office door constitutes, I take it, just one occurence of the sentence. Yet, it seems to me to spread around a multiplicity of utterances - one for each different minute (second?) the note spends hung on my door. Of course, to get across a content, someone needs to read the note, so that that someone grasps the proposition, but I don't think it's so implausible that there are propositions out there that no one grasps :-) (This is similar to the gravestone poem example, right? What does Egan say about that - how does he reject it? Sorry for being lazy...) So maybe Kaplan's talk of occurences gets things wrong, but defining truth for utterances instead of truth for occurences would fix it - in such a way that the multiple-utterances view is compatible with the "modified" Kaplan. Even if one finds ungrasped propositions hard to swallow, each grasped proposition gets expressed in a context in which there is both a "speaker" and an audience, located in time and space, which seems to me to be the notion of context Kaplan is using.
If you find this too confusing or stupid, please ignore it. :-)


Dan López de Sa said...

Hi Dan!

The first part of the paper was actually to argue (with Egan) against the interpretation of his audience’s “positional” context as context of assessment, in MacFarlane’s sense ;-).

Re the core or your post, though, I interpreted Egan as suggesting that the phenomenon of audience-sensitivity motivates a refinement of the Lewisian basic notion of a sentence being true at a (Lewisian) context (= Kaplan context, according to my (and Lewis’) interpretation of Kaplan), positing both a context for the speaker and a context for the audience. (I contended that contexts were already flexible enough as to accommodate the phenomenon, a position that Egan himself considers in a couple of footnotes, acknowledging discussion to Carrie Jenkins.)

Part of the discussion did concern how the phenomenon could seem to pose a problem, or allow for an alternative form of accommodation, when (various conceptions of) particular acts of uttering, as opposed to sentences in context, are involved. Perhaps this a good occasion to invite people to elaborate ;-)!

m g-c said...

I thought the discussions about this on the occasion of Dan LdS's presentation at the LOGOS-Jean Nicod meeting were clarifying. I would say that the examples only show that with one and the same "utterance" (individuating utterances relative to the physical properties of the uttering-output) one can make different "utterances" (now individuating them on the basis of constitutive semantic properties, such as the contribution of indexicals). Egan also appears to go for this (pp. 269-71), although he is not ultimately fully committal. But I agree with Dan's main claim, that there is a natural way of interpreting Lewis' framework of sentences, indexes and context that can also accommodate the examples.