But now consider this free ref. in Levinson, in "Legacy of Grice", (ed. K. Hall):
'semantic informativeness of an utterance'
'what Bar-Hillel & [sic] Carnap called'
no such reference in the reference -- but of course he is wanting us to figure out, since this is basic stuff: it's Bar-Hillel of vintage 1946 Mind (pragamtics as waste-basket) and Carnap of God knows when (I know who does!).
But consider yet again "semantic informativeness".
In my PhD I called this 'byte by byte'. Surely informativeness _is_ a *semantic*, not pragmatic, notion. Yet Carnap, by applying it to 'utterance' and in the realm of empirical pragmatics, may be thinking otherwise. For Grice it is all semantic alright. After all, part II of is WoW is "Explorations in semantics and metaphysics: no place, really, for pragmatics.
Quantity and informativeness.
If Bar-Hillel and Carnap had something definite in mind (in their minds, rather) when they spoke, Levinson says they did, of the 'semantic informativeness of an utterance' what was _Grice_ having in mind?
Two thoughts:
i. Vintage Grice: 1961. His example
(a) The pillar box seems red.
(b) The pillar box _is_ red.
Doesn't carry, because he _is_ forced to admit or grant that none is more 'informative' than the other: the one is what Berlin, et al, would call an utterance in a phenomenalist language; the other in a physicalist language. And neither entails the other. Grice's idea that 'strong' and 'weak' may apply here, while it did appeal to K. DeRose, it does not appeal _me_. Horn's idea of formalising this onto a scale
ii. The Grice of the William James. There's really no mention of 'weak' and 'strong' in WJL, but just 'informativeness'. Grice is on 'informative', 'fails to be informative enough', is 'underinformative', or 'overinformative'. No need to retreat to what he possibly thought superseded theory of the inconclusive weak and strong of 1961.
The question remains what _Levinson_ is thinking about. He is a very clever fellow with a taste for formalising the correct things in the correct terms, so, despite his unphilosophical illustrations with furrin lingos (I love him!) he should give us a clue! (But will he?)
Etc. (When I say he has the good taste of formalising what is worth formalising, I'm having in mind his co-work with J. D. Atlas, vintage as 1973, What is an implicature?, presented at the Performadillo -- and making use for the first time in the history of Oxford philosophy -- Atlas was Wolfson -- of the "Practical Inferences" of KENNY in reply to Hare! Fiat lux!)
Posts
In connection with informativeness, I would look not at how much information is supplied but how much curiosity is unsatisfied. If I ask you whether the horse is green, "the pillar box seems red" is not "informative" at all. (And what's a pillar box anyway?) I still know no more about the color of the horse than I did before. But if I ask "What color does the pillar box seem to you to be?", "The pillar box seems red." is a perfect answer, and it would be impossible to think of one that is more "informative."
ReplyDeleteI'm sure that "2+2=4 and the capitol of France is Paris" contains more information than "2+2=4," but I don't know why anyone would ask about their relative informativeness absent a specific question to be answered.
Of course, not all utterances respond to questions, but all should address a perceived ignorance in A that U wishes to remedy, so all utterances should be tested for informativeness against the state of A's justifiable ignorance before and after the utterance.
Errata:
ReplyDelete1. capital, not capitol
2. so all utterances should be tested for informativeness against the state of A's ignorance before the utterance vs. his presumable lack thereof after it.
Sorry.