I do find this aspect of Grice particularly hard to get a grip on, and suspect that in our Carnap/Grice conversation it will prove tricky.
First, the distinction he is working from seems strange.
The Neo-traditionalists are the ones who perceive there to be important differences between formal logic and ordinary discourse, whereas the modernists have a clutch of characteristics which I have never seen exemplified (except perhaps in Grice), which include logical parsimony and the belief that a standard formal vocabulary matches ordinary discourse better than is held by neo-traditionalists.
I am perhaps better acquainted with non-philosophical logicians than with philosophical logicians. Mathematical logicians and logicians in computer science contain few who could be described as traditionalists.
They are the non-philosophical face of modern logic, and I would say that it is almost universally supposed by such people that formal languages differ from natural languages in quite radical ways among which some of the simplest are in the semantics of the propositional connectives.
I would guess that most of them would consider these differences a matter of design. As in the case of Russell, the formal languages are designed to be superior to natural languages, following millenia of tradition in the evolution of mathematical language and notation.
Its not hard to see why Grice might want to take a moderating position on the extent of the differences. He is a moderating influence on ordinary language philosophy, and part of the moderation involves denying that certain features of usage should be taken as reflections of meaning. This does provide a plausible alternate account of why we do not say "she got into bed and took off her nickers" even if this might not be literally false, and it is a worthwhile project to see how far this kind of reconciliation might go in relation to logical connectives.
However, he does seem to be adopting a rather extreme position, and is to that extent exposed. The extent of the exposure depends upon what we take to be his central thesis.
Let us put aside his characterisation of neo-traditional verses modernists on logic, which is at risk of creating a fruitless diversion.
There seem to me to be two distinct theses which he might be advancing.
The first is that the certain features of ordinary language do in fact correspond more faithfully in their semantics to the corresponding features of "standard" logic than has been alleged by their critics.
The second is that though no such simple correspondence exists it is nevertheless possible in some more roundabout way to give a good account of ordinary langage using modern symbolic logic.
Do either of these seem plausible condensations of his project?
Sometimes he seems to talk as if he meant the first, but then he does things which suggest at best the second. System G fits this case, for this is an aknowledgement surely that equality normally used in modern logic is not adequate for personal identity, but an attempt to give an account through a modern system of something closer to the ordinary concept of identity?
Even with his project construed in the more liberal terms there are general difficulties which strike me, and one wonders how they can be overcome.
Of which the relevant variants on the paradox of analysis are perhaps the most sweeping. Surely no precise account can be faithful if the original is vague?
What does Grice have to say about this kind of issue?
RBJ
Monday, July 5, 2010
Subscribe to:
Post Comments (Atom)
Thank you, Roger -- As Roger knows, most of the references to this somewhat mysterious System G-HP -- are indeed a nod to our joint work on the conversation Carnap/Grice so I'm glad he offers his input. On a first reading of his comment, I would think that Grice would endorse the first of the theses Jones exposes, to wit:
ReplyDelete(in Jones's apt wording):
certain features of ordinary language do in fact correspond more faithfully in their semantics to the corresponding features of "standard" logic than has been alleged by their critics.
Since Jones has easy access to Grice's WoW, I would suggest that the issue can be formulated in terms of what Grice calls 'specifications' of meaning -- we can expand on these in further blog posts, since I'll try to be brief here this being a commentary. Only ONE of those meaning specifications -- in my view --, i.e. those involving the meaning of expressions of a timeless kind for a part of an utterance -- would fit the bill.
"She got into bed and got off her nickers" -- is an utterance. By uttering it, an utterer may mean whatever he pleases, basically. Including, "It is raining". This is applied, not timeless, and it's in terms of 'that'-clauses. E.g. Utterer meant, on this occasion, that she FIRST got into bed and THEN got off her nickers".
But Grice seems to be wanting to say -- recall this is Lecture V I am referring, in WoW, which is thus seen as an expansion of Lecture II where he does mention the 'divergence' (alleged, of course) between 'and' and "/\" --.
So, whereas what an Utterer may mean by uttering such an utterance is pretty variable, we can still claim,
"and" means '/\'.
---- Note that the specification of expression meaning, for Grice, involves the use of 'direct quotation' marks. So, to use the list of seven devices he has in Lecture II, which is in fact an expansion of the prolegomenon Lecture I, where he mentions explicitly Strawson on 'if', would go:
"not" means '~'
"and" means '/\'
"or" means '\/'
"if" means ')'
"all" means '(x)'
"some (at least one)" means '(Ex)'.
"the" means '(ix)'
I agree with Jones that one feels Grice's neotraditionalist/modernist debate is a bit contrived, and I wouldn't be surprised, then with a surname like that he possibly deserved it, that Grice is using Strawson as his Strawman!
Thanks for the input, Roger. I will try to comment on a post blog. I think your thesis I fits Grice's bill nicely.
ReplyDeleteThank you, Roger -- As Roger knows, most of the references to this somewhat mysterious System G-HP -- are indeed a nod to our joint work on the conversation Carnap/Grice so I'm glad he offers his input. On a first reading of his comment, I would think that Grice would endorse the first of the theses Jones exposes, to wit:
ReplyDelete(in Jones's apt wording):
certain features of ordinary language do in fact correspond more faithfully in their semantics to the corresponding features of "standard" logic than has been alleged by their critics.
Since Jones has easy access to Grice's WoW, I would suggest that the issue can be formulated in terms of what Grice calls 'specifications' of meaning -- we can expand on these in further blog posts, since I'll try to be brief here this being a commentary. Only ONE of those meaning specifications -- in my view --, i.e. those involving the meaning of expressions of a timeless kind for a part of an utterance -- would fit the bill.
"She got into bed and got off her nickers" -- is an utterance. By uttering it, an utterer may mean whatever he pleases, basically. Including, "It is raining". This is applied, not timeless, and it's in terms of 'that'-clauses. E.g. Utterer meant, on this occasion, that she FIRST got into bed and THEN got off her nickers".
But Grice seems to be wanting to say -- recall this is Lecture V I am referring, in WoW, which is thus seen as an expansion of Lecture II where he does mention the 'divergence' (alleged, of course) between 'and' and "/\" --.
So, whereas what an Utterer may mean by uttering such an utterance is pretty variable, we can still claim,
"and" means '/\'.
---- Note that the specification of expression meaning, for Grice, involves the use of 'direct quotation' marks. So, to use the list of seven devices he has in Lecture II, which is in fact an expansion of the prolegomenon Lecture I, where he mentions explicitly Strawson on 'if', would go:
"not" means '~'
"and" means '/\'
"or" means '\/'
"if" means ')'
"all" means '(x)'
"some (at least one)" means '(Ex)'.
"the" means '(ix)'
I agree with Jones that one feels Grice's neotraditionalist/modernist debate is a bit contrived, and I wouldn't be surprised, then with a surname like that he possibly deserved it, that Grice is using Strawson as his Strawman!
I think you are suggesting that the conjunction ("and" &, inverted carrot, etc) functions as a sort of "time-keeper" at times in natural language, or shall we say seems to be a...narrative device--not strictly conjunctive. A sequencing of sorts, like "she hopped into the sack, AND ripped off her knickers", OR "she ripped off her knickers, and bent over the bed." :). She did one act, then another. Time--temporality--, or describing events a real problem for logicians of course. But the ...implicature, if you will, would probably be, " She did this Act A, AND then she did that Act B" (or other things too).
ReplyDeleteYes-- The rule, brief as it is, is 'be orderly'. Of course it's not a rule. Horn has identified a Greek grammarian, who he thinks is the first Gricean, since he had noted this 'regulation' -- to report events as they succeed. The Gricean explanation (or locus classicus) took Horn and I some time to locate. I finally found it in Grice's "Presupposition and Conversational Implicature" (where Grice uses Urmson's 1956 example, "He went to bed and took off his trousers"). Since this is a precis of his theory in an independent lecture, when Grice compiled the thing for publication in 1988 (he died in 1989, and the thing came out in 1989) he omitted the paragraph. But Horn quotes it in full in his "Implicature" entry for his own edited "Handbook of Pragmatics" (Blackwell) with a credit to J. L. Speranza. So, yes, it's the 'be orderly' thing. Urmson is, himself, relying on Strawson who has the example, "She got pregnant and married" (or something) and recall this is 1952, and already crediting H. P. Grice twice (in that book, "Introduction to logical theory"). So the pragmatic 'explanation' was 'in the air', and for good!
ReplyDeleteThe "then" is then a conversational implicature attached to 'and' in "She went to bed and stripped" (on occasion). There are further implicatures of the 'post hoc, propter hoc' type in Ryle-type examples, "He died, and drank the poison" -- Example by Levinson, Pragmatics: "The lone ranger ran away and jumped on his horse".
ReplyDeleteok--in other words, AND's not just a conjunction or addition, but like sequencing, and causality, not to say intention, human acts, etc. But I don't think that means "&" can't be used, but about the...propositions, language--not the connectives. First order logic works fine when you strip it of like all hints of intention and psychology--, or probability. Logic's like circuitry, not the vague suggestions and psychologism of prosecutor-speak (or...most "conversationalists"). Im not against psychology, but it's a different kettle of fish...or maybe IM missing something (the issue seems a bit related to Carnap's state descriptions...)
ReplyDeleteIt's a bit like your idiosyncrasy -- I don't like the idea of an idiolect because most of my idiosyncrasies are SIGNALLINGS which are not 'lects'. So, I would think that I do share with Grice this idea that 'and' should be used without any relic or unwanted implicature to the effect that 'then' is suggested:
ReplyDelete"I went to Paris and to London and to Munich and to Geneva and to Venice and to London and to Paris and to Manchester and to Rome".
I wouldn't think there is any order in that!
So, the point is we DO need something like the inverted wedge ('p /\ q') -- I use this sign because it's the one used by Grice in WoW, Logic and Conversation, Lecture III. Oddly, the first use of this sign is NOT to concatenate or add, as J may be suggesting. For surely I don't need "and" to do that. I can say:
ReplyDelete"I went to Paris. I went to London".
I mean, I do not NEED to say, "and"".
But Grice suggests that we do need to say 'and', to be courteous to any possible co-conversationalist. Suppose he says:
"That's not true!"
"What isn't?"
"That you went to London and to Paris".
At this point, a concatenation is pretty vague. If I find Grice's paragraph I'll append.
This comment has been removed by the author.
ReplyDeleteWell, the implicature might be...time itself. Say Jr was suspected of, oh, running heroin for the Mob, and when he returns to LAX, the Feds whisk him off for interrogation. And they break out a calendar. So bad cop says, "JUNIOR, we want you to tell us exactly what you did, at all hours of the day, ..and who you did it to...". Then, Jr. says "I flew to London. The next day, June 14, I flew to paris," etc. So normal conversation doesn't usually require the temporal specificity, but it can be added--Grice and ord-lang just seem to be saying the conjunction doesn't describe time, which is accurate. So, context. But certainly events could be evidence of a sort and added up, even if not strictly quantitative...Jr.'s affadavit. or, I guess one concatenates it, unless a program says not to....
ReplyDeleteWedge works for FO-L, but so does Ampersand (Boolos's choice).
Yes -- 'and' should not involve ANY reference to time. I think it's the commutativeness of 'and' that counts:
ReplyDeletep & q iff q & p
The conditions that make 'p & q' true have nothing to do with the order. There are zillions of uses of 'and' and the use of 'and' to report events must be a small percentage of that use. Grice is not being original here. He is just defending the use of '&' to represent 'and' even against the commentary by Strawson in "'And' and '.'" in his early (1954) Introduction to Logical Theory.