"You're the cream in my coffee!"-- Grice on disimplicature

Speranza

I am using this example, because in unpublished (but never unwritten) notes, Grice has this marginal comment,

"You're the cream in my coffee!"

-- the 'disimplicature' being TOTAL!

--- I may have written something about this elsewhere in the Grice Club. Now for some running commentary on Jones, as I focus then on this notion of

"disimplicature"

(which may be behind much talk on presupposition, presupposition cancellation, loose, undeterminacy, and so on...)

Jones writes:

"I thought I might say a little about why I was curious about Grice's views on presupposition. This is in the context of our "conversation between Carnap and Grice", which has been dormant for a while. It seemed to me, though I don't believe this comes from Carnap, that Carnap's position on ontology as given in his "Empiricism Semantics and Ontology" is a bit like saying that ontological questions external to some language framework (which Carnap considers meaningless) might well be considered common presuppositions to any assertion in the language."

I would think, but we would need to trace this, that this is Collingwood's idea of 'presupposition'. Collingwood is particularly interest in that, like Grice and Strawson, he is "Oxford" (or Oxonian). It MAY relate, since Collingwood was a bit of a continental, with Husserl:

"Philosophy without presuppositions."

I never understood that term! But you are right, also that it connects with Grice's appeal to English -- an English philosopher works with the categories of the English language, and so on.

"Presupposition to any assertion" is a good one. Horn once wrote on this, on what he called 'assertive inertia," I think. I pointed out to him that, for Grice, this extends BEYOND assertion: questions ("Who killed Cock Robin?") and orders ("Do it!") also _presuppose_. While there is this idea that

presupposing

conflicts with _stating_ or asserting (what Grice, as Neale notes, calls central speech acts), as opposed to 'implying', it also applies to central speech acts which are NOT assertion. There is a recent book by Palgrave Publishers on "Assertion" which might consider at this pont.

Jones continues:

"The question is then whether Grice would have any objection to that way of talking about a non-natural language. My impression is however, that Grice's objection to presuppositions is exclusively related to his objecting to truth value gaps in natural languages."

Indeed. For we can see that System G, as we call it, or System G_HP, strictly: a highly powerful version of System G, is about:

syntax
semantics
pragmatics

Most of the alleged divergences between 'and' and '&', say, can be dealt with syntactically. In SOME cases, we need an appeal to the 'semantic' truth-tables. The whole enterprise starts to loose sense if we allow that in truth-value assignment we are allowed NOT to assign a truth-value.

By distinguishing syntax-semantics here I am being a traditionalist. By syntax, Grice and I refer to Gentzen-type of 'natural' deduction, introduction and elimination of symbols. Consider the introduction of 'and', or the elimination of 'and', and so on.

The 'semantics' for "and" is a different animal. Here we see that "p & q" receives the same truth-table that "q & p" does. The commutability of "and" then is 'semantic' rather than 'syntactic'. In the case of 'negation', perhaps it's both. And so on. For each operator, we need to consider this. In the case of the definite description, we need to focus on the introduction/elimination (syntax) of the iota operator, and the truth-table for the iota operator. Grice considers this in terms of what he calls the

Russellian expansion

in WoW. I.e.

"The king of France is not bald"

thus gets expanded in THREE conjuncts. A & B & C. In none of them he considers the 'iota operator', but we have to trace the 'iota' operator to its very definition. E.g. its reliance on Leibniz's Law of the equivalence of indiscernibles, and so on. For 'iota' is a DERIVED symbol. But it DOES depend on the symbol for "=", identity. There is a closure clause for defining iota-x, such that there is no other "y" such that... "=" itself is a derived sign. So we need to consider all the primitive syntax behind "the". Then, we identify the minimal syntactic rules, and THEN we deal with the semantics.

The pragmatics for System G would allow for things like:

"Jones's butler was clumsy."

He meant, "Jones's gardener," since U wrongly thought that the man with protruding ears was Jones's butler (he was dressed as one, and Jones indeed had him dress as a butler for that special party). Grice wants to say that 'clumsy' attaches truthfully to that individual that the utterer describes as "Jones' butler". The utterance in itself is _false_. And so on. In any case, there is a pragmatic explanation for any divergence as regards the truth values of things like descriptions. Neale does not consider this, but I would think the expansion,

"whoever he is"

works very well. In some cases, when we use a description, such as "the president of Ruritania", "whoever he is", "must have it easy". In other cases, it is totally uncalled for. You can NOT add, "whoever he is". "The president of the USA, whoever he is, is meeting the president of Italy, whoever he is." This sounds as _too loose_. Grice sometimes, as Neale notes, introduces some devices that LOOK like formal, but they ain't. In some cases, the typography fails. Grice uses square-brackets, the WoW reprint does not. The joke is lost. In other cases, Grice uses SMALL CAPITALS: "JONES'S BUTLER", as oposed to "Jones's butler". In one case, it is IDENTIFICATORY; in the other it is not -- Grice unimaginatively, but he would rather be seen dead than using Donnellan's vocabulary -- says 'non-identificatory' (Donnellan: attributive/referential).

And so on...

Jones continues:

"In the case I had in mind, the presupposition does not result in truth value gaps.
Not at least in the language itself, though Carnap's view on these external questions, that they are meaningless, does result in no truth value for the external question."

Good. We would need to discuss this in

Husserl: title of his book: philosophy without presuppositions. What are the presuppositions of philosophy? Suppose materialism versus empiricism. The existence of sense-data, or the mind. Is that a 'presupposition'? What is Husserl aiming at? He thinks that Phenomenology discusses things, philosophically, and does not need any presupposition as to what's there.

Then there's Collingwood. He is considering things like the idea of history, etc. And he notes that there are 'presuppositions'. Grice et al, in "Metaphysics" in Pears, The nature of metaphysics, discusses this particular use of 'presupposition' as it was influential in Oxford pre-analytic philosophy of metaphysics. Collingwood was the prof. of metaphysics, at the time Grice was writing that.

And then there's Carnap, and Strawson misusing a word with recognised pedigree, as 'presupposition' was, to simple things like,

"By uttering 'the king of France is not bald,' the utterer IMPLIES that there is a king of France, which happens to have hair." And so on.

Typically Grice would not stand such loose use!

Jones continues:

"Carnap himself I think would consider the question of truth value gaps in natural languages (and indeed most questions about natural languages) as
belonging to some science (linguistics?) because natural languages are contingent, rather than to philosophy."

Grice would of course aim at deeper, more committed view! Why leave it to linguistics, or some empirical science, what philosophers DO with language or lingo? He wants to say that truth-assignment (central meaning, in Strand 5) is so basic, that without it, there's no room for the analytic-synthetic distinction. It all depends on the alethic view on things. This is NOT something that can be refuted, say, by a Chinese speaker who says, "Hey, in my dialect, "The moon is made of cheese" displays a truth-value gap."

--- And so on!

Jones continues:

"However, he softened on the question of the limits of philosophy, and would problably not want to press that kind of objection to ordinary language philosophy later in his life."

To think that the Carnap papers somewhere deposited in a uni deploy such varied views! Grice, too, later in his life, changed his views. And to think that what he left in the Grice papers could fill VOLUMES!

Jones:

"I just had a look at the Strawson paper in the Carnap Schillp volume, which is about the relative merits of constructed and natural languages in analytic philosophy. Strawson seems rather dismissive of the case for formality. He misses however a central issue, which is the difference between his (Strawson's) conception of philosophical analysis and Carnap's. He is assuming that in advocating formal languages for philosophy Carnap is advocating their use for the kind of philosophy which Strawson is doing, i.e. the analysis of ordinary language. But Carnap doesn't consider that to be philosophy,"

Lovely. Grice, at least, KNEW that! He relished (if that's the word) in that anecdote. When Gustav Bergmann (whoever he was) was invited to attend a Saturday morning with Austin, Bergmann allegedly replied:

"I rather do things other than spending a Saturday morning with a group of English futilitarians."

---- Grice was being reactionary, and by sticking to the implicatures of idioms he KNEW he was not doing philosophy (or Philosophy) but he loved not doing it!

----

Jones:

"and so its very unlikely that he intended to advocate formal languages specifically for that purpose. Grice's neo-traditional line of argument sometimes sounds more like an exploration than a statement of conviction on Grice's part.
He wants to contradict those who argue that the logic of ordinary discourse must diverge from "classical" logic in various respects, but it's not clear from my own reading of Grice that he is not playing "devil's advocate", perhaps having a less dogmatic view on this than appears."

Too true. I LOVE your reading of him!

"In "Vacuous names" which is part of that enterprise, Grice looks very exploratory and experimental. It has a flavour of "lets see how far we can take this" (can we have our cake and eat it?) and devotes very little energy to the question whether ordinary language really is like "system Q"."

Too true. He is especially exploratory in that he starts the thing by noting that Quine disallows "names" in his system. So, seeing that there are no NAMES, who (i.e. Quine) cares if some of them (i.e. names) are vacuous? It is very sad that Quine, in his "Reply to Grice", notes that he rather not even GO there. He finds all the subscript notation totally otiose, when there is a re-write to hand. He notes, too, that for "Fa" there's Ax & Fx -- i.e. a name like "Marmaduke Bloggs" (a) becomes the predicate, "being-Marmaduke Bloggs", and so on. He should at least pay some attention to the later segment on descriptions as such. But Quine was very confused about propositional attitudes. He thought that 'believing that the sun is red' is an unanalysed predicate (notably in "Words and objects"). The logical form of

"Peter thinks Paul is good"

is Fa -- where 'a' is Peter, and "F" is "-- thinks that Paul is good"; to be further analysed by denying status to "a" and replacing it for the further predicate 'being Peter"

"There is something that holds the property of being-Peter and such that it also holds the property of thinking that Paul is good".

We know that Quine never understood philosophy as Carnap did. We know that Quine made BAD use of some ideas he just took from Carnap, and we KNOW that Quine used Carnap to bring in his own programme against anaytic philosophy, by proposing a scepticism towards what makes the venture of analytic philosophy worth being pursued, but that's another topic!

Jones continues:

"One trick [Grice] misses in this is the possibility of retaining a two-valued logic by "looseness"."

I love that. I point that using 'disimplicature' helps as well.

"U implicates" -- Grice spends some time elucidating. Let us revise what he says about "disimplicate", in the Grice Club, using the search engine, "disimplicature", some other time?

----

Grice's examples:

Hamlet saw his father on the ramparts at Elinore.

Hamlet's father is dead.

In a context where utterer assumes his addressee assumes that the utterer assumes that Hamlet's father is dead, the utterer "is not committed to the usual entailment", as we may say, that Hamlet's father was on the ramparts.

Bill intends to climb Everest next week.

When context (looseness?) makes it obvious tahat there are forces that prevent us from fulfilling an intention, as we may say, one may utter

"Bill intends to climb Everest next week."

WITHOUT committing to the usual entailment. The utterer DISIMPLICATES that Bill is SURE, for example, or KNOWS, that he will climb Everest, "just because everyone knows of the possibly prohibitive difficulties involved," as we might say.

Implicature is a matter of meaning more than we say -- to entailed meaning, as it were.

With disimplicature, Grice is allowing that the MEANING conveyed by an utterer (who utters 'x') in a context may in fact be LESS than is entailed (as we might say) by 'x'. The entailment, to use Grice's word, is "dropped" in context.

Grice notes that 'disimplicature' is "total", as in "You are the cream in my coffee"" (Grice Papers). The remark appears in parentheses. In a context which makes the WHOLE of 'what is said', as it were, untenable, it can be replaced with an implicated (metaphorical) meaning ("You're my pride and joy"). The mechanism of disimplicature does not diverge from that of imlpicature. Disimplicature is explained in terms of the assumption that the conversationalist is abiding by the conversational 'maxims' (and principle), particularly those enforcing 'qualitas' (truthfulness) -- submaxim 1 in particular. If one of all the entailments of 'what is said' are PLAINLY *false*, they can be assumed NOT to arise on that particular occasion of use. S. Yablo noted, with scepticism: Implicatures happen. But then they often don't. Disimplicatures happen, then.

And so on.

Jones goes on:

"One obtains a loose semantics by using principles to constrain the interpretation of one or more constructs without making the constructs completely definite."

This, charming, looks like a semantic (or systemic) regimentation for what may be eventually, a process of a pragmatic nature (as disimplicature is). But it is welcomed in allowing to observe what type of looseness is involved, and so on.

"This is one way of dealing with partial functions in classical first order set theory. Under this scheme a definite description denotes some value even if the description is not satisfied, but in that case we have no way of knowing which function it denotes."

Very good. It may relate to 'dossier'. Neale does not mention this, but Evans does in his "Ways of Reference". Evans takes up the notion of a 'dossier' that Grice introduces in "Vacuous Names". When conversationalist A meets conversationalist B they have to work on the overlap of a dossier. An example from Urmson, "Intentions and Intentions", Aristotelian Society, may apply:

"I saw Mary's husband today."

----- It turns out that he saw the postman. "Why didn't you say, "the postman", rather than "Mary's husband"? -- I.e. A and B, as they converse, have dossiers:

Philip: Mary's husband, the postman.

--- When we hold a conversation, we apply the definite description that we think fits best "in the context of utterance", as it were. This may relate to what Strawson, before Grice, for once, called,

Principle of the Presumption of Ignorance
Principle of the Presumption of Knowledge
Principle of Relevance ("Identifying reference and truth-values").

Without a bit of each, conversation could not even start. If we all share the same knowledge, what's the point of INFORMING, say? And so on. These pragmatic dimensions may implinge on the semantic looseness that Jones is referring to. Or not!

Jones:

"This works fine in a two valued logic and is the usual way of treating Hilbert's choice function (which corresponds to indefinite description)."

Lovely. Grice was OBSESSED with 'indefinite descriptions'. Apparenly, he never saw farther than:

"He is meeting A WOMAN this evening." Implicature: not his wife.

----

"I found a tortoise in a house." (Not my tortoise in my house).

"She broke a finger" -- the nurse. Not hers?

And so on. This is in WoW: III. He is discussing the logical form of indefinite descriptions, and is like precisely refuting a view like Cohen's:

Cohen would like to say that 'a' (as in "an x") has DIFFERENT _senses_ according to context. Grice sticks to just the logical form of the canonical system, with glosses for any further implicature (or disimplicature) that may arise in context.

I never use "a" or "an" in indefinite descriptions. To me, "a" means "one" ("He is meeting ONE woman this evening."). I use 'some' -- I follow Warnock there, "Metaphysics in Logic" -- "Some kings of France are not bald," say.

--- Note that in the paraphrasis for "(Ex)" in WoW:II Grice gives "some (at least one)". The use of 'one' to replace (Ex) seems colloquial.

Jones concludes his interesting post:

"Though this way of dealing with the problem is quite old, the possibility of "looseness" of this kind is not often alluded to by philosophers, but might possibly be the best way to interpret natural languages as two valued logics."

Indeed. Keyword then: LOOSENESS. Cfr. VAGUENESS, FUZZY, DISIMPLICATURE, and keep counting!

Cheers.