The reading assignment is Grice (1989a). This was first published in 1975 in Davidson and Harman (1975). The version I am sharing with you is from a collection published in 1989:
This paper introduces the theory of “Conversational Implicature”. This is a theory about how you can say one thing but mean another, and how others can understand what you mean even when it is different from what you say.
The paper does not begin with this. Instead, it begins with a question about the relation between formal logic and natural language.
| Formal logic | English |
|---|---|
| ~\(p\) | not \(p\) |
| \(p \wedge q\) | \(p\) and \(q\) |
| \(p \vee q\) | \(p\) or \(q\) |
| \(p \supset q\) | if \(p\) then \(q\) |
| \(\forall x\) | for all \(x\) |
| \(\exists x\) | for some \(x\) |
| ℩\(x\)1 | the \(x\) |
Grice tells us that pretty much everyone agrees that the constants of formal logic don’t have quite the same meaning as their natural language counterparts.
In another paper, he gives examples to illustrate this divergence in meaning. Here is one:
…Because it would be incorrect or inappropriate to say “He got into bed and took off his trousers” of a man who first took off his trousers and then got into bed, it is part of the meaning, or part of one meaning, of “and” to convey temporal succession. (Grice 1989b, 8)
The logical connective, \(\wedge\), does not convey temporal succession: ‘He got into bed \(\wedge\) he took off his trousers’ and ‘He took off his trousers \(\wedge\) he got into bed’ mean the same thing: that both component sentences, ‘He got into bed’ and ‘He took of his trousers’, are true.
Here is a second example, involving ‘or’:
The fact that it would be inappropriate to say “My wife is either in Oxford or in London” when | know perfectly well that she is in Oxford has led to the idea that it is part of the meaning of “or” (or of “either…or”) to convey that the speaker is ignorant of the truth-values of the particular disjuncts. (Grice 1989b, 8–9)
Again, the logical connective, \(vee\), does not carry this extra meaning. ‘My wife is in Oxford \(\vee\) my wife is in London’ means that at least one of the two disjuncts is true, and conveys nothing about the speaker’s ignorance.
He gives on more example, involving conditionals:
Strawson maintained that, while “if \(p\) then \(q\)” entails “\(p \rightarrow q\),” the reverse entailment does not hold; and he characterized a primary or standard use of “if…then” as follows: “each hypothetical statement made by this use of ‘if’ is acceptable (true, reasonable) if the antecedent statement, if made or accepted, would in the circumstances be a good ground or reason for accepting the consequent statement; and the making of the hypothetical statement carries the implication either of uncertainty about, or of disbelief in, the fulfillment of both antecedent and consequent.” (Grice 1989b, 8–9; quoting Strawson 1952, III pt. 2)
Those of you who have taken a logic class may remember discussion there of the weird features of the “material” conditional. In the quote above, Strawson is attempting to articulate the meaning of the English conditional, ‘if…then’, in a way that diverges from the simple material conditional of formal logic.
Returning to Grice: pretty much everyone agrees that the constants of formal logic don’t have quite the same meaning as their natural language counterparts. But they disagree about what this shows.
The formalists think that, at least for the purposes of logic and science, we should replace our messy natural language with the purified language of formal logic. Frege was a formalist in this sense.
The informalists resist this: science is not the sole purpose of language; the logic of ordinary language is important in its own right.
Grice rejects the shared assumption:
The common assumption of the contestants that the divergences do in fact exist is (broadly speaking) a common mistake, and that the mistake arises from inadequate attention to the nature and importance of the conditions governing conversation. (Grice 1989a, 24)
His theory of conversational implicature is just that: a theory of the conditions governing conversation.
Slides
Additional Reading
For a compelling application of the theory of conversational implicature, see Saul (2012). She argues that there is no morally significant difference between lying and telling a misleading truth, as both involve essentially the same abuse of conversational trust.
Grice’s theory of conversational implicature might help explain what is going on in the cases raised by Donnellan (1966): a speaker successfully refers to a man using the definite description, ‘The man with martini,’ even though he is not actually drinking a martini. These cases are discussed by Kripke at the beginning of Lecture I. For further development of this, see Kripke (1977).
Originally, I was toying with the idea of also assigning Grice (1957). This is an earlier article, and in it, Grice offers a “theory of meaning” that reduces facts about meaning to facts about speaker intention. Here, in a nutshell, is the account he defends:
\(a\) means \(p\) by uttering \(u\) if and only if \(a\) intends in uttering \(u\) that:
- his audience come to believe \(p\),
- his audience recognize this intention, and
- occur on the basis of (2).
There are interesting connections between this and his theory of conversational implicature, but it would be too much to try to discuss both in one week!
For a fairly accessible overview of the idea, and some objections, see Speaks (2021, sec. 3.1.1).
References
Footnotes
This symbol, an inverted lowercase Greek iota, was introduced by Russell as a notation for definite descriptions: ‘℩\(x\)F(x)’ means ‘The \(x\) such that F(x)$’.↩︎