H. P. Grice, answered from the texts and cited to the page.
Speaker meaning — or what I have called meaning_NN, nonnatural meaning — is, at its core, the phenomenon of one creature producing an effect in another not merely by causing that effect but by intending the other to recognise the intention behind the production. That recognition is not incidental; it is the mechanism. The analysis I have worked toward across successive redefinitions runs roughly as follows.
For an utterer U to mean something by an utterance x, U must intend an audience A to think that x possesses some feature f, intend A to think that U has that first intention, intend A to think of f as correlated in some way c with a type of response r, and intend A to think that U has that third intention as well.1 The modes of correlation — iconic, associative, conventional — are left open; what matters is that U intends A to arrive at the response through recognising U's intention to produce it, not independently of that recognition.
This is what separates genuine communication from mere causation of belief. Consider the tailcoat case: many people, on seeing someone put on a tailcoat, would conclude that person is about to go to a dance, and the tendency to produce that belief is in some sense regular and learnable — but we do not say the tailcoat means_NN that the wearer is going to a dance.2
The Stevensonian causal account, which tries to capture meaning in terms of a tendency to produce attitudes dependent on conditioning, fails precisely here: if the qualifying phrase "attending the use of the sign in communication" is taken seriously, the account is circular; if it is not, it admits the tailcoat case as a case of meaning, which is absurd.3
The little-girl example I have used to test the analysis is instructive. I address a French sentence to a child who believes it means "Help yourself to a piece of cake," when in fact it means something quite different. She helps herself, as I intended. I intended her to think, and to think I intended her to think, that the sentence meant what she took it to mean.
The fact that the sentence conventionally means something else is no obstacle to my having meant something by my utterance — namely, that she was to have some cake.4 What matters is the intention-structure, not the conventional meaning of the vehicle. This is why the reductive direction runs from the semantic to the psychological, not the other way. Nonnatural meaning, as I distinguish it from natural meaning, is nonfactive and tolerates quotation marks around the specification of what is meant; natural meaning is factive and does not.5
The underlying idea behind both is that of consequence — in natural meaning, consequences are states of affairs; in nonnatural meaning, consequences are conceptions or complexes involving conceptions — and this asymmetry is what makes nonnatural meaning the more specialised and the more in need of further analysis.6 Speaker meaning is the basic notion; sentence meaning and word meaning are derivative, and the programme of Utterer's Meaning, Sentence-Meaning, and Word-Meaning is precisely to show how the latter two are connected to the former.7
U uttered x intending (1) A to think x possesses f (2) A to think U intends (1) (3) A to think of f as correlated in way c with the type to which r belongs (4) A to think U intends (3)Studies in the Way of Words, pp. 110–111
Does this satisfy us that putting on a tailcoat means_NN that one is about to go to a dance (or indeed means anything at all)? Obviously not.Studies in the Way of Words, pp. 223–224
if we have to take seriously the second part of the qualifying phrase ('attending the use of the sign in communication'), then the account of meaning is obviously circular.Studies in the Way of Words, pp. 223–224
I would say that the fact that the sentence meant and was known by me to mean something quite different is no obstacle to my having meant something by my utterance (namely, that she was to have some cake).Studies in the Way of Words, pp. 110–111
we may, for example, inquire whether a particular occurrence of the verb 'mean' is factive or nonfactive, that is to say whether for it to be true that so and so means that p it does or does not have to be the case that it is true that pStudies in the Way of Words, pp. 357–358
The idea behind both uses of 'mean' is that of consequence; if x means y then y, or something which includes y or the idea of y, is a consequence of x. In 'natural' meaning, consequences are states of affairs; in 'nonnatural' meaning, consequences are conceptions or complexes which involve conceptions.Studies in the Way of Words, pp. 357–358
My aim in this essay is to throw light on the connection between (a) a notion of meaning which I want to regard as basic, namely the notion which is involved in saying of someone that by (when) doing such-and-such he meant that so-and-so (in what I have called a nonnatural sense of the word 'meant'), and (b) the notions of meaning involved in saying (i) that a given sentence means 'so-and-so' (ii) that a given word or phrase means 'so-and-so.'Studies in the Way of Words, pp. 124–125