2.3 Grice on natural and non-natural meaning

Ironically, the word ‘meaning’ has many different meanings. There are four occurrences of ‘mean’ (or ‘meaning’ or ‘meant’, etc.), italicised, in the following paragraph:

Here is the paragraph again, with each of the four occurrences of ‘mean’, ‘meant’ and ‘meaning’ replaced with an appropriate synonym.

There is no good reason to demand of a theory of meaning that it give an account of every kind of meaning. It may turn out that there is some underlying unity to the uses of ‘mean’ (or ‘means’, ‘meant’, ‘meaning’) on display here, but that is not something on which we should insist.

In view of the plethora of meanings of 'meaning’, Grice proposes to set aside those that are not his immediate concern and to focus on understanding the nature of those that are. One kind of meaning that is left over is defined by Grice in terms of the speaker's intentions. This is a good candidate for being the kind of meaning we are interested in, i.e. the meaning utterances have that accounts for the role they play in communication. But before coming to what he says about this kind of meaning, we need to see which kinds of meaning he sets to one side as confusing distractions.

Grice begins his paper (Reading 1) by making an important distinction between two species of meaning that it is particularly easy to confuse, which he labels natural meaning and non-natural meaning. The kind of meaning he later defines in terms of speakers’ intentions is non-natural meaning. Natural meaning is the kind being attributed in claim (a):

(a) ‘Those spots on your face mean you have measles.

This claim could be true only if the italicised sub-sentence is true, i.e. only if you really do have measles. If you had spots but you didn't have measles, the spots would not mean that you had measles. They would have to have some other source. Contrast this with claim (b), which uses ‘mean’ in its non-natural sense:

(b) ’The spots in the arrangement below mean you have measles.’

This whole assertion would be true even if the italicised sub-sentence was false. That is, an assertion of (b) would still be correct even if you did not have measles. Generalising, the difference between the two kinds of meaning is this: it is consistent with something's having non-natural meaning that what it non-naturally means is false; but it is not consistent with something's having natural meaning that what it naturally means is false. (In the paper, Grice notes other differences, but this is the main one.)

Grice's purpose in making this distinction is merely to avoid confusion. He sets natural meaning (or ‘meaning_n’ as he calls it) to one side and moves on to developing a theory of non-natural meaning (‘meaning_nn’), the kind he is more interested in. His partiality has to do with the fact that examples of meaning that involve language are typically cases of meaning_nn, and no one has so far come up with a good theory of meaning_nn. Natural meaning, by contrast, does not really have much to do with the meaning of words or utterances, and is in any case relatively non-mysterious. ‘X means_n that p’ can be understood as a substitute for one or other of various simple phrases, including:

• X causes it to be the case that p

• X is conclusive evidence for p

• X is not possible unless p is true

• X entails that p

Natural meaning is mentioned later in his paper, but only in order to clear up potential confusions, not because Grice is especially interested in it.

The distinction between natural and non-natural meaning is, Grice notes, not always clear cut. The same entity can sometimes have both natural and non-natural meaning. Here is an illustration:

This is plausible on both readings of ‘meant’. But in most cases the distinction seems to be reasonably easy to apply. Let us move on, then, to Grice's theory of non-natural meaning. (Henceforth in this section, ‘meaning’ should be read as ‘non-natural meaning’ unless specified otherwise.)