Grice's "Presupposition and conversational implicature"

 In “presupposition and conversational implicature”, Grice is interested in the question of whether what Strawson called “presupposition” was something that could be accounted for in terms of conversational implicature or some similar  (what we would now call) pragmatic mechanism. He considers several examples, but cares mostly about the existential presupposition of (utterances of sentences containing) definite descriptions, factives and what Robert Stainton and myself have called “Quasi-factives”, i.e., sentences like:

1. Peter does not regret inviting Joanna to the party.

which, when asserted, defeasibly but strongly imply the truth of their complement (in this case, that Peter invited Joanna to the party).

According to Grice, the existential presupposition of definite descriptions can be easily accounted for by appealing to a norm of manner according to which it is bad manners to hide the main point of your assertion (my phrasing, not his). This is what he calls the principle of conversational ‘tailoring’. 

As far as I can understand the idea, we can divide the semantic entailments of a sentence in two kinds: those that it explicitly expresses and those that it abbreviates. For example, “There is cat here and it does not like visitors” and “This cat does not like visitors” both imply that there is a cat here, but the first one expresses it explicitly, while the second one contains it in abbreviated form.

Since it is bad manners to hide the main point of your assertion, whenever we prefer the abbreviated form over the explicit, it must be because we are signaling that there is something there we think is part of the conversation’s common ground; we do not want to be distracted by it.

Now, Grice follows Russell in taking 

(2) The present king of France is bald.

to abbreviate that there is currently a king of France and that it is unique and that he is bald. Then, he wonders what the significance of us using (2) instead of the expansion (2x) “there is currently a king of France that it is unique and bald”. He appeals to the aforementioned principle of manner to argue that we use (2) instead of (2x) when the three conjuncts are not a la par on the conversation, i.e., some are  part of the conversation’s common ground and some are the point of the assertion. From the semantic point of view, they are all the same, so a pragmatic process is required in order to determine which one is put forth as the point of the conversation and which one is assumed to be part of the common ground. 

Thus, its negation

(3) The present king of France is not bald.

implies the falsity of at least one of the above implicated propositions. However, these propositions are not logically independent. If there is currently no kind of France, then, there is no unique king of France and also there is no current, unique, bald king of France. Thus, (3) is true either because there is currently no king of France, or because it is not unique or because it is not bald, where each one of them is logically more precise than the others. Thus, he concludes that the point of (3) cannot be but the strongest one, and that the utterer of (3) uses the abbreviated form to signal the speaker that she takes the other two possibilities to be part of the conversation’s common ground (p. 276).

He then remarks  that this cannot be easily extended to a case like (1) – I assume because there is no similar logical hierarchy among the possible grounds for it truth – and thus that the presupposition there most likely will be conventional. He writes:

“I do not see that it is going to be particularly easy to represent the implication in the case of regret as being one of a conversational kind.” (p. 280) According to Grice (p. 280), (1) can be true either because Peter does not think he invited Joanna to the Party or because he does not have a negative attitude towards such an invitation. But here the truth of the compliment, i.e., that Peter invited Joanna, is not necessary for these possible grounds to be the case, thus the aforementioned pragmatic mechanism does not apply.