1 Introduction
All semantic and pragmatic approaches to meaning share the idea that meanings are diverse, and that some of them are structured. First, one of the basic distinctions is between literal and non-literal meaning, of crucial relevance to the analysis of speech acts (Searle 1969, 1979). In speech act analysis, an utterance like (13) literally means (14), but communicates (15):
(14) Are you able to pass the salt?
(15) I request that you pass the salt.
In speech act theory, there is a clear and explainable relation between literal and non-literal meaning. The main assumption of indirect speech act theory (Searle 1979) is that one way of realizing an indirect speech act is to ask one of the addresseeās condition defining the meaning of an act of requesting. This is the preparatory condition 1 (Searle 1969: 66): āH[earer] is able to do A[action], S[peaker] believes that H is able to do Aā. So, asking whether the hearer is able to pass the salt is equal to a request for the salt.1
Second, in Gricean pragmatics, it is common in most situations to associate what is implicated with what is said. For instance, (16) is a classic example of conversational implicature, where B implicates in her answer to A that C āis the sort of person likely to yield to the temptation provided by his occupationā (Grice 1989: 24), put succinctly in (17):
(16) A: How is C getting on his job?
B: Oh, quite well, I think; he likes his colleagues, and he hasnāt been in prison yet.
Third, when a speaker asserts a sentence, such as I stopped smoking in (18), she does not just assert something she believes to be true, but also presupposes a true proposition, which is not communicated but is necessary for the assertion to receive a truth value.
(18) | Doctor: | Are you a smoker? | |
| Patient: | I stopped smoking seven years ago. | |
(19) The patient used to smoke.
The assertion I stopped smoking cannot be true if it is false that the patient smoked in the past. So, the patient in this situation should have answered No, or I never smoked; but in the case of (18), her answer presupposes that the doctor thinks, or has reason to think, that the patient used to smoke. Similarly, if the doctor had asked (20) instead of (18), the patient ā if she never smoked ā is placed in a less-than-optimal situation where she is required to cancel the relevant presupposed proposition that she used to smoke:
(20) | Doctor: | When did you stop smoking? | |
| Patient: | I did not stop smoking, because I have never smoked. | |
Last but not least, some implicit content is even more covert than presuppositions, which cannot be false if the assertion is true ā that is, which are not defeasible without contradiction. Buying a Chow certainly entails ābuying a dogā, and if a proposition is true of a Chow, it is true of a dog, as (21) shows:
(21) # I bought a Chow, but not a dog.
These facts are uncontroversial, but they seem to belong to different categories of issues: the direct vs. indirect speech act is a pragmatic issue; the difference between what is said and what is implicated shows the difference between semantic and pragmatic meanings; the assertion/presupposition distinction is a semantic and pragmatic issue (the target is a presupposition, but the trigger is an utterance); and finally, entailment is a semantic relation between what is said and what is entailed.
So why is the distinction between what is literal and non-literal not of the same type? The usual answer is that an indirect speech act has a pragmatic meaning as well as an implicature, whereas presuppositions and entailments are defined as semantic. So, the literal/non-literal distinction is not homogeneous, whatever the property of the trigger and the target, as Table 2 shows:
Table 2:Different types of meaning.
| target |
| trigger | semantic | pragmatic |
| semantic | entailment | presupposition |
| pragmatic | implicature | indirect speech act |
This initial analysis is unfortunately incomplete: first, the list of types of meaning is not exhaustive; second, at least one type of meaning (presupposition) is not always considered as semantic, but as pragmatic (Stalnaker 1977); third, nor is there any consensus on the nature of one type of implicature ā that is, conventional implicature, defined in Potts (2005) as semantic meaning. The primary goal of this chapter is to fix the definition of these meaning relations. To do so, I will propose a method of disentangling meaning relations from one another. The first step is to propose certain tests; the second provides a truth-conditional analysis, which is to say a contrastive analysis of entailment, presupp...
