Grice’s most enduring contribution to the philosophy of language is his distinction between ‘what is said’ and ‘what is implied’. Those propositions which the speaker means to communicate but which are not part of what she explicitly says, are what Grice calls ‘conversational implicatures’. However, what interests us here, for the moment, is Grice’s ‘what is said’, and how it is related to the semantic content of the utterance. ‘What is said’ can be taken in one of two ways: either it refers to “what is asserted by the speaker”, or to the (meaning of the) words uttered (Allott 2010: 197). Given the nature of linguistic underdeterminacy, both cannot be true at the same time – either ‘what is said’ includes pragmatic intrusions, or it is strictly semantic but subpropositional (see Blakemore 2002: 47). Grice recognized that ‘what is said’ had to somehow go beyond the words uttered (what is known as ‘sentence meaning’25) – i.e.,
that the words uttered, on their own, could not make up a full, truth-evaluable proposition. As a result, he argued that hearers have to go beyond sentence meaning in order to get at ‘what is said’, that is, what the speaker meant to communicate directly. The hearer has to be able to assign reference to referring expressions, disambiguate ambiguous constituents, and determine the time and location of the utterance before an utterance can be adjudged to be true or false (Clark 2013: 77; Neale 1992: 520).26
According to Grice, then, ‘what is said’ has to be pragmatically established, but is still very close to the conventional meaning of the words uttered, i.e., to sentence meaning (Clark 2013: 167).
Grice’s conversational implicatures, on the other hand, are almost fully pragmatic in nature. Although they do follow from “the conventional meaning of the words used” (i.e., ‘what is said’), they are also dependent on “context” and rely on the hearer’s knowledge and application of what Grice calls the ‘Cooperative Principle’ (Grice 1989: 31):
25 See Recanati (2004: 6) for the notion of ‘sentence meaning’.
26 Technically, an utterance cannot be true or false (only propositions can) – what I should say here is ‘before
“Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.” (Grice 1989: 26)
Grice assumes that conversation is a rational exchange, where both parties observe the Communicative Principle. He argues that communicators also assume that hearers will assume that the speaker will observe the Cooperative Principle. Based on the Cooperative Principle, Grice proposes four maxims – norms which we are expected to follow in conversation. Briefly, these maxims assume that speakers will be “informative but not too informative” (Maxim of Quantity); truthful and “based on adequate evidence” (Maxim of Quality); relevant (Maxim of Relation); and “to speak in an appropriate manner” (Maxim of Manner) (Clark 2013: 57). The Maxim of Manner leads the speaker to “avoid obscurity of expression”, “ambiguity”, and leads her to “avoid unnecessary prolixity” and “be orderly” (Grice 1989: 27).
The combination of ‘what is said’, context, and the Cooperative Principle and the accompanying maxims, leads to the generation of implicatures (Neale 1992: 525; Clark 2013: 59). As such, implicatures can never be semantically encoded in an utterance (unless they are conventional, cf. infra) (Blakemore 2002: 74). According to Grice (1989: 30), implicatures arise in one of four ways – either the maxims are being followed; there is a “clash” between more than one maxim; the speaker ‘flouts’ a maxim; or the speaker ‘opts out’ of the Cooperative Principle. The last one will not concern us here – the other three are described in detail by Grice (1989: 32-33).
(24) A: I am out of petrol.
B: There is a garage round the corner. (from Grice 1989: 32)
In this case, B would not be acting in accordance with the Cooperative Principle if he did not mean to imply that the garage around the corner will have petrol available. In other words, the only way in which A will take B to be acting in accordance with the Cooperative Principle, is if he generates the implicature that the garage will have petrol available. If this implicature is not generated, the maxims would be violated (Clark 2013: 59).
(25) contains an instance of ‘clashing’ maxims:
(25) [A is planning with B an itinerary for a holiday in France. Both know that A wants to see his friend C:]
A: Where does C live?
B’s answer would seem to violate the Maxim of Quantity – A obviously wants to know more than the general vicinity where C lives if he is to plan his trip right. However, B “is aware that to be more informative would be to say something that infringed the […] maxim of Quality”, as he does not have “adequate evidence” for a more detailed description (Grice 1989: 33). The implicature here is, then, that B has no exact idea of where C lives.
The speaker can also ‘flout’ a maxim:
(26) [In a recommendation letter for a “philosophy job”:] “Mr. X’s command of English is excellent, and his attendance at tutorials has been regular.” (Grice 1989: 33).
In this case, there is no clash between maxims, but there is a clear violation of the maxim of Quantity: the person to whom the letter is directed, obviously wants to know more than what the writer informs him of in (26). Grice states that the writer of (26) “must, therefore, be wishing to impart information that he is reluctant to write down”; in this case, his belief that Mr. X “is no good at philosophy” (ibid.). The hearer, in assuming that the writer of (26) is observing the Cooperative Principle, is “led to the conclusion that” the writer “is trying to convey something else, something more relevant to the purposes at hand” (Neale 1992: 525). In cases like (26), Grice (1989: 33) argues that the Cooperative Principle operates at the level of what is implied, and not ‘what is said’ – Neale (1992: 525) states that “the primary message” in (26) is to be found in the implicature, and that ‘what is said’ doesn’t really matter for the purposes of what the writer is trying to communicate. Its only purpose is to lead the hearer to the conversational implicature that Mr. X would not be a good candidate for the philosophy job.
In sum, Grice assumes that ‘what is said’ is very close to the conventional meaning of the words uttered. Pragmatic inference plays a limited role in establishing ‘what is said’, and is much more active in deriving ‘what is implied’ – the generation of conversational implicatures depends on hearers’ ability to recognize how speakers follow, diverge from and flout the conversational maxims. Note that the conversational maxims are irrelevant in determining ‘what is said’ – it is highly likely that Grice meant for disambiguation and reference assignment to be automatic or ‘mechanical’ processes (Hall 2013: 100).
Grice’s contribution to the field of pragmatics lies in a different direction as well. As stated, his recognition of the crucial role of pragmatic inference in communication is still very much at the center of pragmatics today; however, his focus on speaker intentions has been at least equally influential (Allott 2010: 9). It is important to underline, as Neale (1992: 528) points out, that Grice’s model entails that speakers intend for their audience to derive both ‘what is said’ and ‘what is implied’ – implications which are not speaker-intended are not part of ‘what is implied’, and
cannot be termed ‘implicatures’ (which are always speaker-intended). Conversely, hearers attempt to recover the speaker’s intention in order to infer both ‘what is said’ and ‘what is implied’. In (24), for instance, the speaker wants the hearer to recognize her intention to inform him that there is a garage around the corner which sells petrol; the hearer, on the other hand, aims to find out what it is that the speaker is trying to inform him of (Sperber & Wilson 1995²: 22-23; Neale 1992: 515). In (26), the speaker intends for her interlocutor to recognize her intention to inform him that she thinks that Mr. X is not a good candidate for the philosophy job. In other words, an utterance is a (very specific) clue to the speaker’s intentions – based on the utterance, the hearer is able to ‘read the speaker’s mind’ and infer that her intention was, for example, to inform him about the petrol at the gas station around the corner (Allott 2010: 126). The speaker, for her part, uses her specific utterance to make clear this specific intention – she intends for the hearer to recognize her intention by means of her specific utterance.27 I will go
into more detail on this point later on.
There are several fundamental problems with Grice’s account of both ‘what is said’ and conversational implicatures, and, more specifically, of where to draw the line between those two. In particular, many theorists have pointed to the fact that a much larger role should be carved out for pragmatics in establishing ‘what is said’ – that there is, in more technical terms, much more pragmatic intrusion in Grice’s ‘what is said’ (Huang 2007: 189).