Contextualist approaches: relevance theory

Contextualist approaches are committed to the view that the underlying syntax of what is said should not determine pragmatic meaning representations. In other words, what a speaker is taken to be saying should not be limited to the compositional meaning that arises from what is said1. This move away from a notion of what is said that is constrained by syntactic structure amounts to an emphatic rejection of Grice’s stance that what is said should be treated as primarily a semantic concern, it is for this reason contextualist approaches are generally characterized as post-Gricean. Instead, it is argued by contextualists that we need pragmatics to properly account for what is said. In relevance theory, which is one of the most prominent of the contextualist approaches, this pragmatically enriched notion of what is said is termed an explicature, thereby contrasting it with the notion of implicature (Carston 2002; Sperber and Wilson [1986]1995). The distinction between explicature and implicature rests, in turn, on the (intuitive) distinction between explicitly and implicitly communicated meaning, that is, between explicitness and implicitness.

Explicatures arise, just like implicatures, through inference. However, the former are tied to what is explicitly communicated because they are always “a development of (a) a linguistically encoded logical form of the utterance, or of (b) a sentential subpart of a logical form” (Carston 2002: 124); where logical form refers to a skeleton that can be derived from word meanings and syntactic structure of an utterance. Implicatures, on the other hand, arise solely through inference without any direct contribution from the logical form of the utterance. Notably, Relevance theorists prefer the term logical form over the Gricean notion of what is said, in order to avoid committing themselves to the claim that there is any psychologically plausible layer of pragmatic meaning representation below that of explicatures.

The issue of cognitive plausibility is one that Relevance theorists take very seriously. We will introduce in more detail the key claims made in relevance theory about how pragmatic meaning is understood or processed by hearers. At this point, however, it will suffice to note that Relevance theorists claim that what is explicated (explicature) and what is implicated (implicature) are both figured out by hearers with reference to the Communicative Principle of Relevance, namely, the assumption that “Every act of ostensive communication conveys a presumption of its own optimal relevance” (Sperber and Wilson 1995: 260). Relevance itself is defined in terms of a balance between the cognitive value of particular information, and the processing effort required to gain that information. In other words, something is more relevant when it involves more useful information that takes relatively little effort for the hearer to figure out, and is less relevant when it involves less useful information or takes relatively more effort to figure out for that hearer. In relevance theory, then, the neo-Gricean Cooperative Principle and attendant conversational maxims, which involve normative expectations by which what is implicated and what is said are made available by the speaker, are all replaced by a single cognitively-grounded notion of relevance. This means that for Relevance theorists what is implicated (i.e. implicature) and what is explicated (i.e. explicature) depends on a general tendency in cognitive processing for individuals to maximise the extraction of information at the same time as minimising their efforts in doing so.

The Relevance theoretic notion of explicature, as we have noted above, builds on the claim that we need pragmatics to properly account for what is said2. In deriving these explicitly communicated meanings, or explicatures, Relevance theorists argue there are various different kinds of inferential processes involved. In that sense, relevance theory goes significantly beyond the neo-Gricean account of what is said, which is much more tightly aligned with what a speaker is literally saying1, in arguing for the pragmatically enriched notion of explicature. In order to see the difference between the two positions, let us consider the following quote from the witness, Kato, at the trial of O. J. Simpson about the latter’s state of mind leading up to the murder of his wife:

While seemingly contradictory at the level of what is said1, here Kato is evidently saying2 that while Simpson was emotionally disturbed (i.e. was upset) he wasn’t disturbed to the point of reaching a murderous state of mind (i.e. wasn’t upset). It is also evident that Kato is talking about not just any point in time prior to making this claim (what is strictly speaking said1), but is specifically referring to the time leading up to the murder. The explicature that arises here is that Simpson was, at the time leading up the murder, upsetx but not upsety (where x refers to emotionally disturbed, while y refers to a murderous state of mind). Grice’s notion of what is said is arguably not well equipped to deal with inferences of this nature.

Relevance theory also advances a number of important theoretical claims about implicatures. The first is the observation that not only “conclusions” but also “premises” can be implicated (Sperber and Wilson [1986]1995). In order to see the difference, let us consider the following excerpt from the television series, Everybody Hates Chris. At the point the excerpt begins, Chris and his mother, Rochelle, have just entered a clothing store to buy some new clothes for Chris to wear for taking his picture at school.

We can see here that when the shop assistant responds to Rochelle’s request, he also supposes that Chris is facing, or likely to face, criminal charges and will have to appear in court. This assumption is treated as an implicature, which arises through his response to Rochelle’s request in her first turn, on a Relevance theoretic account. However, this implicature does not logically follow from what he says, but rather is presumed by it, a point we earlier addressed in passing in relation to unpresupposed presuppositions. In other words, it is an implicated premise because it is supplied by others (i.e. Rochelle and Chris) figuring out what assumptions are required for the utterance to be sufficiently relevant on this occasion. That is, for the shop assistant’s response to count as relevant in relation to Rochelle’s request, she and Chris must infer that he is assuming Chris needs the outfit to appear in court. An implicated conclusion subsequently arises through Chris’s response, where he implies that he doesn’t need the outfit for appearing in court, something which naturally follows from his explicitly communicated claim that he is not facing any criminal charges. Carston (2002: 142) argues that the Relevance theoretic distinction between implicated premises and implicated conclusions points to the different “inferential roles that implicatures may play in the derivation process”. However, it is arguably more than just this. The distinction between implicated premises and implicated conclusions in fact refers to two distinct ways in which implicatures can arise, that is, through speakers implying, indicating, suggesting and the like (what Grice termed “implicating”), versus them supposing or assuming and so on (which does not appear to fall under implicating), a point to which we will return.

Here Erica is suggesting that Mark is taking the issue of getting into a good final club far too seriously, to which Mark responds by implying that he is motivated rather than obsessed through claiming there is a difference between the two. However, it is Erica’s response, which Mark subsequently casts as “cryptic” and “speaking in code”, that is of particular interest here. While Erica ostensibly agrees with Mark’s claim that there is a difference between being obsessed and being motivated, it appears that she is implying something else as well. However, what exactly she might be implying is left up to Mark. It could be taken as implying that Mark is indeed simply highly motivated, thereby agreeing with what he has both explicated and implicated. However, it is equally plausible that she is implying Mark is in fact obsessive, thereby agreeing with what he has explicated, but disagreeing with what he has implicated. Given what she has implied here is (deliberately) left up to him to decide, it constitutes an instance of a weak implicature, and for this reason it is difficult for Mark to hold Erica responsible for it. Moeschler (2012) claims that the continuum from strong through to weak implicatures represents one of the most important post-Gricean developments in theorizing implicature, as it is through it that a distinction can be made between implicatures for which the speaker is held primarily responsible (strong implicatures) and those for which are left to the responsibility of the hearer (i.e. weak implicatures). We will also be returning to this point when we consider how pragmatic meanings arise in interactional contexts.

Finally, Relevance theorists also propose that meaning representations can be further embedded within other meaning representations. For example, in the following excerpt from the fi lm, Four Weddings and a Funeral, the shop assistant’s suggestions for a wedding gift are, at first, taken by Charles to be helpful. It subsequently turns out, however, that the suggestions were, in fact, meant sarcastically

In order for Charles to understand that the shop assistant is being sarcastic, he must reconcile the ostensibly helpful suggestion that he buy a pygmy warrior or carrier bags, with the dismissive and evidently unhelpful stance on the part of the shop assistant about Charles’s attempt to buy something in that shop. In other words, a meaning representation involving a thought (i.e. a proposal about what Charles could buy) is here embedded within an attitudinal stance involving the shop assistant’s beliefs (i.e. about Charles’s financial capacity, or lack thereof, to buy something in that shop). Instances where a representation of a thought, belief and the like is embedded within another layer of thought, belief and so on are termed metarepresentations. In the case of pragmatic meaning this involves cases where a meaning representation is embedded within a further, high-order meaning representation. These are sometimes termed higher-order explicatures or implicatures in relevance theory. Metarepresentations appear to be most useful when examining irony and other echoic phenomena, that is, when speakers attribute thoughts to others or when utterances are about the thoughts or utterances of others. Relevance theorists argue these are critical for developing a more nuanced understanding of pragmatic meaning (Wilson 2000). Our view, however, is that such phenomena are better understood from a metapragmatic perspective, as we discuss further, given that they often involve attitudinal stances and other indicators of interpersonal meaning that go beyond an account of pragmatic meaning proper.