1 Introduction
An āimplicatureā is any meaning a speaker can convey beyond the meaning of what is explicitly said. This meaning can be more than, or different from, the conventional, semantic meaning of the words uttered, as illustrated by Levinsonās classic example (1989, p. 89):
- (1)
A: ĀæPuede decirme la hora?
āCan you tell me the time?ā
B: Bueno, el lechero ya ha pasado.
āWell, the milkman has come.ā
Assuming Bās response is relevant, he conveys (i.e., āimplicatesā) not knowing the precise time, but that it is approximately the time the milkman comes. Bach (2012) notes that the distinction between the meaning of the words used and what a speaker means in uttering them āis perhaps the distinction most basic to pragmaticsā (p. 47). The philosopher H. Paul Grice, who coined the term implicature, proposed his theory of conversational implicature (1975, 1989) to explain how speakers can mean more than what they explicitly state.
In the following section, key distinctions underlying Griceās theory are outlined. Subsequently, I describe the theory, the defining features and types of conversational implicature, and how conversational implicatures can arise. Next, I discuss the class of implicated meaning Grice called āconventional implicatureā and debates it spurred, before identifying some criticisms and misconceptions about Griceās model. I then outline Hornās (1984) and Levinsonās (2000) neo-Gricean theories, which constitute revisions of Griceās original proposal. Afterward, I review various studies on Spanish, which apply classical or revised Gricean frameworks and heuristics, and then discuss some methodological issues associated with Gricean pragmatics. Finally, I offer suggestions for future research utilizing constructs from Gricean and neo-Gricean models.
2 Review of existing research
2.1 Saying, meaning, and implicating
Griceās theory of conversational implicature is rooted in his distinction between saying, meaning, and implicating. Grice (1975) associated what someone has āsaidā with the conventional meaning of the words uttered, and any enriched or inferred content that was not part of the propositional (semantic) content of an utterance could not be, according to Grice, part of what was said.
Grice (1957) also distinguished between ānatural meaning,ā or meaning-N, and ānon-natural meaning,ā or meaning-NN. Meaning-N emphasizes the relationship between observable elements and what they signify, as illustrated by Griceās classic example in (2) (1957, p. 377, my Spanish translation):
By contrast, meaning-NN refers to oneās intentional communication expressed via an utterance. Thus, if a speaker utters
she might intend to communicate (i.e., āmean-NNā) any number of things, from conveying the olfactory sensing of smoke, to suggesting checking the oven. In distinguishing this type of meaning, Grice elucidated the fact that speakersā intentional meanings depend not only on the words used, but also on contextual information and the assumption that we generally aim to make our intentions recognized by our addressees. To capture these ideas, Grice (1975, 1989) proposed a system of conversational logic for the effective, efficient, and rational use of language in conversational interactions known as the theory of conversational implicature.
2.2 Griceās theory of conversational implicature
Griceās theory of conversational implicature is based on the recognition that general principles of language use guide speakers in everyday conversation. The originality of his theory rests on his account of how conversational implicatures arise, their features, and the role of his overarching Cooperative Principle (1975, p. 45):
-
(4) Cooperative Principle:
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.
The Cooperative Principle (henceforward, the CP) captures the idea that speakers generally try to make their utterances appropriate to the conversational context, since otherwise, successful conversations would be impossible. In other words, felicitous conversations rely on the assumption that interlocutors intend to be cooperative and, according to Grice, speakers implicate messages by exploiting this assumption. He instantiates the CP via four categories of conversational maxims, three of which have sub-maxims (1975, pp. 45ā46):
Griceās maxims are principles speakers generally follow and expect others to follow, which together, spell out the CP. The Quantity sub-maxims account for our assumption that speakers should be no more or less informative than necessary. The Quality sub-maxims capture our expectation that speakers will not make specious or unfounded statements, while Relation (also called the āRelevance maximā) captures the assumption that speakers will contribute relevantly to conversations, just as we would offer ibuprofen for a headache and not a band-aid. Finally, the Manner maxim does not concern how much information is conveyed, but rather how speakers express intended meanings via more usual, simpler, clearer, and oftentimes shorter expressions (e.g., casa āhouseā), versus less usual, prolix or obscure, and thus marked expressions (e.g., residencia āresidenceā). Finally, the Manner sub-maxim of ābe orderlyā captures our tendency to relate events in chronological order, and explains differences in meaning between (6a) and (6b), which express the same truth-conditional meaning (i.e., if (a) is true, (b) is true, and vice versa) (from Reyes, 1994, p. 66):
- (6)
Pepa se casĆ³ y tuvo dos hijos.
āPepa got married and had two children.ā
Pepa tuvo dos hijos y se casĆ³.
āPepa had two children and got married.ā
In uttering (6a) or (6b), one conversationally implicates that the events described happened in the order reported. Therefore, saying (6b) implicates that Pepaās children are, in Reyesās words, āextramatrimonialesā āillegitimateā (1994, p. 66). This distinction in meaning is one that truth-conditional semantics could not account for, but which speakers can cancel by adding a phrase like pero no en ese orden ābut not in that order,ā cancelability being one defining characteristic of conversational implicatures. In other contexts, however, no such implicature arises from using y āandā (e.g., Ayer comĆ un montĆ³n y no hice ejercicio āYesterday I ate a lot and didnāt exerciseā), or a cause-consequence interpretation could be generated (CogĆ la gripe y no pude ir āI caught the flu and couldnāt goā). Grice maintained that such inferable meanings were not part of āwhat was saidā but rather what was āconversationally implicated.ā The following describes the features and types of conversational implicature comprising Griceās model.
2.3 Features and types of conversational implicature
Unlike other types of implicational meaning (e.g., presuppositions, semantic entailments), conversational implicatures are derived from Griceās CP and maxims. Additionally, they do not contribute to the truth conditions of the utterances that produced them. Thus, (6a) and (6b) share the same truth conditions but generate different implicatures. Conversational implicatures are also nonconventional meanings. For instance, my use of un millĆ³n (āa millionā) in (7) conversationally implicates muchĆsimo āa lot,ā instead of, literally, 1,000,000:
Conversational implicatures are calculable based on the utterance, the maxims, and contextual knowledge. Furthermore, they are cancelable or defeasible by adding a phrase that nullifies the implicature in certain linguistic or non-linguistic contexts. Cancelability characterizes the two types of conversational implicature identified by Grice (1975): āparticularized conversational implicaturesā (PCIs) and āgeneralized conversational implicaturesā (GCIs). PCIs depend on the context for their generation, while GCIs arise regardless of the context, as (8) illustrates (+> indicates āimplicatesā):
- (8)
Este cafƩ estƔ templado.
āThis coffee is lukewarm.ā
PCI: +> ĀæPuedes calentarme el cafĆ©?
āCan you heat up my coffee?ā
PCI: +> ĀæMe trae un cafĆ© mĆ”s caliente?
āWill you bring me some hotter coffee?ā
PCI: +>ĀæTiro este cafĆ© y hago mĆ”s?
āShould I throw this coffee out and make more?ā
GCI: +> Este cafƩ no estƔ caliente.
āThis coffee is not hot.ā
The PCIs (8bād) generated by (8a) depend on the conversational context (e.g., at home, in a restaurant, in the office, and participantsā shared background knowledge); however, following Gricean theory, the same GCI (8e) arises regardless of context due to the speakerās use of templado āwarm/lukewarm.ā1 Yet, each implicature can be canceled. For instance, after saying (8a), one could add pero me lo tomo igual ābut Iāll drink it anyway,ā canceling the potential particularized implicatures in (8bād); and, one could say, having sipped the coffee, de hecho estĆ” bastante caliente āactually itās quite hotā to nullify the GCI in (8e). The implicature no estĆ” caliente āitās not hotā arising from saying the coffee is templado āwarm/lukewarmā is a type of GCI known as a āscalar generalized Quantity implicature.ā
Scalar GCIs are based on Griceās Quantity 1 maxim and arise from sets of linguistic alternates of the same grammatical category, which can be ordered by degree of informativeness or semantic strength (e.g., <excelente, bueno>, <āexcellent, goodā>). By asserting a stronger item to the left on the scale, the weaker item is entailed. However, using a weaker expression in the set implicates the negation of any stronger members to its left. Thus, Es un buen alumno āHeās a good studentā generates the GCI No es un...