Implicature

I am often asked to explain the difference between my notion of impliciture (Bach 1994) and the relevance theorists’ notion of explicature (Sperber and Wilson 1986; Carston 2002). Despite the differences between the theoretical frameworks within which they operate, the two notions seem very similar. Relevance theorists describe explicatures as “developments of logical forms,” whereas I think of implicitures as “expansions” or “completions” of semantic contents (depending on whether or not the sentence’s semantic content amounts to a proposition). That is (...) not much of a difference. We agree that implicitures/explicatures go beyond what is said (in a strict sense) and yet fall short of being implicatures. So, what is the difference, or is it just terminological? As we will see, the real differences emerge when the two notions are situated in their respective theoretical frameworks with their contrasting conceptions of what is involved in linguistic communication. (shrink)

1. Sentences have implicatures. (11, 14, 19)** 2. Implicatures are inferences. (12. 14) 3. Implicatures can’t be entailments. 4. Gricean maxims apply only to implicatures. (16, 17) 5. For what is implicated to be figured out, what is said must be determined first. (12, 13) 6. All pragmatic implications are implicatures. 7. Implicatures are not part of the truth-conditional contents of utterances. (20) 8. If something is meant but unsaid, it must be implicated. (20) 9. Scalar “implicatures” are implicatures. (11) (...) 10. Conventional “implicatures” are implicatures. (shrink)

Grice’s Razor is a methodological principle that many philosophers and linguists have used to help justify pragmatic explanations of linguistic phenomena over semantic explanations. A number of authors in the debate over contextualism argue that an invariant semantics together with Grice’s (1975) conversational principles can account for the contextual variability of knowledge claims. I show here that the defense of Grice’s Razor found in these “Gricean invariantists,” and its use against epistemic contextualism, display all the problems pointed out earlier in (...) Davis (1998). The everyday variation in acceptable knowledge claims is better explained in terms of implicature than indexicality, but general conversational principles shed little light on whether ‘know’ is used hyperbolically, meiotically, or loosely in a context, although this issue is crucial in deciding what if anything ‘S knows p’ implicates. I present reasons favoring an account of the representative bank case in terms of loose use, making clear how they differ from Grice’s Razor. (shrink)

Gricean communication is communication between utterers and their audiences, where the utterer means something and the audience understands what is meant. The weak transmission idea is that, whenever such communication takes place, there is something which is transmitted from utterer to audience; the strong transmission idea adds that what is transmitted is nothing else than what is communicated. We try to salvage these ideas from a seemingly forceful attack by Wayne Davis. Davis attaches too much significance to the surface structure (...) of sentences of the type ‘S communicates the belief (desire …) that p to A’ by assuming that the communicated entity is denoted by the grammatical object following ‘communicates’. On our proposal, what is communicated in all Gricean cases is a thought. And since S communicates the thought that p to A only if S means that p and A understands what S means, the thought that p will be transmitted from S to A. (shrink)

To address complications involving ambiguity, presupposition and implicature, three processes underlying natural language interpretation are isolated: translation, entailment and attunement. A meta- language integrating these processes is outlined, elaborating on a proof-theoretic approach to presupposition.

The neo-Gricean account: the source of these scalar implicatures is a reasoning process (undertaken by the hearer), which culminates in an inference about the belief state of the speaker.

We present an argument for revising the theory of alternatives for Scalar Implicatures and for Association with Focus. We argue that in both cases the alternatives are determined in the same way, as a contextual restriction of the focus value of the sentence, which, in turn, is defined in structure-sensitive terms. We provide evidence that contextual restriction is subject to a constraint that prevents it from discriminating between alternatives when they stand in a particular logical relationship with the assertion or (...) the prejacent, a relationship that we refer to as symmetry. Due to this constraint on contextual restriction, discriminating between alternatives in cases of symmetry becomes the task of focus values. This conclusion is incompatible with standard type-theoretic definitions of focus values, motivating our structure-sensitive definition instead. (shrink)

An utterer may convey a message to her intended audience by means of an explicit statement; or by a non-conventionally mediated one-off signal from which the audience is able to work out the intended message; or by conversational implicature. I investigate whether the last two are equivalent to explicit testifying, as communicative act and epistemic source. I find that there are important differences between explicit statement and insinuation; only with the first does the utterer assume full responsibility for the truth (...) of what she communicates to her audience. (shrink)

This paper proposes that both weak and strong NPIs in English are sensitive to the downward entailingness of their licensers. It is also proposed, however, that these two types of NPIs pay attention to different aspects of the meaning of their environment. As observed by von Fintel and Chierchia, weak NPIs do not attend to the scalar implicatures of presuppositions of their licensers. Strong NPIs see both the truth-conditional and non-truth-conditional (scalar implications, presuppositions) meaning of their licensers. This theory accounts (...) for the puzzling inability, noted by Rullmann and Gajewski, of Strawson anti-additive operators to license strong NPIs, as well as for the effects of Zwarts’s hierarchy of negative strength. Additional issues concerning comparative quantifiers, few, and proportional quantifiers are addressed. (shrink)

Is language understanding a special case of social cognition? To help evaluate this view, we can formalize it as the rational speech-act theory: Listeners assume that speakers choose their utterances approximately optimally, and listeners interpret an utterance by using Bayesian inference to “invert” this model of the speaker. We apply this framework to model scalar implicature (“some” implies “not all,” and “N” implies “not more than N”). This model predicts an interaction between the speaker's knowledge state and the listener's interpretation. (...) We test these predictions in two experiments and find good fit between model predictions and human judgments. (shrink)

[p. 45] I wish to represent a certain subclass of nonconventional implicatures, which I shall call CONVERSATIONAL implicatures, as being essentially connected with certain general features of discourse; so my next step is to try to say what these features are. The following may provide a first approximation to a general principle. Our talk exchanges do not normally consist of a succession of disconnected remarks, and would not be rational if they did. They are characteristically, to some degree at least, (...) cooperative efforts; and each participant recognizes in them, to some extent, a common purpose or set of purposes, or at least a mutually accepted direction. This purpose or direction may be fixed from the start (e.g., by an initial proposal of a question for discussion), or it may evolve during the exchange; it may be fairly definite, or it may be so indefinite as to leave very considerable latitude to the participants (as in a casual conversation). But at each stage, SOME possible conversational moves would be excluded as conversationally unsuitable. We might then formulate a rough general principle which participants will be expected (ceteris paribus) to observe, namely: 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. One might label this the COOPERATIVE PRINCIPLE. On the assumption that some such general principle as this is acceptable, one may perhaps distinguish four categories under one or another of which will fall certain more specific maxims and submaxims, the following of which will, in general, yield results in accordance with the Cooperative Principle. Echoing Kant, I call these categories Quantity, Quality, Relation, and Manner. The category of QUANTITY relates to the quantity of information to be provided, and under it fall the following maxims. (shrink)

Indefinites face competition at two levels: Presupposition and content. The antipresupposition hypothesis predicts that they signal the opposite of familiarity, or uniqueness, namely, novelty, or non-uniqueness. At the level of descriptive content, they are pressured from two sides: definites expressing identity and another phrases expressing difference, and Gricean reasoning predicts that indefinites signal both difference and identity and are infelicitous when definites and another phrases are felicitous. However, occasionally a space opens between the and another, for a to fill. This (...) is in part due to conditions handicapping the or another semantically, in part to another’s phonological handicap. The division of labor between determiners in the field of difference and sameness is thus the result of an intricate competition. We model this competition in a version of game-theoretic pragmatics. (shrink)

Grice’s Razor is a principle of parsimony which states a preference for linguistic explanations in terms of conversational implicature, to explanations in terms of semantic context-dependence. Here I propose a Gricean theory of knowledge attributions, and contend on the basis of Grice’s Razor that it is superior to contextualism about ‘knows’.

To appear in a volume in honor of Grice edited by Klaus Petrus.

1. Implicature: some basic oppositions IMPLICATURE is a component of speaker meaning that constitutes an aspect of what is meant in a speaker’s utterance without being part of what is said. What a speaker intends to communicate is characteristically far richer than what she directly expresses; linguistic meaning radically underdetermines the message conveyed and understood. Speaker S tacitly exploits pragmatic principles to bridge this gap and counts on hearer H to invoke the same principles for the purposes of utterance interpretation. (...) The contrast between the said and the meant, and derivatively between the said and the implicated (the meant-but-unsaid), dates back to the fourth century rhetoricians Servius and Donatus, who characterized litotes—the figure of pragmatic understatement—as a figure in which we say less but mean more (“minus dicimus et plus significamus”; see Hoffmann 1987 and Horn 1991a for discussion). In the classical Gricean model, the bridge from what is said (the literal content of the uttered sentence, computed directly from its grammatical structure with the reference of indexicals resolved) to what is communicated is constructed through implicature. As an aspect of speaker meaning, implicatures are by definition distinct from the non-logical inferences that the hearer draws; it is a category mistake to attribute implicatures either to hearers or to sentences (e.g. P and Q) and subsentential expressions (e.g. some). But we can systematically (at least for generalized implicatures; see below) correlate the speaker’s intention to implicate q (in uttering p in context C), the expression p that carries the implicature in C, and the inference of q induced by the speaker’s utterance of p in C. (shrink)

Lepore and Stone seek to replace the rationality-based Gricean picture of coordination between speaker and hearer with one leaning more strongly on the roles of convention and speaker knowledge while doing away with conversational implicature. Focusing on the phenomena of indirect speech acts, asymmetric conjunction, and scalar inferencing, I argue that the case for abandoning implicature as an analytical tool is not ultimately compelling. I seek further to demonstrate the utility of the classical Gricean distinction between what is said and (...) what is implicated in the construction of an intuitively appealing approach to the perennial question of how to distinguish lying from misleading inside and outside the legal domain. (shrink)

This article argues that what Grice termed ‘particularized conversational implicatures’ can be divided into two types. In some cases, it is possible to reconstruct the inference from the explicit content of the utterance to the implicature without employing a premise to the effect that that the speaker expressed that content (by means of an utterance). I call these ‘material implicatures’. Those whose reconstruction relies on a premise about the speaker's verbal behaviour, by contrast, I call ‘behavioural implicatures’. After showing that (...) the division is theoretically significant, I ask whether current pragmatic theory is able to accommodate it. I conclude that, while (neo)-Gricean pragmatics cannot do so straightforwardly, the distinction is already implicit in Relevance Theory. The article ends by considering the question of whether, in the light of previous discussion, speaker meaning really is always an instance of non-natural meaning. (shrink)

There is an intuitive and seemingly significant difference between lying and falsely implicating. This difference has received scrutiny both historically and recently, mostly in the context of the following two questions. First, how should lying be defined so as to distinguish it from falsely implicating? Second, is the difference between lying and falsely implicating really significant, and if so, how and why is it significant? Answers to the first question typically invoke assertion, claiming (roughly) that to lie is to assert (...) something you take to be false, as opposed to merely implicating something you take to be false. The task is then to spell out the account of assertion that is needed for a satisfactory definition of lying. Answers to the second question divide over whether the difference between lying and falsely implicating has moral significance. This chapter will review and critically evaluate the current state of the literature on these two questions. One conclusion will be that it is not clear that lying can be defined by breaking it down into the components of asserting something, on the one hand, and taking that something to be false (or not taking it to be true), on the other. Another conclusion will be that a puzzle has emerged about the significance of the difference between lying and falsely implicating. There is strong reason to think that the difference is not a moral one, yet alternative explanations of its felt significance are unsatisfying. Finally, an effort will be made to link the two questions by examining the role that norms for assertion might play both in defining lying and in explaining the significance of the difference between lying and falsely implicating. (shrink)

Várias tentativas foram feitas pelos teóricos da referência direta para acomodar o dado intuitivo da opacidade referencial— a não ocorrência de substituição mútua salva veritate de nomes próprios co-referenciais nas orações subordinadas, precedidas por ‘que’, nas orações em que se atribuem atitudes proposicionais. A teoria defendida por Nathan Salmon, em seu livro de 1986 Frege’s Puzzle , é provavelmente a versão mais bem elaborada daquilo a que adiante nos referimos como ‘a Teoria Implicativa’. Salmon sustenta que a opacidade referencial é (...) uma ilusão decorrente de nossa incapacidade de distinguir o conteúdo semântico de atribuições de crenças das suas ‘partilhas pragmáticas’, como Salmon as chama. Lamentavelmente, seu trabalho deixa de todo misterioso o mecanismo rotineiro de produção de tais partilhas pragmáticas. Salmon limita-se a sugerir vagamente que estão envolvidas implicaturas conversacionais griceanas. Minha tese central neste artigo é a de que Salmon está equivocado, visto que as partilhas pragmáticas necessárias à sua teoria não satisfazem o critério de cancelabilidade , o que deveria ocorrer sempre que estamos às voltas com uma genuína implicatura conversacional. O argumento apresentado, até onde o saibamos, é inteiramente original. DOI:10.5007/1808-1711.2010v14n3p405. (shrink)

In this paper I discuss some of the criteria that are widely used in the linguistic and philosophical literature to classify an aspect of meaning as either semantic or pragmatic. With regards to the case of scalar implicature (e.g. some Fs are G implying that not all Fs are G), these criteria are not ultimately conclusive, either in the results of their application, or in the interpretation of the results with regards to the semantics/pragmatics distinction (or in both). I propose (...) a psychologically relevant criterion, that of the primary or secondary role of context. This criterion applies to sub-personal processes that derive the interpretation of a scalar term rather than to the eventual interpretation of the term, and there exist well-established experimental paradigms that can generate quantitative data. I present recent studies on scalar implicature which employ such off-line and real-time paradigms, aiming to demonstrate how research on the semantics/pragmatics distinction can benefit from experimental investigation. (shrink)

The computation of both Scalar Implicatures (SI) and Association with Focus (AF) is characterized with reference to sets of alternatives. However, it has generally been assumed that the relevant alternatives are determined in different ways for the two processes. Specifically, it has been assumed that the alternatives for SI – scalar alternatives – are computed by a special procedure specifically designed for implicatures, whereas the alternatives for AF – focus alternatives – are determined by the general theory of association with (...) focus – focus semantics. As far as we know, the only attempt to connect the two is Krifka (1995), under which scalar alternatives and focus alternatives are identical and determined by focus semantics. However, Krifka’s result is based on a specific stipulation about scalar items, which he borrows from Horn and incorporates into focus semantics, namely that scalar items are inherently focused and have their Horn Scale as their lexically specified focus values. (shrink)

This paper is concerned with the conditions under which a person can be said to have told someone or predicted (the answer to a question like) ‘who came’.

I distinguish between the classical Gricean approach to conversational implicatures, which I call the action-theoretic approach, and the approach to CIs taken in contemporary cognitive science. Once we free ourselves from the AT account, and see implicating as a form of what I call “conversational tailoring”, we can more easily see the many different ways that CIs arise in conversation. I will show that they arise not only on the basis of a speaker’s utterance of complete sentences but also on (...) the basis of sub-sentential clauses—cases of so-called embedded implicatures—as well as from discourse segments containing several sentences—cases that Geurts calls ‘multiplicatures’. I will argue that they arise also from contents that are themselves implicit, such as presupposed contents or other implicatures. All but the first sort of case are difficult for the traditional Gricean AT account to handle, whereas they fall naturally out of an account that sees conversational participants as engaged in conversational tailoring—i.e., as engaged in a process of shaping informational and discourse structural properties of utterances in their successive conversational turns, and hence shaping their interlocutors’ cognitive environments. (shrink)

Grice’s so-called ‘theory of conversation’ (Grice 1967a) establishes a basic distinction between two aspects of utterance meaning: what is said and what is implicated. Some authors (Carston (1988), Recanati (1989), Sperber and Wilson (1986)) have criticized this distinction and, particularly, its application to the pragmatic analysis of several linguistic phenomena, giving rise to an interesting debate on the delimitation of the different aspects of utterance meaning. Bach (1994) enters the discussion with a proposal of revision of Grice’s original distinction, including (...) a new category: what is implicited. The aim of this paper is to participate in this debate paying attention to some questions concerning the Gricean ‘tests’ of cancelability and non-detachability for the different aspects of utterance meaning. More specifically, our claim is that these tests support Bach’s (1994) triple distinction among what is said, impliciture, and implicature, because we can establish the following results. (shrink)

Classical Gricean pragmatics is usually conceived as dealing with far-side pragmatics, aimed at computing implicatures. It involves reasoning about why what was said, was said. Near-side pragmatics, on the other hand, is pragmatics in the service of determining, together with the semantical properties of the words used, what was said. But this raises the specter of ‘the pragmatic circle.’ If Gricean pragmatics seeks explanations for why someone said what they did, how can there be Gricean pragmatics on the near-side? Gricean (...) reasoning seems to require what is said to get started. But then if Gricean reasoning is needed to get to what is said, we have a circle. (shrink)

The British philosopher Herbert Paul Grice observed that the total significance of an utterance embraces not only “what is said” but what is implied. His term of art for the latter was “implicature，” and he identified conversational implicature as an important type of implicit meaning or signification.

Predicates such as tall or to know Latin, which intuitively denote permanent properties, are called individual-level predicates. Many peculiar properties of this class of predicates have been noted in the literature. One such property is that we cannot say #John is sometimes tall. Here is a way to account for this property: this sentence sounds odd because it triggers the scalar implicature that the alternative John is always tall is false, which cannot be, given that, if John is sometimes tall, (...) then he always is. This intuition faces two challenges. First: this scalar implicature has a weird nature, since it must be surprisingly robust (otherwise, it could be cancelled and the sentence rescued) and furthermore blind to the common knowledge that tallness is a permanent property (since this piece of common knowledge makes the two alternatives equivalent). Second: it is not clear how this intuition could be extended to other, more complicated properties of individual-level predicates. The goal of this paper is to defend the idea of an implicature-based theory of individual-level predicates by facing these two challenges. In the first part of the paper, I try to make sense of the weird nature of these special mismatching implicatures within the recent grammatical framework for scalar implicatures of Chierchia (Structures and beyond, 2004) and Fox (2007). In the second part of the paper, I show how this implicature-based line of reasoning can be extended to more complicated properties of individual-level predicates, such as restrictions on the interpretation of their bare plural subjects, noted in Carlson (Reference to kinds in English. Doctoral dissertation, University of Massachusetts at Amherst, 1977), Milsark (Linguistic Analysis 3.1: 1–29, 1977), and Fox (Natural Language Semantics 3: 283–341, 1995); restrictions on German word order, noted in Diesing (Indefinites, 1992); and restrictions on Q-adverbs, noted in Kratzer (The Generic Book, ed. Carlson and Pelletier, 125–175, 1995). (shrink)

Paul Grice coined the term "implicature" in his 1967 William James Lectures. A speaker implicates a proposition if it is part of what he or she communicates, but not part of what he or she literally says. For example, when you direct a stranger to a gas station, you implicate that it is open. This is communicated, but not literally said. The only viable theory of implicature in the philosophical literature is Grice's own theory, which appeals to a "cooperative principle" (...) and a set of "maxims" to explain implicatures. The first chapter of this dissertation argues that Grice's theory does not produce detailed and testable predictions. The remaining three chapters develop a "plan-based theory of implicature", using ideas from computational linguistics. In Chapter Two, I consider the general problem of plan recognition, developing a representation for plans and rules for reasoning about them. Chapter Three discusses the key concept "a speaker's plan", and applies it to several of Grice's examples. Chapter Four extends the analysis to conversations whose point is to construct a domain plan. (shrink)

This paper offers an analysis of metalinguistic comparatives such as more dumb than crazy in which they differ from ordinary comparatives in the scale on which they compare: ordinary comparatives use scales lexically determined by particular adjectives, but metalinguistic ones use a generally-available scale of imprecision or ‘pragmatic slack’. To implement this idea, I propose a novel compositional implementation of the Lasersohnian pragmatic-halos account of imprecision—one that represents clusters of similar meanings as Hamblin alternatives. In the theory that results, existential (...) closure over alternatives mediates between alternative-sets and meanings in which imprecision has been resolved. I then articulate a version of this theory in which the alternatives are not related meanings but rather related utterances, departing significantly from Lasersohn’s original conception. Although such a theory of imprecision is more clearly ‘metalinguistic’, the evidence for it from metalinguistic comparatives in English is surprisingly limited. The overall picture that emerges is one in which the grammatical distinction between ordinary and metalinguistic comparatives tracks the independently motivated distinction between vagueness and imprecision. (shrink)

The paper concerns the expression of non-strict comparison, focusing in particular on constructions of the form [no(t) . . .-er than] in modified numerals. The main empirical finding is the observation that negated comparatives contrast with regular comparatives in that the former but not the latter can give rise to (scalar) implicatures. It is shown that such a contrast falls out of theories of exhaustive interpretation that claim alternatives to form dense scales. An important result is that the paper sharpens (...) the desiderata for theories of exhaustification. (shrink)

The connective or can be treated as an inclusive disjunction or else as an exclusive disjunction. Although researchers are aware of this distinction, few have examined the conditions under which each interpretation should be anticipated. Based on linguistic-pragmatic analyses, we assume that interpretations are initially inclusive before either (a) remaining so, or (b) becoming exclusive by way of an implicature ( but not both ). We point to a class of situations that ought to predispose disjunctions to inclusive interpretations and (...) to situations that encourage exclusive interpretations. A disjunction's ultimate interpretation is based on its potential informativeness, where the interpretation of the disjunctive utterance having the smallest number of true conditions is considered most informative. Our investigation leads to five experiments employing arbitrary materials. Among the problems expected to encourage inclusive interpretations are those that present disjunctions in the antecedents of conditionals and in question forms. The best candidates to produce implicatures are those disjunctions that underdetermine an expected conjunctive conclusion, although other disjunctive utterances that are more informative as exclusive are discussed and tested. (shrink)

Williamson defiende la regla del conocimiento, RK, sobre las aseveraciones: debemos aseverar que p sólo si sabemos que p. En este trabajo exploro algunas consecuencias interesantes de RK: (a) en ocasiones, al hacer una aseveración correcta transmitimos (como implicatura) un significado no literal verdadero, que \sin embargo\ no podría ser correctamente aseverado; (b) ese tipo de implicatura se da, entre otros casos, en una cierta subclase de las implicaturas: las implicaturas argumentativas; (c) RK y la noción de implicatura argumentativa permiten (...) explicar la tendencia a tratar de forma equivalente diferentes tipos de argumentos antiescépticos inspirados en Moore.Williamson defends the knowledge rule, KR, about assertions: one must: assert p only if one knows p. In this work I explore some interesting consequences of KR: (a) sometimes, when making a correct assertion we transmit (as an implicature) a true non-literal meaning which \nevertheless\ could not be rightly asserted; (b) this kind of implicatures are present, for instance, in a certain subclass of implicatures: argumentative implicatures; (c) KR and the notion of argumentative implicature allow for an explanation of the tendency to treat different kinds of Moore-like antiskeptical arguments as if they were equivalent. (shrink)

According to the standard analysis, quantifiers such as , connectives such as , modals such as and a host of other expressions form informational scales (Horn, 1972). In the canonical case, informational scales are defined on the basis of entailment (e.g. p and q asymmetrically entails p or q). Given the Gricean assumption that speakers try to say as much as they truthfully can that is relevant to the conversational exchange, the fact that an informationally weaker term was used in (...) (1)-(3) often gives the listener reason to think that the speaker was not in a position to offer a stronger statement (presumably because such a statement would be false). Thus, even though weak scalar expressions such as some and or have a lower-bounded semantics (for instance some means ‘some and possibly all’), their semantic content is typically upper-bounded by a conversational implicature (e.g. ‘some but not all’). More recently, the precise mechanisms responsible for the computation of scalar implicatures (SIs) have become the topic of much debate. There is considerable disagreement as to whether SIs are derived on the basis of broadly Gricean quantity considerations, as the traditional account would have it, or post- Gricean relevance-oriented computations, whether they are the result of local or global calculations, and whether they are context-specific or generalized, default inferences (for varying perspectives, see Hirschberg, 1985; Sperber and Wilson, 1986/1995; Grice, 1989; Carston, 1990, 1998; Horn, 1992; Levinson, 2000; Chierchia, 2001). In several cases, SIs have been used to motivate and illustrate very different views of the architecture of the semantics-pragmatics interface. Despite their prominent place in the theoretical linguistic literature, scalar inferences have attracted relatively little attention in psycholinguistics.. (shrink)

This paper focuses on the semantic and pragmatic properties of certain aspectual predicates (e.g. start) and degree modifiers (e.g. half). As is wellknown, such terms typically give rise to SCALAR IMPLICATURES (SIs). For instance, an utterance such as (1a) or (2a) is often taken to carry the implicature in (1b) and (2b) respectively.