According to recent work on lexical pragmatics within the relevance-theoretic framework, grasping the intended meaning of a metaphorically used word requires a process of adjusting the linguistically encoded concept to derive an ad hoc concept whose denotation is broader than that of the lexical concept. Metaphorical uses are claimed to be one kind of loose use of language, on a continuum with approximations, hyperboles and other kinds of meaning extension. The question addressed in this paper is whether this account fully captures the processes involved in understanding metaphors and the kinds of cognitive effects they have. We tackle this question by examining the similarities and differences between metaphors and hyperboles and between metaphors and similes. The upshot of our analyses is two proposals, both requiring further investigation: (a) that a distinction should be drawn between the kind of ad hoc concepts derived for hyperbolic and other loose uses, on the one hand, and metaphorical uses, on the other, and (b) that the understanding of some metaphorical uses, in particular extended and/or novel creative cases, is achieved by a different mode of processing altogether, one which gives much greater weight to the literal meaning.

Hyperbolic use of language is very frequent but has seldom been thought worthy of serious analytical attention. Hyperbole is usually treated as a minor trope which belongs with one or the other of the two dominant figurative uses of language, metaphor and irony. In this paper, we examine the range of ways in which hyperbole is manifest, in both its 'pure' uses and its prevalent co-occurrences with other tropes.We conclude that it does not align closely with either metaphor or irony but is a distinctive figure of speech in its own right, characterized by the blatant exaggeration of a relevant scalar property for the purpose of expressing an evaluation of a state of affairs. The relative simplicity of hyperbole enables its exploitation of a range of independent mechanisms of non-literal linguistic communication including loose use, metaphor, simile, and expressions of ironical and other attitudes.Keywords: hyperbole, metaphor, irony, quantitative shift, qualitative shift, evaluative component 3 hyperbole and metaphor are essentially continuous; that is, they are 'not genuinely distinct categories, at least from descriptive, psycholinguistic, or pragmatic points of view' (2008: 95). Irony, by contrast, works quite differently on this account: it is treated as an echoic use of language, and hence as essentially metarepresentational in a way that metaphor and hyperbole are not (Wilson and Sperber 1992, 2012;Wilson 2013).Clearly, at least one of these views cannot be right. So the question presses: where does hyperbole belongwith metaphor or with irony (or with neither)? We shall use this 'metaphor or irony?' question to structure our investigation of hyperbole, but our ultimate conclusion will be that, while the reasoning of each camp provides important insights which any complete account must accommodate, hyperbolic language use is a distinctive phenomenon in its own right and one that combines in interesting ways with a range of other figures of speech. 2The rest of the paper is structured as follows. In section 2, we start with some initial (pre-theoretical) observations about hyperbole and provide some relatively clear examples to work with, including cases that involve hyperbole co-occurring with other figures; in section 3, we set out some necessary background concerning the distinction between metaphor and irony. Then, in sections 4 and 5, we assess the lines of thought that have led some theorists to align hyperbole with irony and others to align it with metaphor, extracting the key insights of these opposing positions. Finally, in section 6, we draw together the conclusions of the earlier sections and argue that there are reasons to think that hyperbole is neither like metaphor nor like irony, but that it functions in a distinctive way that exploits a range of kinds of language use (loose uses, comparisons, categorizations, and expressions of ironical and other kinds of attitude). 2 We want to emphasise from the outset that the uses of language labeled by these terms -'hyperbole', 'metaphor', irony', 'metony...

In this paper, I argue for an account of metaphorical content as what is said when a speaker utters a metaphor. First, I show that two other possibilities-the Gricean account of metaphor as implicature and the strictly semantic account developed by Josef Stern -face several serious problems. In their place, I propose an account that takes metaphorical content to cross-cut the semantic-pragmatic distinction. This requires rethinking the notion of metaphorical content, as well as the relation between the metaphorical and the literal. Semantics, Pragmatics, and What is SaidTo proceed, it is necessary to sharpen the nature of the confl ict over how hearers recover metaphorical content. Two distinctions -between semantics and pragmatics , approaches metaphor on the model of indexicals and demonstratives developed by David Kaplan (1989Kaplan ( , 1979 . It is characteristic of demonstrative expressions that their content changes with the context of utterance, even while there is something uniform about the determination of that content across uses. Determining the content of a demonstrative expression involves the combination of specifi cally linguistic information with information about the context at hand. Using these expressions as a model, Stern offers an account of what is uniform across all instances of metaphorical speech, namely, what he takes to be the specifi cally linguistic knowledge that we bring to bear in understanding metaphors.The key to Stern ' s account is his introduction of an operator, 'Mthat', which acts as the linguistic controller of metaphorical interpretation. 5 In other words, Stern takes this operator to constitute an aspect of our linguistic competence. As with demonstratives, context-dependence is what allows metaphorical content to vary, while the variation is constrained by the Mthat-operator. In placing metaphorical content within the scope of linguistic interpretation in this way, the proposition (metaphorically) expressed by a sentence in the context of utterance is able to constitute what is said.The semantics of the Mthat-operator are as follows. The sub-sentential metaphorical unit, when it is being used metaphorically, does not have its usual representation in the underlying structure of the sentence. Instead, it is represented by the combination of that word or phrase and the Mthat-operator. For example, the metaphor ' Juliet is the sun ' is properly understood as composed of two elements, ' Juliet ' and ' Mthat[ ' is the sun ' ] ' , rather than simply ' Juliet ' and ' is the sun ' . When we hear the words ' Juliet is the sun ' , our linguistic knowledge prompts us to generate a metaphor set of possible interpretations, including all grammatically admissible combinations of Mthat with the various sub-parts of the sentence. 6 5 This operator parallels Kaplan ' s ' Dthat ' operator, introduced to deal with referential uses of defi nition descriptions ( Kaplan, 1979 ). 6 Cf. Stern, 2000 , p. 134. According to Stern, the members of the metaphor set in this case are:

In standard Relevance Theory, hyperbole and metaphor are categorized together as loose uses of language, on a continuum with approximations, category extensions and other cases of loosening/broadening of meaning. Specifically, it is claimed that there are no interesting differences (in either interpretation or processing) between hyperbolic and metaphorical uses (Sperber and Wilson 2008). In recent work, we have set out to provide a more fine-grained articulation of the similarities and differences between hyperbolic and metaphorical uses and their relation to literal uses (Carston & Wearing 2011, forthcoming). We have defended the view that hyperbolic use involves a shift of magnitude along a dimension which is intrinsic to the encoded meaning of the hyperbole vehicle, while metaphor involves a multi-dimensional qualitative shift away from the encoded meaning of the metaphor vehicle. In this paper, we present four experiments designed to test the predictions of this analysis, using a variety of tasks (paraphrase elicitation, self-paced reading and sentence verification). The results of the study support the view that hyperbolic and metaphorical interpretations, despite their commonalities as loose uses of language, are significantly different. IntroductionWhile metaphor is usually treated as a distinctive, even unique, use of language, hyperbole has generally been categorized, by those few theorists and experimentalists who have considered it, as belonging with irony and meiosis (understatement) (Fogelin, 1988; Clark, 1996 3 paying attention to. 3 In section 2, we briefly examine the claim that metaphor and hyperbole are not genuinely distinct from a descriptive point of view. In section 3, we report our experimental study, which challenges the claim that there are no psycholinguistic differences between the two.We conclude in section 4 by clarifying the resulting positions of metaphor and hyperbole. To anticipate: while a single pragmatic mechanism may account for the understanding of both of these non-literal uses of language, significant descriptive and psycholinguistic differences between them must be respected. Descriptive differences between hyperbolic and metaphorical usesAs just noted, the intuition that metaphor is a distinctive, even unique, use of language is widespread. In trying to characterize the special nature of metaphor, many theorists have talked of it as involving a mapping across distinct cognitive domains (e.g.

The pattern of impairments exhibited by some individuals on the autism spectrum appears to challenge the relevance-theoretic account of metaphor (Carston, 1996(Carston, , 2002Sperber and Wilson, 2002;Sperber and Wilson, 2008). A subset of people on the autism spectrum have near-normal syntactic, phonological, and semantic abilities while having severe difficulties with the interpretation of metaphor, irony, conversational implicature, and other pragmatic phenomena. However, Relevance Theory treats metaphor as importantly unlike phenomena such as conversational implicature or irony and like instances of ordinary literal speech. In this paper, I show how Relevance Theory can account for the prima facie incongruity between its treatment of metaphor and the case of individuals with autism.Certain aspects of the pattern of impairments exhibited by some individuals on the autism spectrum seem to pose a challenge to the relevance-theoretic account of metaphor. 1 A subset of individuals with autism-those with so-called 'highfunctioning autism' 2 -appear to have near-normal abilities with respect to syntactic, phonological, and semantic knowledge while having severe difficulties with the interpretation of metaphor, irony, indirect speech acts, fictional discourse, and rhetorical questions. This suggests that there is something importantly different about the comprehension of these phenomena compared with the comprehension of ordinary literal speech, in virtue of which these individuals are able to do the latter but not the former. Any adequate account of metaphor must respect and explain whatever this difference amounts to. 3