Relevance and lexical pragmatics

Tags: Sperber & Wilson, cognitive effects, narrowing, interpretation, account, assumptions, Cambridge University Press, approximations, denotation, blatant violations, precision, metaphor, concept, processing, Robyn Carston, concept construction, mutual adjustment, processes, lexical semantics, Dan Sperber, China Pragmatics Association, Glucksberg, International Pragmatics Association, Relevance theory, Wilson, alcoholic drink, Levinson, pragmatic approaches, blatant violation, metaphorical extension, put to sleep, Sperber, Oxford University Press, literal interpretation, D. & D. Sperber, Working Papers, Department of Linguistics, William James Lectures, Proceedings of the Aristotelian Society, Vega Moreno, Levinson, S., John Benjamins, D. & D. Wilson, UCL
Content: Relevance and Lexical Pragmatics Deirdre Wilson Published in Italian Journal of Linguistics/Rivista di Linguistica 15.2: 273-291(2003) and UCL Working Papers in Linguistics 16: 343-60 (2004) Abstract The goal of lexical pragmatics is to explain how linguistically specified (`literal') word meanings are modified in use. While lexical-pragmatic processes such as narrowing, broadening and metaphorical extension are generally studied in isolation from each other, relevance theorists (Carston 2002, Wilson & Sperber 2002) have been arguing for a unified approach. I will continue this work by underlining some of the problems with more standard treatments, and show how a variety of lexical-pragmatic processes may be analysed as special cases of a general pragmatic adjustment process which applies spontaneously, automatically and unconsciously to fine-tune the interpretation of virtually every word. 1 Introduction Lexical pragmatics is a rapidly developing branch of linguistics that investigates the processes by which linguistically-specified (`literal') word meanings are modified in use.1 Well-studied examples include narrowing (e.g. drink used to mean `alcoholic drink'), approximation (e.g. square used to mean `squarish') and metaphorical extension (e.g. battleaxe used to mean `frightening person'). In much of the literature, narrowing, approximation and metaphorical extension have been seen as distinct pragmatic processes and studied in isolation from each other. I will * Earlier versions of this paper were presented at the European Society for Philosophy and Psychology, the International Pragmatics Association and the 8th China Pragmatics Association. Many thanks to the organisers and participants, to Dan Sperber and Robyn Carston for valuable ideas and inspiration, and to Tim Wharton, Rosa Vega-Moreno, Paula Rubio-Fernandez and the members of the UCL Pragmatics Reading Group. I would like to acknowledge the support of the AHRB under grant MRG-AN9291/APN16356 `A unified theory of lexical pragmatics'. 1 See e.g. Grice 1967; Ducrot 1972, 1984; Searle 1979; Lakoff & Johnson 1980; Sperber & Wilson 1985, 1998; Cruse 1986; Hobbs & Martin 1987; Lakoff 1987; Lahav 1989; Sweetser 1990; Horn 1992, 2000; Aitchison 1994; Bach 1994, 2001; Gibbs 1994; Copestake & Briscoe 1995; Franks 1995; Recanati 1995, 2004; Rips 1995; Bertuccelli Papi 1997; Carston 1997, 1999, 2002; Blutner 1998, 2002; Lascarides & Copestake 1998; Lasersohn 1999; Fauconnier & Turner 2002; Wilson & Sperber 2002, 2004; Merlini Barbaresi 2003.
defend the alternative view that they are outcomes of a single pragmatic process which fine-tunes the interpretation of virtually every word.2 I will adopt a simple model of linguistic semantics that treats words as encoding mentally-represented concepts, elements of a conceptual representation system or `language of thought', which constitute their linguistic meanings and determine what might be called their linguistically-specified denotations.3 The goal of lexical semantics is to investigate the relations between words and the concepts they encode, and the goal of lexical pragmatics is to account for the fact that the concept communicated by use of a word often differs from the concept encoded. NARROWING is the case where a word is used to convey a more specific sense than the encoded one, resulting in a restriction of the linguistically-specified denotation. Approximation and metaphorical transfer may be seen as varieties of BROADENING, where a word is used to convey a more general sense, with consequent widening of the linguistically-specified denotation. The effect of narrowing is to highlight a proper subpart of the linguisticallyspecified denotation. Here are some illustrations: (1) All doctors drink. (2) a. As I worked in the garden, a bird perched on my spade. b. Birds wheeled above the waves. c. A bird, high in the sky, invisible, sang its pure song. d. At Christmas, the bird was delicious. (3) Mary is a working mother. (4) I have a temperature. In (1), drink might convey not the encoded sense `drink liquid' but, more specifically, `drink alcohol', or `drink significant amounts of alcohol'. In (2a-d), each use of bird would highlight a different subset of birds. As noted by Lakoff (1987: 80-82), (3) would generally indicate not just that Mary satisfies the definition `female parent who works', but that she is a stereotypical working mother, bringing up young children while working for money outside the home; and (4) would normally convey not the truism that the speaker has some temperature or other but that her temperature is high enough to be worth remarking on. APPROXIMATION is a variety of broadening where a word with a relatively strict sense is extended to a penumbra of cases (what Lasersohn 1999 calls a `pragmatic halo') that strictly speaking fall outside its linguistically-specified denotation. 2 For elaboration of this view, see e.g. Carston 1997, 2002 chap. 5; Sperber & Wilson 1998; Wilson & Sperber 2002, 2004. 3 For an approach along these lines, see Fodor 1998. For relevance-theoretic accounts of concepts, see e.g. Sperber & Wilson 1986/95 chap. 2, 1998; Sperber 1996 chap 6; Carston 2002 chap 5; Wilson & Sperber 2002.
Loose uses of round numbers, geometric terms and negatively-defined terms are good examples, as in (5-7): (5) This coat cost 1,000 dollars. [`about 1,000 dollars'] (6) The stones form a circle, an oval, a pyramid. [`approximately a circle'] (7) This injection will be painless. [`nearly painless'] As with narrowing (cf. (2) above), different degrees and types of approximation are appropriate in different circumstances; compare the interpretations of flat in (8a-e): (8) a. This ironing board is flat. b. My garden is flat. c. My neighbourhood is flat. d. My country is flat. e. The Earth is flat. A second variety of broadening, which I will call CATEGORY EXTENSION, is typified by the use of salient brand names (Hoover, Kleenex) to denote a broader category (`vacuum cleaner', `disposable tissue') including items from less salient brands. Personal Names (Chomsky, Einstein) and common nouns both lend themselves to category extension (cf. Glucksberg 2001: 38-52). Some more creative uses are illustrated in (9-12): (9) Federer is the new Sampras. (10) Brown is the new black. (11) Mint is the new basil. (12) Is oak the new pine? In (9), Sampras evokes the category of gifted tennis players of a certain type. In (10) ­ a typical piece of fashion writer's discourse ­ black evokes the category of staple colours in a fashion wardrobe; echoes are found in cookery and interior design writing, as in (11) (`herb of the moment') and (12) (`trendy furniture wood'). These examples of category extension are not analysable as approximations. The claim in (10) is not that Federer is a borderline case, close enough to being Sampras for it to be acceptable to call him Sampras, but merely that he belongs to a broader category of which Sampras is a salient member; and so on for the other examples. Metaphor and hyperbole may be thought of as more radical varieties of category extension.4 For example, (13) would be an approximation if used to indicate that 4 See e.g. Recanati 1995, 2004; Carston 1997, 1999, 2000; Sperber & Wilson 1998; Glucksberg 2001; Wilson & Sperber 2002.
the water was close enough to boiling to be described as boiling, and a hyperbole if used to indicate that the water was merely hotter than expected, or uncomfortably hot: (13) The water is boiling. The metaphors in (14-16) are analysable on similar lines, as radical extensions of the linguistically-specified denotation: (14) Mary is a rose, a lily, a daisy, a violet; a jewel, a diamond, a ruby, a pearl. (15) That book puts me to sleep. (16) The leaves danced in the breeze. Thus, violet in (14) might be seen as representing the category of delicate, unflamboyant, easily overlooked things, of which violets are a salient subcategory, and so on for other examples. Neologisms and word coinages provide further data for a theory of lexical pragmatics and shed some light on the nature of the mental mechanisms involved. Experiments by Clark & Clark (1979) and Clark & Gerrig (1983) show that newlycoined verbs derived from nouns, as in (17-19), are no harder to understand than regular verbs: (17) The newspaper boy porched the newspaper. (18) They Learjetted off to Miami. (19) He Houdinied his way out of the closet. This suggests that lexical-pragmatic processes apply `on line' in a flexible, creative and context dependent way, and may contribute to the explicit truth-conditional of utterances (in Grice's terms, `what is said') as well as to what is implicated (Carston 2002; Wilson & Sperber 2002). Any discussion of lexical pragmatics must make some assumptions about the nature of the semantic representations that provide the input to pragmatic processes. Synchronically, where the borderline between lexical semantics and pragmatics falls in individual cases is not always clear, and it may be drawn in different ways in the minds of different individuals. Moreover, the repeated application of lexicalpragmatic processes may lead to semantic change: what starts as a spontaneous, one-off affair may become regular and frequent enough to stabilise in a community and give rise to an extra sense.5 My interest here is not so much in the details of individual cases as in the pragmatic processes that apply spontaneously, automatically and unconsciously to fine-tune the interpretation of virtually every 5 See e.g. Lyons 1977; Hopper & Traugott 1993; Bertuccelli Papi 2000.
word, and I will therefore largely abstract away from the question of whether, or when, a word like drink, or Hoover, or flat may be said to have acquired an extra stable sense.6 My account will be consistent with the view that some words are strictly defined and loosely used, while others have broader, vaguer meanings which are typically narrowed in use. 2 Some existing accounts Many philosophical and Pragmatic Approaches seem to take for granted that narrowing, approximation and metaphorical extension are distinct pragmatic processes, which lack common descriptions or explanations and need not be studied together. For example, narrowing is often analysed as a case of default inference to a stereotypical interpretation,7 approximation has been seen as linked to variations in the standards of precision governing different types of discourse,8 and metaphor is generally treated on Gricean lines, as a blatant violation of a maxim of truthfulness, with resulting implicature.9 These accounts do not generalise: metaphors are not analysable as rough approximations, narrowings are not analysable as blatant violations of a maxim of truthfulness, and so on. Moreover, there are internal descriptive and theoretical reasons for wanting to go beyond these existing philosophical and pragmatic accounts. Levinson (2000: 37-8, 112-34) treats narrowing as a default inference governed by an Informativeness heuristic (`What is expressed simply is stereotypically exemplified'), itself backed by a more general I-principle instructing the hearer to Amplify the informational content of the speaker's utterance, by finding the most specific interpretation, up to what you judge to be the speaker's m-intended point ... (ibid: 114). The I-heuristic might be seen as dealing with stereotypical narrowings such as (3) above, and the I-Principle as dealing with less stereotypical cases such as (4). However, this approach leaves many aspects of the narrowing process unexplained. In the first place, there may be several possible degrees or directions of narrowing, as in (1) (where drink may be narrowed to `drink alcohol' or `drink a lot of alcohol') and (2) (where bird is narrowed in different ways in different contexts). 6 Appeals to polysemy are probably justified in many cases. However, since each encoded sense of a polysemous word may undergo further pragmatic processing, polysemy does not eliminate the need for lexical pragmatics. 7 See e.g. Horn 1984, 1992, 2000; Levinson 2000; Blutner 1998, 2002. For discussion, see Lakoff 1987. 8 See e.g. Lewis 1979; Lasersohn 1999. For discussion, see Gross 2001. 9 See e.g. Grice 1975, Levinson 1983.
Levinson (ibid: 118) notes (and experimental evidence confirms, cf. Barsalou 1987) that even stereotypical narrowing is context dependent. For example, Englishman in (20) would evoke different stereotypes in a discussion of cooking, cricket, sailing, seduction, etc: (20) John is an Englishman. In the second place, stereotypical narrowing also competes with other varieties of narrowing.10 In (21), man might be narrowed to an idealised rather than a stereotypical interpretation, indicating that Churchill is a man worthy of the name rather than a typical man: (21) Churchill was a man. According to the I-Principle, the hearer of (1-4) and (20-21) should choose the appropriate degree and direction of narrowing using his judgement about `the speaker's m-intended point' (i.e. the speaker's meaning) ­ a judgement which is therefore presupposed rather than explained by this account. Levinson acknowledges the context dependence of I-implicatures, but maintains (ibid: 118) that `at a sufficient level of abstraction' they are default inferences (generalised implicatures) which `hold as preferred interpretations across contexts, and indeed across languages.' However, as illustrated in (5-19) above, broadening appears to be just as strong a tendency as narrowing at the lexical level, and it is not clear why narrowing rather than broadening should be seen as the default. Nor is it clear how an approach based on the I-principle could generalise to approximations or metaphors (which Levinson appears to treat as blatant violations of a maxim of truthfulness in the regular Gricean way). It is therefore worth looking for an alternative, more explanatory account. Lewis (1983: 244-45) treats approximation as a type of pragmatic vagueness governed by contextually-determined standards of precision: When is a sentence true enough? [...] this itself is a vague matter. More important [...], it is something that depends on context. What is true enough on one occasion is not true enough on another. The standards of precision in force are different from one conversation to another, and may change in the course of a single conversation. Austin's `France is hexagonal' is a good example of a sentence that is true enough for many contexts but not true enough for many others. 10 For an interesting survey of many varieties of narrowing, see Lakoff 1987.
To shed any light on how approximations are understood, this approach would have to be supplemented by some account of how the appropriate standard of precision is formulated, and how it may change in the course of a conversation. Consider (22): (22) The lecture will start at 5.00 and end at 6.00. As noted in Wilson & Sperber (2002: 592-8), a student with several lectures to attend might accept (22) as `true enough' if the lecture starts ten minutes late and ends ten minutes early, but not if it starts ten minutes early and ends ten minutes late. This asymmetry would somehow have to be built into the standards of precision in force. Moreover, what counts as `true enough' for a student planning a day's lectures would not be `true enough' for a sound engineer getting ready to broadcast the lecture live. This would have to be accommodated in an account of how the appropriate standard of precision is determined. Of course, all this is predictable enough in a common-sense way, given background assumptions and general expectations about human behaviour. The problem with the appeal to `contextually-determined standards of precision' is that it seems to be acting as little more than a placeholder for a more detailed pragmatic account. As suggested above, Lewis's account of approximation does not generalise to category extension, metaphor or hyperbole. Approximations are appropriate in borderline cases; category extension, metaphor and hyperbole are not. Lewis himself proposes separate analyses of approximation and figurative utterances (which he treats as encoding figurative meanings in the traditional semantic way; cf. Lewis 1983: 1983; Wilson & Sperber 2002: 587-9). Separate analyses could be justified by showing that there is a definite cut-off point between approximation and figurative utterances; but it is doubtful that such a cut-off point exists. Thus, (13) above has a gradient of interpretations with clear approximations at one end, clear hyperboles at the other and a range of borderline cases in between; (7) above, which I have treated as an approximation, could equally well be classified as a hyperbole; and examples such as (9-12) above, which I have treated as cases of category extension, are sometimes classified as metaphors (Glucksberg 2001: v, 47). The lack of a clear cut-off point between approximation, category extension, hyperbole and metaphor also raises problems for Grice's analysis of figurative utterances. On Grice's approach, (13-14) above would be analysed as blatant violations of the maxim of truthfulness (`Do not say what you believe to be false'), implicating (23-24), respectively: (23) This water is very hot. (24) Mary resembles a rose in some respects.
As noted by Wilson & Sperber (2002: 593-4), this account of metaphor and hyperbole does not generalise to approximations, which are generally perceived as `true enough' rather than blatantly false. Moreover, it is hard to see where on the gradient of interpretations an utterance of (7) or (13) might stop being `true enough' and start being `blatantly false'. There are also well-known descriptive and theoretical problems with Grice's account. For one thing, the literal interpretations of negative metaphors such as (25) are trivially true rather than blatantly false: (25) Mary is no angel. For another, Grice's analysis, taken as a model of the comprehension process, predicts that hearers should consider the literal interpretation first, and move to a figurative interpretation only if the literal interpretation is blatantly false. Yet there is both experimental and introspective evidence that this prediction is false (e.g. Gibbs 1994; Glucksberg 2001). In interpreting (15) above (an example due to Dan Sperber), for instance, it may never even occur to the hearer to wonder whether the book literally put the speaker to sleep. All this suggests that it is worth trying to develop a more general account of lexical-pragmatic processes, which acknowledges their flexibility, creativity and context dependence, and treats them as applying spontaneously, automatically and unconsciously during on-line comprehension to fine-tune the interpretation of virtually every word. In the next section, I will outline an account which brings together ideas from experimental studies of categorisation and metaphor,11 on the one hand, and from relevance theory and other recent work in pragmatics,12 on the other. Experimental studies of categorisation by Barsalou (e.g. 1987, 1992) support the view that lexical narrowing cannot simply be analysed as default inference to a ready-made stereotype or prototype. In the first place, typicality judgements about existing categories (e.g. BIRD, ANIMAL) are quite variable across individuals, contexts and times, and the appeal to ready-made stereotypes or prototypes does not explain this variability. In the second place, people can readily provide typicality rankings for made-up categories that they could not have encountered before (e.g. THINGS THAT CAN FALL ON YOUR HEAD), or predict the typicality rankings for familiar categories from the point of view of real or imagined individuals for whom they would be most unlikely to have ready-made prototypes stored. Barsalou sees these facts as best explained by assuming that the content of a category on a particular occasion is not determined by accessing a ready-made 11 See e.g. Barsalou 1987; Gibbs 1994; Glucksberg et al. 1997; Glucksberg 2001. 12 See e.g. Recanati 1995, 2004; Carston 1997, 2002; Sperber & Wilson 1998; Rubio Fernandez 2001; Vega Moreno 2001, 2003, 2004; Wilson & Sperber 2002.
stereotype or prototype, but is constructed on-line in an ad hoc, context-specific way, from a reservoir of encyclopaedic information which varies in accessibility from individual to individual and situation to situation, with different subsets being chosen on each occasion. This idea clearly has implications for lexical narrowing. However, apart from noting that the choice of a particular subset of encyclopaedic assumptions is affected by discourse context, the accessibility of information in memory and considerations of relevance, he does not provide a concrete pragmatic hypothesis about how the narrowing process might go. As noted above, experimental studies of metaphor by Gibbs (1994) and Glucksberg (2001) confirm the inadequacy of models of comprehension based on the standard Gricean approach by showing that a literal interpretation does not always have to be considered and rejected before moving to a metaphorical interpretation. Glucksberg proposes that metaphor should instead be analysed as a variety of category extension. On this approach (tacitly adopted in section 1 above), just as Hoover may be used to represent the broader category of vacuum cleaners of which it is a salient member, so Sampras may be used to denote a broader category of gifted tennis players of which he is a salient member, and violet may be used to denote a broader category of delicate, unflamboyant, easily-overlooked things, of which it is a salient member. Glucksberg (2001: 46) comments: Good metaphors ... are acts of classification that attribute ... an interrelated set of properties to their topics. It follows that metaphoric comparisons acquire their metaphoricity by behaving as if they were class-inclusion assertions. On this approach, narrowing and metaphor are complementary processes, one restricting and the other extending the category denoted by the linguisticallyencoded concept. Barsalou's and Glucksberg's accounts share the assumption that encyclopaedic information associated with a mentally-represented category or concept may be used to restrict or extend its denotation in an ad hoc, occasion-specific way.13 Thus, encyclopaedic information about attributes of various subsets of birds may be used to highlight a particular subpart of the denotation of `bird' or evoke a broader category, e.g. the category of flying things. Glucksberg and Barsalou both mention the role of considerations of relevance in selecting an appropriate set of attributes, but make no attempt to develop a full pragmatic account of what factors trigger lexical-pragmatic processes, what direction they take, and when they stop. For many years, relevance theorists have been pursuing the idea that lexical comprehension involves ad hoc broadenings or narrowings of encoded concepts 13 For further discussion of Glucksberg's and Barsalou's accounts, see Carston 2002:chapter 5, especially notes 1, 14.
based on the use of encyclopaedic information constrained by expectations of relevance.14 In the next section, I will consider what light the theory might shed on how the ad hoc concept construction process might go. 3 Relevance theory and lexical pragmatics Relevance theory is based on a definition of relevance and two general principles: a Cognitive and a Communicative Principle of Relevance.15 Relevance is characterised in cost-benefit terms, as a property of inputs to cognitive processes, the benefits being positive cognitive effects (e.g. true contextual implications, warranted strengthenings or revisions of existing assumptions) achieved by processing the input in a context of available assumptions, and the cost the processing effort needed to achieve these effects. Other things being equal, the greater the positive cognitive effects achieved, the greater the relevance of the input to the individual who processes it. However, the processing of the input, the accessing of contextual assumptions and the derivation of positive cognitive effects involves some effort of perception, memory and inference. Other things being equal, the smaller the processing effort required, the greater the relevance of the input. According to the First, or Cognitive, Principle of Relevance (Sperber & Wilson 1995: 260-66), the human cognitive system tends to allocate attention and processing resources so as to maximise the relevance of the inputs it processes. As a result of constant selection pressure towards increasing cognitive efficiency, our perceptual mechanisms tend automatically to pick out potentially relevant inputs, our memory retrieval mechanisms tend automatically to pick out potentially relevant contextual assumptions, and our inferential systems tend spontaneously to process them in the most productive way. Communicators should therefore be able to predict, at least to some extent, what stimuli an addressee is likely to attend to, what contextual assumptions he is likely to use in processing them, and what conclusions he is likely to draw. 14 See e.g. Sperber & Wilson 1983, 1986/95, 1998; Sperber 1989, 1997, 2000a; Wilson 19902003, 1995; Carston 1997, 1999, 2002; Sperber & Wilson 1998; Papafragou 2000; Wilson & Sperber 2002, 2004; He 2003. In earlier versions of the theory, implicatures based on loose and metaphorical uses of concepts were not seen as affecting explicatures via backwards inference during the mutual adjustment of explicit content, context and cognitive effects (for discussion of these notions, see section 3 below). In later versions, with the introduction of the mutual adjustment process (e.g. Sperber & Wilson 1998; Wilson & Sperber 2002, 2004), and particularly thanks to Robyn Carston's work (e.g. Carston 1997, 1999, 2002 chap. 5), the idea of ad hoc concept construction has been more fully incorporated into the theory. 15 For detailed accounts of the current version of the theory, see Sperber and Wilson 1995; Carston 2002; Wilson & Sperber 2002, 2004.
According to the Second, or Communicative, Principle of Relevance (Sperber & Wilson 1995: 266-71), utterances create general expectations of relevance. The addressee of an utterance is entitled to expect it to be at least relevant enough to be worth processing (and hence more relevant than any alternative input available to him at the time), and moreover, the most relevant utterance compatible with the speaker's abilities and preferences. This motivates the following comprehension procedure which, according to relevance theory, is automatically applied to the online processing of attended verbal inputs. The addressee takes the linguistically decoded meaning: following a path of least effort, he enriches it at the explicit level and complements it at the implicit level until the resulting interpretation meets his expectations of relevance; at which point, he stops. This mutual adjustment of explicit content, contextual assumptions and cognitive effects constrained by expectations of relevance is the central feature of relevance-theoretic pragmatics.16 This approach to utterance comprehension has two important consequences for lexical pragmatics. In the first place, there is no presumption of literalness: the linguistically encoded meaning (of a word, a phrase, a sentence) is no more than a clue to the speaker's meaning, which is not decoded but non-demonstratively inferred. In the second place, understanding any utterance, literal, loose or metaphorical, is a matter of seeing its intended relevance, and seeing the intended relevance of an utterance is a matter of following a path of least effort in mutually adjusting explicit content, context and cognitive effects, as specified in the relevance-theoretic comprehension procedure. Relevance theory therefore suggests the following answers to the basic questions of lexical pragmatics: lexicalpragmatic processes are triggered by the search for relevance, they follow a path of least effort, they operate via mutual adjustment of explicit content, context and cognitive effects, and they stop when the expectations of relevance raised by the utterance are satisfied (or abandoned). To illustrate how this account might apply to the narrowing of temperature in (4), consider the following scenario. Peter has just suggested that he and Mary pay a visit to his aunt in hospital, and Mary replies as in (4): (4) I have a temperature. In the circumstances, Peter will have not only a general expectation of relevance but a particular expectation about how Mary's utterance is likely to achieve relevance at this particular point in the discourse: he will be expecting it to achieve 16 Mutual adjustment is seen as taking place in parallel rather than in sequence. The hearer does not first identify the proposition expressed, then access an appropriate set of contextual assumptions and then derive a set of cognitive effects. In many cases (notably in indirect answers to questions, or when a discourse is already under way), he is just as likely to reason backwards from an expected cognitive effect to the context and content that would warrant it. See e.g. Sperber & Wilson 1998; Carston 2002; Wilson & Sperber 2002, 2004.
relevance as a response to his suggestion that they visit his aunt in hospital. Literally interpreted, of course, her utterance is trivially true and achieves no positive cognitive effects. However, temperature is a scalar term, and different points on the scale should yield different implications when combined with easily accessible contextual assumptions. Assuming some version of a spreading activation model of memory, Peter's encyclopaedic assumptions about temperatures, hospital visits and the possible connections between them should be highly activated at this point. It should therefore be a relatively straightforward matter, by following a path of least effort in the mutual adjustment of content, context and cognitive effects, to arrive at an interpretation on which temperature expresses an ad hoc concept TEMPERATURE*, denoting a temperature high enough to make it inadvisable for Mary to visit Peter's aunt in hospital.17 More generally, narrowing is undertaken in the search for relevance. Hearers satisfy their expectations of relevance by looking for true implications (or other positive cognitive effects). Narrowing increases implications. A hearer following the relevance-theoretic comprehension procedure is therefore entitled to narrow the encoded sense to a point where it yields enough true implications to satisfy the general expectation of relevance raised by the utterance, together with any more specific expectations raised by the fact that the utterance has been produced by that speaker, for that audience, at that particular point. If several possible narrowings are available, he follows a path of least effort, using whatever assumptions and expectations are most highly activated (e.g. by the utterance itself and by preceding discourse). If he finds enough true implications to satisfy his expectations of relevance, he assumes that this was the speaker's meaning; if not, he tries another route.18 Similar analyses apply to approximation, category extension, metaphor and hyperbole. Consider the category extension in (9), which was used by many commentators during Wimbledon 2003 (which Roger Federer eventually won): (9) Federer is the new Sampras. For many hearers, the encoded concept SAMPRAS would provide access to a wide array of encyclopaedic assumptions about Sampras, some of which will receive additional activation from the mention of Federer and from the discourse context, including the fact that the utterance was produced during Wimbledon 2003. Although these highly activated assumptions will differ from hearer to hearer, they 17 On the treatment of ad hoc concepts in relevance theory, see Sperber & Wilson 1998; Carston 2002 chap. 5; Wilson & Sperber 2002. 18 It might be interesting to make a detailed comparison between this analysis, based on theoretical characterisations of relevance and expectations of relevance, backed by relevanceoriented accounts of cognition and communication, and an alternative based on Levinson's Iprinciple, with its largely uncharacterised notion of `speaker's m-intended point'.
are likely to include the information that Sampras is a formidably gifted natural player of a certain type, that he has won Wimbledon many times and played a leading role in the tournament over many years, and so on. In these circumstances, a hearer following the path of least effort and looking for true implications (or other positive cognitive effects) via mutual adjustment of content, context and cognitive effects is likely to arrive at an interpretation in which Sampras expresses an ad hoc concept SAMPRAS* which denotes not only Sampras but other players with these encyclopaedic attributes, and conclude that the speaker is claiming that Federer falls into this ad hoc category and is therefore likely to dominate Wimbledon for many years, etc. Finally, consider how an account along these lines might apply to the interpretation of put to sleep in (15): (15) That book puts me to sleep. On the Gricean approach, (15) should have three distinct interpretations: as a literal assertion, a hyperbole or a metaphor. Of these, the hearer should test the literal interpretation first, and consider a figurative interpretation only if the literal interpretation blatantly violates the maxim of truthfulness. Yet as noted above, there is both experimental and introspective evidence against this approach when construed as a model of comprehension. On the relevance-theoretic account, there is no presumption that the literal meaning will be tested first. The encoded concept PUT TO SLEEP is merely a point of access to an ordered array of encyclopaedic assumptions from which the hearer is expected to select an appropriate subset. Let us suppose that Mary has produced (15) in response to Peter's question "What do you think of Martin's latest book?" He will therefore be expecting her utterance to achieve relevance by answering his question: that is, by offering an evaluation of the book. Given this expectation, her utterance is likely to activate the contextual assumption that a book which puts one to sleep is likely to be extremely boring and unengaging. By following a path of least effort in the mutual adjustment of context, content and cognitive effects, he should then arrive at an interpretation on which put to sleep expresses the ad hoc concept PUT TO SLEEP*, which denotes a broader category containing not only literal cases of putting to sleep, but other cases that share with it the encyclopaedic attribute of being extremely boring and unengaging. Only if such a loose interpretation fails to satisfy his expectations of relevance would Peter be justified in exploring further contextual assumptions, and moving towards a more literal interpretation.19 On this approach, broadening, like narrowing, is undertaken in the search for relevance and results from the mutual adjustment of context, content and cognitive 19 For further discussion of this example, see Wilson & Sperber 2004.
effects, constrained by expectations of relevance raised by the utterance itself. In many cases, the mutual adjustment process will converge on a broader or narrower category than the linguistically-specified denotation, with effects that have been roughly classified in the literature as narrowing, approximation, category extension, metaphor, hyperbole and so on. The main point of this paper has been to argue that these taxonomies do not pick out genuine natural kinds. There is no clear cut-off point between the different varieties of broadening. Moreover, as noted by Carston (1997), narrowing and broadening may combine, so that a single word may express an ad hoc concept that is narrowed in some respects and broadened in others. In the domain of lexical pragmatics, these taxonomies have led to a fragmentation of research programmes and obscured some interesting generalisations that hold across the whole domain. It is worth systematically exploring the possibility of developing alternative, more unified accounts. 4 Concluding remark In the Oral Presentation of this paper at the 8th China Pragmatics Association conference in December 2003, I invited collaboration from colleagues in China on the comparative study of lexical pragmatic processes. Chinese scholars in pragmatics have already made a major contribution to our understanding of social, pragmatic and cognitive Factors Affecting the use and understanding of metaphor and vagueness (e.g. He 2003). However, we know rather less about cross-cultural similarities and differences in narrowing, category extension and approximation. In discussion after the paper, (16) and (17) were suggested as examples of narrowing in Chinese: (16) Jane has money. [`enough money to be worth remarking on'] (17) John has a reputation. [`a good reputation'] Both are also possible examples of narrowing in English: indeed, (17) has two possible narrowings, meaning either `a good reputation' or `a bad reputation'. This raises interesting questions about the extent to which lexical-pragmatic processes such as narrowing, approximation and category extension are universal, and how potential variations should be explained. Further collaboration on these issues would be very welcome. References Aitchison, J. (1994) Words in the Mind: An Introduction to the Mental Lexicon. Oxford: Blackwell.
Bach, K. (1994) Conversational impliciture. Mind & Language 9: 124-162. Bach, K. (2001) Speaking loosely: Sentence nonliterality. In French & Wettstein (2001: 249-263). Barsalou, L. (1987) The instability of graded structure: implications for the nature of concepts. In Neisser (1987: 101-140). Barsalou, L. (1992) Frames, concepts, and conceptual fields. In Kittay & Lehrer (1992: 21-74). Bertuccelli Papi, M. (1997). Implicitness. In Verschueren et al, eds. Bertuccelli Papi, M. (2000) Is a diachronic speech-act theory possible? Journal of Historical Pragmatics 1: 57-66. Blutner, R. (1998) Lexical pragmatics. Journal of Semantics 15: 115-62. Blutner, R. (2002) Lexical semantics and pragmatics. Linguistische Berichte 10: 27-58. Carruthers, P. & J. Boucher, eds. (1998) Language and Thought. Cambridge: Cambridge University Press. Carston, R. (1997) Enrichment and loosening: complementary processes in deriving the proposition expressed? Linguistische Berichte 8: 103-127. Carston, R. (2000). Explicature and semantics. UCL Working Papers in Linguistics 12: 1-44. Revised version to appear in Davis & Gillon (forthcoming). Carston, R. (2002). Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford, Blackwell. Carston, R. & S. Uchida, eds. (1998) Relevance Theory: Applications and Implications. John Benjamins: Amsterdam. Clark, E. & H. Clark (1979) When nouns surface as verbs. Language 55: 767-781. Clark, H. & R. Gerrig (1983) Understanding old words with new meanings. Journal of Verbal Learning and Verbal Behavior 22: 591-608. Copestake, A. & T. Briscoe (1995) Semi-productive polysemy and sense extension. Journal of Semantics 12: 15-67. Cruse, D. (1986) Lexical Semantics. Cambridge: Cambridge University Press. Davis, S., ed. (1991) Pragmatics: A Reader. Oxford: Oxford University Press. Davis, S. & B. Gillon (forthcoming) Semantics: A Reader. Oxford: Oxford University Press. Ducrot, O. (1972), Dire et ne pas dire. Paris: Hermann. Ducrot, O. (1984) Le dire et le dit. Paris: Minuit. Fauconnier, G. & M. Turner (2002) The Way We Think. Conceptual Blending and the Mind's Hidden Complexities. New York: Basic Books. Fodor, J. (1998). Concepts: Where Cognitive Science Went Wrong. Oxford: Clarendon Press. Franks, B. (1995) Sense generation: a 'quasi-classical' approach to concepts and concept combination. Cognitive Science 19: 441-505. Gibbs, R. (1994) The Poetics of Mind. Cambridge: Cambridge University Press. Glucksberg, S. (2001) Understanding figurative language: From Metaphors to Idioms. Oxford: Oxford University Press. Glucksberg, S, D. Manfredi & M. McGlone (1997) Metaphor comprehension: How metaphors create new categories. In Ward et al. (1997: 327-50). Grice, H. P. (1967) Logic and conversation. William James Lectures, reprinted in Grice 1989. Grice, H. P. (1989) Studies in the Way of Words. Cambridge MA: Harvard University Press. Gross, S. (2001) Essays on Linguistic Context-Sensitivity and its Philosophical Significance. London: Routledge. He, Z. (2003) Notes on Pragmatics. Guangzhou: CIP. Hobbs, J. & P. Martin (1987) Local Pragmatics. Proceedings, International Joint Conference on Artificial Intelligence: 520-23, Milan: Hopper, P. & E. Traugott (1993) Grammaticalization. Cambridge: Cambridge University Press.
Horn, L. (1984) Towards a new taxonomy for pragmatic inference: Q-based and R-based implicature." In Schiffrin (1987: 11-42). Horn, L. (1992) The said and the unsaid. Ohio State University Working Papers in Linguistics (Proceedings of SALT II) 40: 163-92. Horn, L. (2000) From if to iff: conditional perfection as pragmatic strengthening. Journal of Pragmatics 32: 289-326. Horn, L. & G. Ward, eds. (2004) The Handbook of Pragmatics. Oxford: Blackwell. Kittay, E. & A. Lehrer (eds.) (1992) Frames, Fields, and Contrasts: New Essays in Semantic and Lexical Organization. Hillsdale NJ: Erlbaum. Lahav, R. (1989) Against compositionality: the case of adjectives. Philosophical Studies 55: 111- 129. Lakoff, G. & M. Johnson (1980) Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, G. (1987) Women, Fire and Dangerous Things. Chicago: University of Chicago Press. Lascarides, A. & A. Copestake (1998) Pragmatics and word meaning. Journal of Linguistics 34: 387-414. Lasersohn, P. (1999) Pragmatic halos. Language 75: 522-551. Levinson, S. (1983) Pragmatics. Cambridge: Cambridge University Press. Levinson, S. (2000) Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge MA: MIT Press. Lyons, J. (1977) Semantics, volume II. Cambridge: Cambridge University Press. Merlini Barbaresi, L., ed. (2003) Complexity in Language and Text. Pisa: Editore Plus. Neisser, U. ed. (1987) Concepts and Conceptual Development. Cambridge: Cambridge University Press. Noveck, I., M. Bianco & A. Castry (2001). The costs and benefits of metaphor. Metaphor and Symbol 16: 109-21. Papafragou, A. (2002) Modality: Issues in the Semantics­Pragmatics Interface. Amsterdam: Elsevier Science. Recanati, F. (1995) The alleged priority of literal interpretation. Cognitive Science 19: 207-32. Recanati, F. (2004) Literal Meanings. Cambridge: Cambridge University Press. Rips, L. (1995) The current status of research on concept combination. Mind & Language 10: 72- 104. Rubio Fernandez, P. (2001) Inhibition of core features in metaphor interpretation. RCEAL Working Papers, Cambridge, Research Centre in English and Applied Linguistics. Ruiz de Mendoza, F. (1998) On the nature of blending as a cognitive phenomenon. Journal of Pragmatics 30: 259-274. Schiffrin, D., ed. (1987) Meaning, Form and Use in Context: Linguistic Applications (GURT 1984). Washington DC: Georgetown University Press. Searle, J. (1979) Literal meaning. Reprinted in Searle (1993: 117-136). Searle, J. (1993) Expression and Meaning. Cambridge: Cambridge University Press. Sperber, D. (1989) Lecture on Relevance and Word Meaning. Relevance Theory Workshop, University of Essex. Sperber, D. (1996) Explaining Culture: A Naturalistic Approach. Oxford: Blackwell. Sperber, D. (1997) Intuitive and reflective beliefs. Mind and Language 12: 67-83. Sperber, D. (2000a) Metarepresentations in an evolutionary perspective. In Sperber (2000b: 117- 37). Sperber, D., ed. (2000b) Metarepresentations: An Interdisciplinary Perspective. New York: Oxford University Press. Sperber, D. & D. Wilson (1983) Draft of Relevance, chapter 8, ms. Department of Linguistics, University College London.
Sperber, D. & D. Wilson (1985) Loose talk. Proceedings of the Aristotelian Society LXXXVI: 153-71. Reprinted in Davis (1991: 540-49) Sperber, D. & D. Wilson (1986/95) Relevance: Communication and Cognition. Oxford: Blackwell (second edition with Postface, 1995). Sperber, D. & D. Wilson (1995) Postface to the second edition of Relevance. Oxford: Blackwell. Sperber, D. & D. Wilson (1998) The mapping between the mental and the public lexicon. In Carruthers & Boucher (1998: 184-200). Sweetser, E. (1990) From Etymology to Pragmatics. Cambridge: Cambridge University Press. Verschueren, J., J-O. Ostman, J. Blommaert & C. Bulcaen, eds. (1997) Handbook of Pragmatics Online. Amsterdam: John Benjamins. Vega Moreno, R. (2001) Representing and processing idioms. UCL Working Papers in Linguistics 13: 73-107. Vega Moreno, R. (2003) Relevance Theory and the pragmatic construction of idiom meaning. UCL Working Papers in Linguistics 15 Vega Moreno, R. (2004) Metaphor interpretation and emergence. UCL Working Papers in Linguistics 16. Ward, T., S. Smith & J. Vaid, eds. (1997) Creative Thought: An Investigation of conceptual structure and Processes. Washington DC: American Psychological Association. Wilson, D. (1990-2003) Concepts. Philosophy of Language/Issues in Pragmatics lectures, Department of Linguistics, University College London. Wilson, D. (1995) Is there a maxim of truthfulness? UCL Working Papers in Linguistics 7: 197- 212. Wilson, D. & D. Sperber (1998) Pragmatics and time. In Carston & Uchida (1998: 1-22). Wilson, D. & D. Sperber (2002) Truthfulness and relevance. Mind 111: 583-632. Wilson, D. & D. Sperber (2004) Relevance theory. In Horn & Ward (2004: 607-632).

File: relevance-and-lexical-pragmatics.pdf
Title: Lexical pragmatics is a rapidly developing branch of linguistics which investigates the processes by which linguistically-specified (`literal') word meanings are modified in use
Author: Deirdre Wilson
Published: Fri Sep 27 12:38:17 2013
Pages: 17
File size: 0.22 Mb


Exposure, 128 pages, 0.53 Mb

Old Testament Survey, 3 pages, 0.11 Mb

Education for What, 33 pages, 0.45 Mb
Copyright © 2018 doc.uments.com