Searching for Some Relevance Answers to the Problems Raised by the Translation of Irony M a

Even though the approaches aiid definitions of irony have not been historically homogeneous, it has generally been argued that it implies a certain discrepancy in meaning, be it between what is said and what is meant, or between attitudes such as blame and praise, to refer only to some of the most recurrent tendencies. Moreover, recent literary and pragmatic views on the interpretation of irony seem to have agreed upon the role played by inference, thus stressing the fact that traditional models of communication fall shortto account for the dexterity of possible meanings that may be conveyed by such a proteic resource. It may be said that the inferentially based relevance approach to communication has been offering fruitful insights into the understanding of irony. Thus, the relevance studies on irony go back in time more than twenty years so far, and new proposals keep on being put forward (1978-). However, perhaps not so many suggestions have been made on the problems and possible recurrent traits in the translation of irony. The present paper sets out to explore some of the problems that have been traced in the translation of irony. The relevance analyses of irony and of translation will then be sketched, with a view to testing whether the relevance proposals on communication and translation can shed some light upon these issues..


Introduction
As this paper attempts to analyse the proposals on the translation of irony that may be contributed by relevance theory, the discussion will be stractured as follows. To start with, the most significant problems that contemporary authors have confronted in the translation of irony will be outlined. A synthesis of the relevance approaches to irony and to translation will follow, with a view to revising those aspects of the theory which could shed light on this issue. The main focus of the paper will then be the development of certain proposals on the basis of the relevance approach to communication as an ostensive-inferential process which could improve the present state-of the-art of the issue. These will be then applied to the analysis of existing translations of representative contemporary works of English literature.

Problems in the translation of irony
The contemporary insights into the translation of resources such as humour or irony go back to the very beginning of the present century, and the earliest proposals, with authors such as Henri Bergson (1900), tend to focus upon the issue from a cultural point of view. His approach is also remarkable for the links established between irony and humour, both from an intra-linguistic and an inter-linguistic level. The complexity of the latter does not only lie in a certain discrepancy between the meaning expressed and the meaning intended, but also in the fact that irony and humour are culturally-bound. For him, a problem common to both the translation of irony and of humour spricgs as a result of their cióse attachment to the language, culture and world view of the societies that produce them: "Many comic effects are incapable of translation from one language to another, because they refer to the customs and ideas of a particular social group" (1900: 64-65).
In his view, the "translatability" of resources such as these depends on whether the comic element lies either in situations or in words. As for the latter, a further distinction can be made between the comic expressed and the comic created by language. It is the comic which is created by language which will pose more problems to the translator. If in the case of the former it may be conveyed in the target language despite the loss of certain aspects, in the case of the comic created by language, to render it in the target language may become impossible, because, Bergson says, "it owes its entire being to the structure of the sentence or to the choice of the words", and " it is language itself that becomes comic " (1900:128).
With regard to this, it is also interesting to note that in those cases in which humour and irony may be translatable, that is, if they are just expressed by language, Bergson approaches their translation in terms of "costs", which are connected to the losses that may be produced when airaing at the effects that may be achieved. Thus, he claims that "The former [the comic expressed by language] could, if necessary, be translated from one language into another, though at the cost of losing the greater portion of its significance" (1900:127). Therefore, the ideas of "balance", and of the translator's need to make choices are already present, albeit in a somehow implicit or weak form.
Already in the nineteen eighties, the problems posed by the translation of irony have sometimes been approached from the perspective of communication, even though sometimes the former has not been defined any further. Thus, in one of the monographic works devoted to irony, Graham Dunstan Martin (1983) draws attention to the importance of the context, and notes how, for irony or humour to be understood, speaker and listeners have to share the context. For this author, and following the Sapir-Whorf hypothesis, different languages pattern the world in different ways. Perhaps motívated partly by this, words have different connotations in difíerent languages, which the translator must be acquainted with. Moreover, as literary language is based upon the possibilities of evocation, meaning will be based on potential significances or connotations which may depart from ordinary language. This will affect the ways in which they can be translated, so as to maintain these connotations: as Martin puts it, "the potentiaüties of words are different in different languages " (1983: 424).
In general terms, the translator acts as a mediator between two different socio-cultural and cognitive environments. As such, s/he tends both to construct a model of the intended meaning of the Source Text (henceforth ST), and to form hypotheses and judgements about the probable ünpact of the ST upon its addressees, with a view to transmittíng similar effects to the Target Text (henceforth TT) audience. During the nineteen eighties, when the influence of text linguistics on translation has been on the increase, and key concepts such as the context have been more accurately defined, Hatim and Masón have claimed that "as a text producer, the translator operates in a different socio-cultural environment, seeking to reproduce his or her interpretation of 'speaker meaning' in such a way as to achieve the intended effects on TT readers" (1990: 92). This shows that those insights into the translation of irony, just as the rest of translation studies, have been influenced by the mainstream of pragmatics, which takes meaning to be intentional, and to be necessarily described in a context whose scope goes far beyond the purely linguistic and needs to be approached from a cognitive perspective.
It is also significant to note that Hatim and Mason's general criterion to decide what is to be included in the translated versión follows a balance between effectiveness and processing effort, which seems to be inspired in Beaugrande and Dressler (1981), and may be said to lie at the core of the origin of the notion of relévame itself as a balance between efforts and effects, as defined by Sperber and Wilson (1986/95): "Thus, the guiding principie for deciding what to include in a text and what to take for granted may be stated as: Is the gain in effectiveness sufficient to warrant the extra processing effort involved?" (Hatim and Masón, 1990: 93).
Indeed, as this quotation shows, the guiding principie that must have oriented the translator's task in the translation of irony comes certainly cióse to Sperber and Wilson's definition of relevance as a balance between processing efforts and communicative or contextual effects. On the other hand, Hatim and Mason's approach to irony is in principie based upon Grice's description of irony as a flouting of the maxim of quality, but incorporates Sperber and Wilson's claim that irony impHes the conveyance of the speaker's attitude towards the proposition expressed. On the whole, their approach tends to reconcile both views: Sperber and Wilson criticise Grice's account of irony because such devices as ironic understatement do not in fact flout the maxim of quality. But we believe Sperber and Wilson's view of "echoic second-degree interpretation" (i.e. echoing an imaginary person's view) is not inconsistent with an essentially Gricean view. Thus, ironic understatement, while it may not flout the maxim of quality, does involve apparent violatíon of the maxim of quantity ("make your contribution as informative as required") (Hatím and Masón, 1990: 98).
Thus, even though Hatim and Masón keep on speaking about irony as a violatíon of one or more maxims, this in turn entails that the speaker dissociates herself from the view expressed, echoes a different point of view and shows her attitude towards it. The recovery of the intended meaning is govemed by inferential processes which allow the addressee to grasp this intended interpretation. However, it may also be commented that such a compromise does not fit in the model of relevance theory, one of whose main guidelines has beenprecisely the rejection of Grice's maxims, for which the principie of relevance is meant to substitute.
The problems caused by irony in translation were also accurately pinned down by Hatim and Masón in their first joint work on translation: "Now, it may occasionally be observed that a translation, while faithfully reflecting the propositional content of the source text, fails to achieve the degree of irony perceptible in the source text" (Hatim and Masón, 1990: 99). Even though the notion of" faithfulness" is controversial in translation theory, the approach to the problem of translation may be regarded as accurate: as irony often goes beyond the words encoded or the propositional content and is to be inferred, if the translator is not aware of the intention or cannot recover it from the context, it may be lost altogether. A possible solution they propose is to further amplify the message expressed in the original, especially when the context cannot be accessed:"... Since TT readers cannot be assumed to share the same cognitive environment as ST readers, the translator may feel the need to provide additional cues for recognition of the ironic intention " (1990: 99).
All this means that the perception of irony may be different whenever the context or the potentialaddressees change, which has obvious consequences for translation. As Hatim and Masón point out:" (...) Successful translation will depend on whether or not TT readers are able to achieve second-degree interpretation with niinimal extra processing effort. Recognition of ironic intention is, in all cases, crucial and will condition the translator's output" (Hatim and Masón, 1990: 100).
In synthesis, what all this entails is that the perception of irony goes beyond the propositional content expressed in the utterance, and requires the inferential recognition of the speaker's communicative intention, which if at all may be traced in a certain context or cognitive environment. This context is by definition variable for relevance theoreticians, and it may be assumed that this is precisely what happens in translation. It may be noted how the sort of terminology employed by Hatim and Masón owes much to relevance theory in such crucial aspects as "second-degree interpretation", "processing effort", or "cognitive environment".
In the case of translation, the translator should attempt to perceive the speaker's meaning and convey the intended message and its corresponding implicatures. In an ironic utterance, what the speaker aims to convey is often not made explicit, but remains implicit, and it must be inferred from what is explicitly stated.
The aspect of the recovery of the meaning intended by means of inferential processes has also concerned literary theorists. Thus, Linda Hutcheon (1994) notes that the difficulty in translating irony lies in the common absence of hints which might bave indicated its existence, and on top of that, another aspect will be to reconstruct the ironic meaning in such a way that is intelligible in the target text for the audience. However, it is also certainly the case that irony is the more enjoyable as it Mis to announce itself.
Some authors nave attempted to contribute certain Solutions that can be adopted in the translation of irony. To give just an example, Marta Mateo (1995) makes a synthesis of the different possibilities available: ... The possible translation solutions for a problem of irony will range from the maintenance of the original irony, being it further explained in a footnote, if unincomprehensible in the target language, on the one hand, to the deletion not only of the original but also of the entire text that contains it. The different strategies will be determined by the norrns that may influence the translator's work at a definite moment, among which the addressee occupies an uppermost position (Mateo, 1995: 68, my translation).
In her study, Katharina Barbe (1995) retakes the view on irony as a cultural phenomenon: "We cannot expect that all cultures nave similar understandings and uses of irony" (1995: 144). She goes so for as to claim that, being irony one of the resources which may either be successfully mastered or else misunderstood by the addressee, it may be approached as one of the mechanisms at the disposal of a given culture "to keep others from understandktg" (1995: 146).
For this author, the problems of the translation of irony do not only nave to do with the complexity of meanings that may be expressed and their conveyance in the target language, but also with the assumptions entertained about translation. In her view, the classical dichotomy made between literal mdfree translation only amounts to creating further problems in the translation of irony: "Finally, I revive the problems of dichotomies, exemplified here byfree and literal translation. I establish that this división, even though methodologically justifiable, appears particularly problematic when irony is translated" (1995: 146).
The possibility of the translation of irony is linked by Barbe to the general principies of translatability between the two languages concerned. These principies refer to the contact between the source and target languages, the general evolution of both cultures, as well as the kind of information presented in the text.
Rather than focus upon the classical dichotomy made between literal and free translation, which as shown above she rejects, Barbe poses the problem on the kinds of irony to be distinguished, mainly nonce irony and common irony. Whereas common irony refers to those expressions which tend to trigger an ironic interpretation, even if found out of context and which can give way to fossilized expressions, nonce irony refers to "those instances of irony which that have not habitually been used for ironic purposes and subsequently lost their original status " (1995:148).
In the case of common irony, the translator would aim to find an equivalent idiomatic expression that may be endowed with similar connotations in the target language.
As for nonce irony, its translation tends to present more challenging problems. The basis for determining whether something is meant to be understood as ironic or not will tend to rely heavily upon the background knowledge shared by the speaker and the audience. On the other hand, background knowledge also has to be considered in the case of common irony.
She has also referred to the different translation strategies which may be adopted in the translation of irony. Her proposal is based upon the analogy with the solutions provided by different authors for the translation of metaphors. She refers to the options such as the following, available for the translation of irony: reproduce the same irony; replace the SLimage with an appropriate TL-image; add a potential irony marker (usually a modal marker); add a description; or keep SL-idiosyncracy (1995:166-167). What she refuses ís the omission of irony as a translational option: in any case, irony has to be present in the TT; otherwise, the message is misrepresented.
For her, the translatability of ironic utterances depends on certain conditions of the relationship between SL and TL: whether participants share or are aware of each others'cultural knowledge; whether both texts use similar linguistic realizations of irony; if SL and TL share similar face-saving strategies, or they use irony for similar purposes; or finally whether both cultures nave comparable institutional organizations. (1995: 167) Even though relevance theory has not provided any explicit account of the problems of the translation of irony, and less of the ways to handle them in their theory, their general approach to translation focuses upon the problem of the specificity to each language: Returning to translation, the obvious problem is that in translation we need to talk about resemblances between texts and utterances that belong to different languages. While there is a certain consensus that it is often possible to achieve a fairly good degree of resemblance in semantic representation across languages, the same cannotbe said of stylisticproperties, which often consist in linguistic features that are far from universal (Gutt, 1991: 126-7, italics as in the original).
Resemblance seems to be then a key concept to account for translation from the standpoint of relevance theory, as well as it is for irony. The approach to style is also significant: in this framework, it is seen as imposing certain constraints and showing indications on the way utterances are to be understood: ... One might well argüe that the point of preserving stylistic properties lies not in their intrinsic valué, but rather in the fact that they provide clues that guide the audience to the interpretation intended by the communicator. We shall refer to such clues as communicative clues (Gutt, 1991:127, italics as in the original).
Gutt also stresses the importance of context:"... the meaning communicated by a text is not attributable to the stimulus alone, but results from the interaction between stimulus and cognitive environment" (1991:132). Likewise, the meaning to be grasped from a certain text or discourse goes beyond the semantic representatíons of the utterances. The context is also necessary to arrive at the intended meaning: (...) The semantic representation of an utterance forms an assumption schema that needs to be developed inferentially until it yields the propositional form of the utterance that can be evaluated as true or false of some state of affairs. Thus the usefulness of these semantic representatíons is that they serve as a "source of hypotheses" about the communicator's intention -that is, they provide communicative clues (Gutt, 1991: 131).
The communicative clues as distinguished by tbis author guide the reader towards the interpretation signalled by the speaker: "Communicative clues are properties built into the text to guide the audience to the intended interpretation" (Gutt, 2000b: 155). Henee, it may be assumed that the recognition of a certain utterance as ironic may be guided by the identifícation of certain communicative clues that signal the meaning intended by the speaker.
These communicative clues are to be handled with special care by the translator. Gutt himself makes a distinction in the role played by them either in intra-lingual or inter-lingual communication: If one were looking only at intra-lingual communication, one could simply say that communicative clues are a subset of the textual properties that are significant for the intended meaning. There would not be any difference in essence between a textual property and a communicative clue.
However, the situation changes when considering cross-lingual communication, and this is where it seemed helpful to form a more abstract concept than a textual property. The reason is that languages differ in the inventory of linguistic features or properties they have (Gutt, 2000b: 153).
In a certain sense, the expression of an ironic attitude would alter the balance effort-effect, in the sense that it could possibly require an extra effort on the addressee, who would tend to expect to obtain further pleasure; that is, as Gutt claims, and as can be derived from the application of the theory itself, the additional processing effort should be outweighed by the achievement of extra effeets: Put in general terms: if a communicator uses a stimulus that manifestly requires more processing effort than some other stimulus equally available to him, the hearer can expect that the benefits of this stimulus will outweigh the increase in processing cost -otherwise the communicator would have failed to achieve optimal relevance (Gutt, 1991: 141).
If translation is a form of communication, however peculiar, it follows that the principies of communication as described by relevance theoreticians should also hold applicable.
So as to successfully transíate irony, it must be firstly spotted and tdentified in the original. Being a cultural mediator, the translator undertakes to recognise the irony present in the original, to identify the meaning intended by the speaker and to convey it in the target language. This may be guided by aspects such as: -the irony should be conveyed in such a way that no much ftirther explicitation than in the original is required. -the reader or addressee of the target text should be able to reach similar effects te those of the source text audience without any further demand of processing efforts.
Basil Hatim (1997) focuses upon the problems to be found in the translation of irony from a discourse and semiotic perspective. On the basis of the primacy of semiotics avowed by the author, it is argued that the study of irony should not be constrained to literary texts, and, as well as Sperber and Wilson have already proposed in the first edition oí Relévame (1986), the break of the dichotomy between literary and non-literary language, which should be best dealt with in terais of a continuum, is also proposed.
He claims that irony can be found at the three different semiotic categories distinguished by Ian Masón and Hatim himself in their 1997 joint work, The Translator as Communicator, -namely, text, genre and discourse. These three socio-cultural units were actually already introduced in their 1990 work, but they become of prime importance in their new proposals for translation, being the basic criteria against which translation equivalence will be assessed. The text refers to the speaker's rhetorical purpose, and to its structure; it is defmed as "a set of mutually relevant communicative functions that hang together (> texture) and are constructed (> structure) in such a way as to respond to a particular context and thus achieve an overall rhetoric purpose" (1997:224, bold as in the original). Genres are the ways in which linguistic conventions cater for or respond for particular social occasions, and are defined as "conventional forms of texts associated with particular types of social occasion" (1997:218). Finally, the notion of discourse embraces all those aspects connected with the ideology expressed in a definite text, they embody a certain attitudinal expression, and are defined as "modes of speaking and writing which involve social groups in adopting a particular attitude towards áreas of socio-cultural activity" (1997: 215).
For Hatim, the problems raised by the translation of irony are based on the contrast existing between meanings which are universal, on the one hand, and their expression, which will vary in the different languages, on the other hand: The probiem for the translator arises whendifferentpragmatic and institutional-communicative procedures are empíoyed by different languages in the expression of almost universally recognized attitudinal valúes which are essentially discoursal and semiotic. That is, while almost all languages have at their disposal the potential ultimately to relay, say, a disparaging attitude (a semiotic concern), what constitutes this in terms of rules of politeness, types of implicature, etc. (which are pragmatic concerns) and those of register appropriateness, level of formality, etc. (institutional-communicative concerns) can and does vary from one language to another (Hatim, 1997:187). Therefore, whereas the meanings and attitudes conveyed when speaking ironically are perhaps universal, each language may make use of different strategies and resources to express them. Henee, the translator must be trained to recognise and deal with all those cross-cultural differences which can be found in texts for the expression of otherwise universal beliefs, ideas, attitudes or feelings. The question he seeks to answer is, going irony often beyond what is expressed into words, whether the appreciation of irony can be made more systematic, and his approach will aim to synthesise what the most significant theories have said about irony; it is therefore broad and eclectic.
Some of the views held on irony by this author regarding the critical stances taken against the traditional accounts of irony agree with Sperber and Wilson: in their view, these theories have failed to determine what is to be understood by "figurative meaning", how it is to be derived from its literal meaning counterpart, and to explain why, and in what circumstances, speakers should prefer ironical utterances instead of their literal counterparts. To do so, a different conception of communication and of meaning is altogether necessary, and for Hatim, in contrast to Sperber and Wilson, Grice will be the first step towards a new explanation. For Grice irony is explained as a violation of the first Maxim of Quality, which states, namely, "Do not say what you beheve to be false". Hatim will refer to Sperber and Wilson's criticism that Grice's approach, just in the same way as it happened with traditional theories, fails to explain why a speaker should prefer an ironical utterance to its literal counterpart. However, he will come to question whether Sperber and Wilson's approach comes to shed any light anyhow. His view will be that these authors' view is not really in open conflict with Grice's proposals, but rather come to enhance them, and he will draw attention to the contributions made by Sperber and Wilson which can be applied to account for the semiotic status of irony, which he upholds.
Besides referring to the distinction drawn between use and mention by Sperber and Wilson in the initial papers where they deal with irony (1978,1981), Hatim refers to another one made by the Relevance authors, which has often gone unnoticed. This is the difference established between the so-called first-order interpretation and second-order interpretation. Irony would be an instance of the latter, which is characterised by the fact that "the thought of the speaker which is interpreted by the utterance (...) is itself an interpretation" Wilson, 1986: 238, apudUaúm, 1997: 194).
For Hatim, moreover, these attitudinal meanings that characterise irony (as not only Grice and Sperber and Wilson have noted, but also Leech, Levinson or politeness theorists) can be approached as semiotic cotegories: by this he means that they are concemed with the interaction between different texts, and of the users with these texts, which are then endowed with a certain intertextual potentidl. What is echoed may be so from other different texts (intertextuality), from other aspeets or fragments of the same text (intratextuality), or from instances of either which are questioned (contratextuality). For Hatim, an echo -in Sperber and Wilson's ternas-may then be of any of these three different kinds: intertextual, intratextual, or contratextual, depending on its source and the rektionship that it maintains with it, which is usually an attitude -which constitutes the sign-held by somebody else than the speaker:" But the more important aspect of semiotic meaning is that which reveáis the speaker's attitude to the opinión echoed". (1997:194-95). Hatim also coincides witli Sperber and Wilson in that the attítude shown by ironical utterances is usually one of rejection or disapprovaí.
However, tiie approach followed by him to the expression and translation of irony will be essentially Gricean, as it is based upon the floutíng of the maxims. For him, both Grice's and Sperber and Wilson's theories are complementary rather than mutually excluding. He will propose, moreover, that irony is not necessarily a floutíng of the Maxim of Quality, but it may also be a flouting of the Maxim of Quantíty. He claims that this will be subject to cross-cultural differences, that is, a given culture may prefer to express irony by means of a certain maxim, whereas a different one may make use of different resources. He will apply this hypothesis to what happens between Engüsh and Arabic, but, as we shall see next, some general conclusions can be reached: For socio-cultural and linguistic reasons (...), English has developed a particular preference for understatement and the cryptic. Irony seems to be one of those aspects of verbal behaviour which benefit considerably from this kind of attítude to language use. In the context, the speaker can leave so much unsaid, yet express the attítude in question. But what is unsaid by no means leaves the utterance incomplete; on the contrary, the utterance will be 'pregnant' with meaning as a result. That is, English seems to allow some economy with the truth (Hatim, 1997: 196).
These assumptions confirm the view that irony goes beyond the words uttered, and that so as to adequately grasp it, it is necessary to infer the speaker's attítude and communicative intention, which are viewed by Hatim as sign systems or semiotic categories. This will be the translator's first step, previous to bis/ her conveyance of the text in the target language. The translator needs to understand the irony present in the source text, and convey it in the target text so that it can be grasped by the target audience, with no extra processing effort, on the one hand, but at the same time, with not any greater explicitness which might spoil the whole effect, due to the fect that the charm of irony relies precisely in that it goes apparently unnoticed, without calling any attention to itself.
The general conclusión reached by Hatim reads as follows: translation should aim at preserving pragmatic or communicative functíons, as they show those meanings which have been intended by speakers. In any case, so as to preserve the same semiotic meaning, the same attítude present in the text, translation may make use of different pragmatic functíons. It seems, then, that Hatim approaches attítude as a semiotic, rather than a pragmatic, construct: ... The hypothesis entertained in this study may be phrased as follows. Translation is a signfbr-sign act of transfer which does not necessary entail the need to preserve the very same pragmatic or register valúes of the source text. Put slightly differently, translation is not necessarily the transposition of a given field, mode or tenor by identical register valúes, ñor is it necessarily the transposition of a given speech act, implicature, etc. by an identical pragmatic manifestation (Hatim, 1997: 197).
In any case, tfae translator may feel it necessary to approach the meanings intended by speakers in the source text, (semiotic analysis of the signs employed), and grasp what the meaning intended might be (pragmatic dimensión). Perhaps both the semiotic analysis and the pragmatic analysis are best seen as complementary.
The important thing is that, in whichever forms the translator chooses to use, the target text may achieve to communicate those aspects of meaning -and irony among them-which are crucial to understand the source text and which should be equally crucial to grasp the meaning of the target text.
At the root of the different ways to convey irony -and any other meanings-in different ways by different languages, cultural reasons and conventions are to be found. Moreover, whenever there exists a time span between the source text and its translation, diachronic changes will also be necessarily taken into account: Rhetorical moves might change as time goes by. Diachronic criteria are helpful in attempting to account for the way languages evolve in dealing with aspects of use such as irony. English prose of the nineteenth century, for example, shows unmistakable tendencies to flout Quantity and not Quality in relaying irony, a tendency which seems to have been on the wane in modern English (Hatim, 1997: 198-99).
Hatim also offers some practical considerations when dealing with irony and its translation: the first thing a translator should do is to be able to appreciate the irony present in the source text. Then, the translator should try to convey the communicative meaning intended by the source text author, and to avoid opting for "some literal rendering that simply defies the ultímate objective of a given text in this regard " (1997: 195). That is, just as irony often goes beyond the literal words avant la lettre of the text, so the translator will have to read between the lines to perceive the speaker's communicative intention.
Hatim's analysis also draws attention to the fact that the ironic meaning intended may go beyond the propositional contení expressed by words. Among the linguistic devices that can be used to convey irony, he refers to the following: lexical repetition, modality, text type, textual devices such as ellipsis or conjunctions, parallelism, or the textual informativity of the unexpected.
The translator's role is twofold: s/he is both a reader of the source text and a writer of the target text. As such, in the case of irony, his role is first to appreciate the irony of the source text and convey it in the target language. As Hatim claims, "...preserving irony becomes a problem not only of reception but of production too" (1997:196).
fn synthesis, the main problems presented by the translation of irony to be faced by the translator that the analysis has shown may be summarised as follows: if it is admitted that irony may be translatable at all, first, the conveyance of irony may be linked to the linguistic cultural aspects of the text, which have to be transposed or adapted to the new cultural and linguistic environment; as a result of this, the context has to be approached in wide cognitive and cultural terms; second, irony may reíy on the linguistic configuration of the text (on words), or rnay go beyond it; third, as irony is the more enjoyable the less it announces itself, the translator should ideally fínd a balance between the processing effort demanded on the target text addressee and the effects to be achieved.
In what follows, I aim to revise the main proposals on irony and on translation advanced by relevance theory, with a view to proposing some possible ways in which it can cope with these problems.

Towards the definition of the relevance approach to irony and translation
The proposals on how to cope with irony from the standpoint of relevance theory are perhaps one of the most widely discussed issues in this framework, and date back to 197 8. Then, in a joint paper by Sperber and Wilson, "Les ironies comme mentions", which also appeared in an English versión, "Irony and the Use-Mention Distinction" (1981), classical definitions of irony as "meaning the opposite of what is said" or "something different from what is said" are discarded. Their criticism of traditional models of rhetoric goes further beyond these notions of irony, since they come to reject the notion of figurative meaning altogether, a standpoint that is also taken by Fish (1989) more than two decades afterwards.' It may be remembered here that the reason why Fish rejects this notion has to do with the relativisation of what is a literal and a figurative versión, since both of them are equally interpretations. In turn, Sperber and Wilson's main argument has to do with the fact that the potential choice between a figurative expression or its literal counterpart is left unexplained by traditional theories. 2 In their view, this also holds trae for other pragmatic approaches, mainly Grice's, as the only thing he does in their opinión is to substitute the notion of implicature for the traditional dichotomy referred to above.
If there is a feature of pragmatic theories which is retaken by Sperber and Wilson is the link of irony with the expression of a certain attitude, usually of rejection and disapproval: Furthermore, the crucial íact that ironical utterances convey not only propositions (which can be accounted for in terms of meaning and implicature), but also vaguer suggestions of images and attitudes, fmds a natural description in the framework we propose (Sperber and Wilson, 1981:296-7).
The basic pillars of their theory of irony are also established now. These may be said to be related to the following three notions: use, mention and echo. The concepts of use and mention are defined as follows: "USE of an expression involves reference to what the expression refers to; MENTION involves reference to the expression itself (1981: 303, capitals as in the original). As for the concept of echo, it is closely related to the above mentioned feature shared by most of contemporary pragmatic approaches: the connection of irony to a certain attitude. It is defined as follows:".. .meant to indicate that the preceding utterance has been heard and understood, and to express the hearer's immediate reaction toit" (1981: 306).
However, it is in their seminal work, Relevance (1986/95), as well as in their 1992 (1989) paper "On verbal irony" that perhaps a certain connection between their approaches to irony and to translation may be found. In Relevance echoic interpretatíons are approached as second-degree interpretatíons: "... an utterance used as an interpretatíon of someone else's thought is always, in the first place, an interpretatíon of one's understanding of that otherperson 'sthought" (1995:238). Afurtherstepistakeninthepaper "On verbal irony", where they suggest replacing the notion of mention (Sperber and Wilson, 1981), by that of interpretive resemhlance (Sperber and Wilson, 1986), to account for irony. They claim that the approach proposed in their earlier paper was too restrictive: "It was too restrictive because certain types of irony do not fit in the analysis of irony as echoic mention proposed in 1981" (1989: 96). In synthesis, the former approach to irony as echoic mention is now seen to be a particular instance of interpretive resemblance, which Blakemore defines as follows:"... An utterance intended as an interpretatíon does not have to be a representatíon of what someone has said. It could be an interpretatíon of someone else's thoughts or opinions" (1992:107). Besides, itmay also be noted that when Sperber and Wilson claimed that irony may be characterised by "looser forms of resemblance" (1992:66), these can be associated with interpretive resemblance.
This form of analysis also allows them to intégrate the study of irony into the general framework of relevance as a balance between cognitive efforts and contextual effects. The indeterminacy which characterises weak implicatures and forms of interpretive resemblance is also retaken by Blakemore to characterise irony: What is the point of producing an utterance which is an interpretative representatíon of another person's thought? In some cases the relevance of an echoic utterance simply lies in the information it gives about the content of an attributed thought. (...) However, in other cases the relevance of an echoic utterance lies in the information it gives about the speaker's attitude towards the attributed thought. This attitude may be one of endorsement. (...) Alternatively, an echoic utterance may convey an attitude of rejection (Blakemore, 1992: 167).
Likewise, translation is also analysed in terms of interpretive resemblance in the framework ofthe theory by Gutt 2000aGutt (1991. The departíng point for his analysis seems to be House's 1981 distínction between covert and overt translations, which is based on whether the translation can" enj oy the status of an original source text in the target culture" (1981( :194, apud Gutt, 1991. Such a difference is what Gutt sets out to account for in terms ofthe distínction between descriptive and interpretive use. An intermediate step is the approach to translation as a process of ostensive-inferentíal communicatíon which may unfold under two different communicatíon situations, primary and secondary, depending on whether the audience can make use ofthe contextual assumptíons envisaged by the speaker or not. Thus, one of his basic aims is to test the validity of relevance for a general theory of translation which can account for both primary and secondary situations. The basis for doing so is to approach translation as interlingual interpretive use, by which it is basically meant that "the translation is presented in virtue of its resemblance with the original in relevant respects" (1991:112), which also means that the translator is to make assumptíons about the cognitive environment of the audience "and about the potential relevance that any aspects of the interpretatíon would have in that cognitive environment" (1991:110).
Nonetheless, for Mateo Martínez (1998), such a distinction should not be read as the poles of a mutually exclusive dichotomy: (...) I believe that we cannot trace clear limits between what Gutt defines as descriptive or "incidental" (1990:142) (covert) translation and interpretive (real) translation (1990,1991). I would rather think of a translating continuum that ranges from translations that, originating in an SL idea, make a TT thatbears little resemblance to the original ST but which manages to convey a similar presumption of relevance to texts that keep all ST assumptions but fail to produce similar cognitive efforts in the TT readers because they maintain an excessive literal format. (...) Within this framework, there can be a flow of communication where translated texts can be sometimes descriptive and at other points interpretive depending on their readers' expectations but without marking off clear boundaries between the two (Mateo Martínez, 1998:177).
Indeed, it appears that the conclusions ultimately reached by Gutt in bis study point in the same direction: after having established a working dichotomy between direct and indirect translation and the former being precisely based on the preservation of the stylistic features or communicative clues of the original, direct translation is seen as a special case of interpretive use, which becomes the general framework to account for translation in the theory.
Some authors have also attempted to complete the relevance proposals on irony by applying certain tenets of the theory. This is the case of Yus Ramos (1997-98), who applies the relevance balance between efforts and effects as well as the role to be played by the context, and refers to the interpretation of irony as a balance between context accessibility and processing effort. 3 His approach is therefore cognitive and relevance-oriented. His most remarkable contribution has three different aspects: first, the tracing of different contextual sources, whose activation by the addressee is crucial for his understanding of irony, on condition that there are certain incompatibilities which may make him actívate an ironic interpretation. The contextual sources distinguished by Yus Ramos to which the addressee will attempt to have access to for the interpretation of ironic utterances are the following: encyclopaedic factual information; mutually manifestphysical context or environment; the speaker's non-verbal behaviour; the addressee's background knowledge of the addresser's biographic data; mutual knowledge; previous utterances of the conversation; and linguistic cues.
The second aspect has to do with the formulation of a criterion ofoptimal accessibility to irony, which springs as a consequence of the application of these contextual resources, and which goes as follows: The processing effort required for the interpretation of the intended ironic meaning of an utterance decreases in proportion to the increase in the number (and quality) of incompatibilities (detected by the addressee) between the information supplied by the inferential integration of simultaneously activated contextual sources (leading or leading plus supportive) and the informatiotí provided by the propositíon expressed by the utterance (Yus Ramos, 2000a: 12).
For Yus Ramos himself, there is a direct application of the criterion ofoptimal accessibility to irony, quoted above, and its role in the processing of irony, to Gutt's translational model. According to this model, the translation of a given text should be guided by the balance between the effort required to interpret the target text and the positive cognitive effects that the individual may achieve. Moreover, at least ideally, such a balance should become as cióse as possible to the balance between processing efforts and cognitive and contextual effects achieved by the ST reader. Therefore, a basic criterion to be followed by the translator is the relevance balance between efforts and effects, such that the TT reader should be able to acquire very similar effects to those grasped by the ST audience, without being required any further efforts. At the same time, this criterion may be applied to translation and to translation assessment in the sense that the target text should be intended to achieve similar effects to those of the original, without requiring any further processing efforts. It can also justify any departures from the literal or propositional translation of the original text, as long as a similar effect is achieved. As we shall see below, this lies at the core of the interpretive resemblance that translation should aún at. This may be illustrated with the following example, taken from an advertisement of a fully-equipped Hi-Fi set: (Mateo Martínez, 1998: 179): "Listof features as long as your arm" (...) "Butnotan arm and aleg". This was adapted as "Abra bien los ojos y descubra todas sus prestaciones" (...) "Pero sin costarle un ojo de la cara".
The achievement of a similar balance between the effects achieved by the pun or word play of the original text and the efforts required on the addressee may be said to lie at the core of the translation above. Thus, the effect pursued in the ST may be said to rely on the effect of surprise, in such a way that an initial propositíon, "List of features as long as your arm" is enriched by the following one, which tends to contradict or cancel some of the assumptions that might have been entertained by the audience: "But not an arm and a leg". The translation also relies on very similar resources to those of the original: here, the pun or wordplay "Abra bien los ojos (...) Pero sin costarle un ojo de la cara" combines idiomatic expressions, culturally grounded, and thus easily accessible or retrievable to the audience on the basis of sbared knowledge, in which, similarly to what happened in the original, the second one caneéis at least part of what the audience might have initially thought.
The contrast between the interlinguistically different encoded propositional forms, on the one hand, and the similar assumptions that they are intended to créate in each of the source and the target audiences respectively, on the other hand, can be accounted for in the framework of relevance theory as a combination of both descriptive and interpretive resemblance based translation criteria, according to Mateo Martínez: The translator has translated the text in an interpretive manner: similar linguistic uses convey the intended idea and créate the same contextual effects and the same inferential processes in both ST and TT readers. And he has also used a descriptive approach by adapting language and intentions to the recipients' expectations (Mateo Martínez, 1998: 179, my italics).
A very similar case can also be found in the following reference taken from a passage from Fowles' The French Lieutenant 's Woman, which will be analysed with greater detail below. Thus, an idiomatic expression such as "The programme was unrelievedly religious", (ST: 111) has been conveyed in the target text by an idiomatic counterpart which in Spanish may be said to give way to similar contextual implications, "El programa era sacro de cabo a rabo" (TT: 126). As in the case of the example described by Mateo Martínez (1998), the translation may be said to have combined resources which point at both descriptive and interpretive resemblance, in that the translator has searched to créate in the target language (henceforth, TL) similar effects to those found in the source language (henceforth, SL), on the basis of the adaptation of language and intentions to the potential TT readers. More generally, it may be hypothesised that such a combination of both kinds of resemblance as described by relevance theory may be particularly useful in the translation of idiomatic constructions, which for the creation of similar contextual effects to those intended by the SL may have to be rephrased into cultural and linguistic equivalent counterparts in the TL. From a relevance point of view, it may be concluded that the above commented search for equivalence may be taken to be relevance oriented in several ways: first, in what concerns the balance between processing efforts and cognitive effects; second, in the notion of accessibility, as described by the Yus Ramos' concept described above: one of the clues that the translator may take as a reference when translating irony, so that it can be adequately grasped by the TT readership, is to hypothesise about the accessibility that they may have to the contextual sources of irony and their possible incompatibilities, and compare them to those of the ST audience. Such a hypothesis would be a guiding principie to decide the possible changes that may be introduced in the TT (Yus Ramos, personal communication).
On the whole, it may be concluded that these examples have shown that Yus' cñteñon ofoptimal accessibility to irony may be a guiding principie for the translator to decide the expression of irony in the target text, on the basis of the hypotheses s/he may make both with regard to the ST and the TT readership.
Finally, the fact that the fester or slower identifícation of irony depends upon the number of incompatibilities detected in the contextual sources available leads Yus Ramos to propose a terminological distinction between the bypassed 'proposition expressed' and the entertained 'proposition expressed'. This is introduced in relation to the question of the role which literal meaning is likely to play in the interpretation of ironic utterances. The proposition expressed may be either bypassed or entertained, depending on whether the speaker's intended interpretation remains implicit and inferable from the context, or else the absence of adequate contextual information does not give way to possible implicatures, and the proposition expressed is processed as the speaker's intended interpretation.
Thus, on the whole, the sketch of the main trends in the analysis of irony and translation from a relevance perspective has shown interesting recurrent tendencies, which may be said to spring from the consistent relevance approach to communicatíon as an ostensiveinferentíal process. In what follows, some of the relevance proposals here outlined will be further explored, with a view to determining the contributions that they may bring forward to the solution of certain problems that the translation of irony confronts the translator with.
3. Some proposals on how to cope with the translation of irony from a relevance perspective The proposals to be dealt with here aim to suggest ways to cope with some of the problems created by the translation of irony: namely, the possible translatability of irony; the importance of context, here approached as a cognitive entity; the different, specific and linguistically and culturally based (pragmatic) connotations that may be expressed in the source and in the target language to refer to what may be taken to be (semiotic) universal meanings; or the problems created by the different levéis of explicitation that may be found in either the source or the target text.
The suggestions proposed below are related to certain basic aspects of the theory: first, the communicatíon of irony and its translation must tend to follow a balance between efforts and effects; second, it will also tend to be governed by the ostensive-inferential nature of such a process, and in connection to this the possible relationships between code and inference must also be explored; third, the way how the notions of interpretive and descriptive resemblance apply wül also be dealt with; fourth, the role played by the context as a cognitive entity has to be taken into account to understand the process of communicatíon itself.

The translation of irony as a balance between efforts and effects
The reader familiar with relevance theory need not be remembered of the importance in communicatíon of the balance between these two terms, efforts and effects, which lies at the very core of the notion of relevance itself. Likewise, translation may be approached as communicatíon, even though it presents peculiar features, as will be shown below. Gutt himself draws attention to this basic fect when he claims that "the application of relevance theory entails that translation is being looked at as part of communicatíon" (1991:21). As such, it is an ostensive-inferential process. However, not only two basic participants, addresser and addressee, take part. There is obviously a third partícipant, who is both addresser and addressee, and this is, no doubt, the translator, who acts basically as a reader or interpreter (addressee) of the source text and a writer of the target text (addresser). In principie, his cognitive environment embraces both texts as well as the cultures where they have emerged, and it is his task to enable communicatíon between the two. In the modeí of communicatíon put forward by relevance theory, his role is not limited to the encoding in the target language of the meanings codified in the source language. From the point of view of the theory followed here, he must aim at the inferential recognition of the meaning being ostensively intended by the speaker in the source language, and at the conveyance of these effects in the target language in such a way that the target text addressee can trace and achieve similar communicative effects without being asked any expenditure of further cognitive efforts. A further implication of this is the relationship between explicatures and implicatures. For Gutt, a translation, so as to communicate the same interpretation as that intended in the original,"... should convey to the receptors all and only those explicatures and implicatures that the original was intended to convey" (Gutt, 1991:94, italics as in the original), since "the intended interpretation of anutterance consists of its explicatures and/ or implicatures" (1991:94). He claims that any further "explications" which are not made explicit in the original and that the translator decides to include "are motivated by the assumption that certain implicatures of the original are highly relevant to the audience, but cannot be derived by them from Üie semantic contents alone, due to contextual differences" (1991:115). As a result, the translator may decide to communicate as explicatures certain items of information that might have been just implicated in the source text. In this sense, Gutt himself seems to favour a weaker reading, which entails that it is on the basis of the whole of the information that may be explicit and implicit in either the source or the target text that the balance of information provided by both texts is to be assessed. Furthermore, in an application of the balance of explicit and implicit communication to the explanation of misunderstandings, Yus Ramos (1999) has proposed two different continua, the econtinuum and the i-continuum, each of which comprise different degrees of implicitness, instead of a single continuum ranging from the explicit to the implicit.
In an ironic utterance, what the speaker aims to communicate is not usually made explicit, but remains implicit, to be inferred from what is explicitly stated. On the whole, it may be hypothesised that the target text should attempt to maintain a degree of explicitness similar to that of the original text, since an over-explicitation of irony might spoil the entire effect. In relevance terms, this is accurately accounted for as a balance between the processing effort required by the addressee to understand the utterance and the communicative effect intended to be achieved by the addresser. In the case of translation, it appears that the target text should aim to convey a similar communicative effect, that of distancing oneself from the proposition expressed, without making this communicative intention too explicit.

Code vs. inference
The balance between implicit and explicit communication outlined above has led to the conclusión that it cannot be approached in terms of a sharp dichotomy, and that the communicator's (and also the translator's) choice to decide what to leave implicit and what to make explicit may have a direct influence upon the way in which the addressee can cope with the message, and how far the latter may be (mis)understood.
The communicators' attempt to understand one another is also highly constrained by the ostensive-inferential processes whereby the speaker makes her communicative intention ostensively manifest 4 to the addressee, one of whose main tasks is precisely to set out to infer this intention. With all this, a new framework for the explanation of communication is required, and supplied by Sperber and Wilson themselves, for whom it does not stop at the addressee's decodification of the message encoded by the speaker. Such a model does not account for the intentions made manifest. Therefore, an extra layer of inference is required. However, as Sperber and Wilson also admit, at least in the case of verbal communication, both inferential and codification/decodification processes are rather complementary than mutually exclusive: "Verbal communication involves both code and inferential mechanisms" (1995: 13). Interestingly enough, this aspect had already been noted by one of the most representative forerunners of the code model of communication, Jakobson, by claiming that "Um ein Zeichen zu interpretieren, kann man sich sowohl auf den Kode ais auf den Kontext, (...) beziehen" (1956: 327). 5 Thus, the question may well be the way in which code and inference interact in the production, interpretation and translation of irony. If, as shown above, the translated text should be geared to maintain a degree of explicitness similar to that of the source text, so that the balance between processing efforts and contextual and cognitive effects to be achieved by the audience should not be substantially altered, then it follows that it is precisely one of the translator's tasks both to perceive the ironic meaning intended in the original, and convey it in the target text in such a way that its addressees may infer similar implicatures to those intended in the original text. It is here that a balance must also be found between what is codified and what is left for the addressee to be inferred. It may be argued that if the purpose of the target text is to allow the addressee to reach or infer similar conclusions, this may be done precisely through the introduction of differences in the ways messages are codified.
Next, the relationship between codification and inference will be explored in the analysis of a fragment from Fowles' The FrenchLieuíenant's Woman and one of its Spanish translations. It has been noted (Onega, 1989) that one of the most remarkable aspects of this work is that it emerges as a parody of the Victorian novel for aspects such as its blurring of the boundaries between history and fiction, the break (or perhaps enhancing) of the illusion of the narrativisation of time, or the questioning of the taken-for-granted narrator's omniscience. As a result, there we find an increasingly unreliable narrator, who tends to combine the most minute omniscience with the freedom he struggles (or at least pretends) to endow his characters with. This can be illustrated with the following passages of the work: I do not know. This story I am telling is all imagination. These characters I créate never existed outside my own mind. If I nave pretended until now to know my characters' minds and innermost thoughts, it is because I am writing in (justas I have assumed some of the vocabulary and 'voice' of) a convention universally accepted at the time of my story: that the novelist stands next to God. He may not know all, yet he tries to pretend that he does. But I live in the age of Alain Robbe-Grillet and Roland Barthes; if this is a novel, it cannot be a novel in the modern sense of the word. (...) (...) I know in the context of my book's reality that Sara would never have brushed away her tears and leant down and delivered a chapter of revelation (ST:85).
What tbis passage shows is the existence in the novel of the different levéis of codification that are being employed by the narrator, who may either be totally omniscient or else sceptic about what he is telling, or by the author himself, or even by the different characters, depending on whether we have direct access to their minds or actions, or else only through the omniscient narrator. The translation has reflected these different levéis of codification. Likewise, the clashes to be observed between the different contextual clues (Yus Ramos, 2000a) can be traced very similarly both in the source and the target text. The breaking (or enhancement) of the illusion of the relationship between fiction and reality somehow clash against the encyclopaedic or commonsense assumptions about fiction; it is not usually expected to find a narrator who casts very explicit doubts about what he is telling and ultimately leaves the choice open to the reader, without having perhaps provided him/ her with sufficient evidence to decide. Likewise, the role of the previous utterance to the beginning of this fragment, "Who is Sarah?/ Out of what shadows does she come?"; "¿Quién es Sarah? ¿De qué sombras ha surgido?" mighthave triggered its interpretation as a rhetorical question, or else as a question to be coherently answered right afterwards, thus compensating the expectation created. Nevertheless, the reply provided both in the English text and in the Spanish translation does away with these assumptions, and probably contradicts the sort of answer that might have been expected.
On the whole, it is only the external readership, together with the narrator, who might possibly been offered the sufficient contextual evidence to identify the underlying contradiction between what is made manifest here, the above commented break or enhancement between reality and fiction. As a result, there will be a clear contradiction between the avowed freedom that both the narrator and consequently the author himself claim to have endowed their characters with and the manifest impossibility for these characters to have had access to all the different ontological levéis presented, which will be another of the sources of irony and humour in the work, as the following passage illustrates: Perhaps you see very little link between the Charles of 1267 with all his newfangled French notions of chastity and chasing after Holy Grails, the Charles of 1867 with his loathing of trade and the Charles of today, a computer scientist deaf to the screams of the tender humanists who begin to discern their own redundancy. But there is a link: (...) (ST:257).
In any case, it may be argued that in these instances the basic contextual clues and the features of context accessibility are maintained, and perhaps the reason for this may be searched in that the target text descriptively resembles the original, for the stylistic properties of the text are maintained and no significant instances of context changes may be traced. Perhaps it may be concluded that an interesting preliminary study to be carried out by the translator in order to render a versión which comes cióse to maintain the communicative intentions of the source text may precisely be to pin down the contextual sources on which meaning relies. Then, it is probably left to the translator's criterion to decide the way in which the communicative clues may be expressed in the target text and how cióse it may resemble the original.

Irony and translation as interpretive resemblance. Descriptive and interpretive resemblance as a continuum
As shown above, the notion of "interpretive resemblance" is the most important trait that recurs in the relevance theoretical account of both irony and translation. Then, it follows that this concept should play an important role in any proposals put forward to cope with the translation of irony from a relevance perspective. For Gutt, the notion of interpretive resemblance has two important implications for translation; first, it is "defined in terms of (...) shared expücatures and implicatures" (1996); and second, it allows to establish comparisons between original and translation without comparing both in terms of questioneddichotomies suchas "literal" or "free". Otherauthors (MateoMartínez, 1998) have also argued that the distinction made between descriptive and interpretive resemblance should be best read as a translation continuum.
If the translation of irony is approached as interpretive resemblance, then, it appears that its conveyance in the target text would tend to follow a similar balance between the explicatures and the implicatures of the original, such that the expenditure of efforts by each of the different audiences may be maintained at a similar level.
The application of these tenets will be tested on the basis of the analysis of the translation of a fragment taken from David Lodge's Changing Places (1975), the story of an initially academic interchange of university posts between a British and an American professor which eventually brings forward much deeper changes for them both: She shot him a shrewd glance. 'How many women have there been in your Ufe?' He stirred uncomfortably in the tepid water, and ran some more into the tub. 'That's an unfair question. At a certain age a man can find satisfaction in one woman alone. He needs stability'. 'Besides, Philip will be coming back soon'. 'I thought you said he wasn't?' 'Oh, that won't last. He'll be back, with his tail between his legs. 'Now there's someone who really does need stability'.
In this dialogue, we find that once and again both Morris and Hilary, who are having a love affair, after having swapped couples, keep on teasing and distancing from each other. Even though the Spanish translation tends to reflect the shades of irony and humour present in the original, there are also some striking changes. To start with, Hilary asks Morris a question which she probably knows that may be embarrassing for him, "How many women have there been in your life?", and indeed, Morris' answer makes his unrestlessness manifest for Hilary, an assumption which is also reinforced or strengthened for the external audience through the narratorial comment. But the most significant aspect of Morris' reply is perhaps his deüberate attempt to distance not only from Hilary 's rather direct question, but also from himself, both of which assumptions he makes explicit: "That's an unfair question. At a certain age a man can find satisfaction in one woman alone. He needs stability". In a sense, this can also be read as a self-directed comment, as if he were trying to defend himself against Hilary's implicature that seems to indicate that she feels there may probably have been some, and even too many, and as if she did not really care about Morris' reply. What is more, Hilary does not really take it long to distance herself from Morris, although it may be said that there are slightly different connotations in either versión. To start with, she sounds just a bit more impersonal in the English text, "Now there 's someone who really does need stability", which might be justified on the grounds of Morris' previous self-reference in the third person, and that would be what Hilary would be trying to distance herself from. In contrast, the Spanish versión attempts to convey a similar shade of meaning, albeit by means of the opposite procedure, that is, by making use of an emphatic form of the personal pronoun whose reference is clearly "PhUip": "¡Él sí que necesita estabilidad!" However, the analytic and contextual impücations of either versión really do not differ much from one another, and therefore, it may be concluded that the Spanish versión has aimed at the conveyance of the emphatic, marked form, which in the English text has been realised by the anaphoric relative adverb, there. Moreover, the fact that she is indirectly and implicitly alluding to her own husband, and showing her somehow contemptuous attitude towards her incidental couple is retaken in her next reply. This is motivated by Morris' next sarcastic comment," 'Maybe we could fix himup with Désirée', Morris joked", which shows that he seems to take it for granted that Hilary shared his standpoint, even if her former comment had shown at least some affection for her real husband. And she sets out to distance herself from Morris even more by sympathetically adopting Morris' real wife standpoint, which must have shocked Morris a bit more. This is conveyed in a very similar form in both texts: "Poor Désirée. Hasn't she suffered enough?"/ "Pobre Désirée. ¿No ha sufrido ya bastante?" In synthesis, it may be claimed that if the notion of interpretive resemblance as put forward by relevance theory is applied to translation, this has consequences for the analysis of the ways in which information is conveyed, and whether it is made in an explicit or in an implicit form. It appears that in every communicative act, there are items of information that are conveyed explicitly and others implicitly. In both cases, the addressee is supposed to infer the communicative intention of the speaker. The reasons to leave certain parts of the information implicit may be very varied, but they undoubtedly convey an important part of the message. As we have seen, the fact that the target text resembles the source text interpretive]y then means tliat the same flow of information, in either an explicit or implicit. form must be maintained. However, from the instance analysed it can also be concluded that the translator must assess the sort of information that is being left implicit, and find out the communicator's reasons for doing so. Perhaps in those cases where there is some information that is being deliberately left implicit, and which is related to the ironic message, it is important for the translator to assess the contextual resources available for the external audience, which should be maintained, so that similar conclusions may be reached, and avoid any kind of over-explicitation, which might have spoilt the whole effect intended by Üie communicator.

The role of context
The analysis of the passages referred to above has shown that a remarkable aspect to be taken into account both in the understanding and translation of irony is the context. An important trait of the context as characterised by relevance theory is that it is above all a cognitive entity, which is further defined by its choice by the participants depending on their accessibüity to the assumptions being entertained: By "context" here I mean not simply the preceding linguistic text, or the environment in which the utterance takes place, but the set of assumptions brought to bear in arriving at the intended interpretación. These may be drawn from the preceding text, or from observation of the speaker and what is going on in the immediate environment, but they may also be drawn from cultural or scientific knowledge, common-sense assumptions, and, more generally, any item of shared or idiosyncratic information that the hearer has access to at the time (Wilson, 1994: 41).
It may be argued that for the addressee to be able to grasp the ironic meaning intended by the speaker, he should be able to bave access to the set of assumptions that his interlocutor has aimed to make manifest. Otherwise, he may miss the whole point of the utterance and misunderstandings may arise. As Blakemore says, "misunderstandings occur when there is a mismatch between the context envisaged by the speaker and the one selected by the hearer" (1992:31).
The following example shows a combination of different contexts, which is characteristic of Postmodernist fiction and which may be summarised like this: the narrator/ internal author addresses both the fictional character and the external addressee or reader. At the same time, and precisely by doing so, he is both a fictional, internal author, and Fowles himself, "in the flesh". Thus, on the whole, there is a merging of real or historical with fictional contexts, which are ordy wholly mastered or accessed by the author /narrator, and which the reader musí get acquainted with so as to understand what is going on. It might be said that all these different levéis must be accurately spotted by the translator, who must also reflect them in the target versión, so as to enable the target text addressee to achieve effects similar to those of the source text, without being required any extra expenditure of effort: There was a very clear suggestion in the sharp look sideways that Charles should keep his eyes to himself. He hastily directed his gaze outside his window and consoled himself that at least the person shunned intimacy as much as he did. (...) (...) But when he sank back into his slumbers, the eyes fastened on him again in the same leechlike manner.
You may one day come under a similar gaze. (...) In my experience there is only one profession that gives that particular look, with its bizarre blend of the inquisitive and the magistral; of the ironic and the soliciting. Now could I use you? Now what could I do with you? It is precisely, it has always seemed to me, the look of an omnipotent god -if there were such an absurd thing-should be shown to have. Not at all what we think of as a divine look; but one of distinctly mean and dubious (as the theoreticians of the nouveau román have pointed out) moral quality. I see this with particular clarity on the face, only too familiar to me, of the bearded man who stares at Charles. And I will keep up the pretence no longer. Now the question that I am asking, as I stare at Charles, is not quite the same as the two above. But rather, what the devil am I going to do with you? I have already thought of ending Charles's career here and now; of leaving him for eternity on his way to London. But the conventions of Victorian fiction allow, allowed no place for the open, the inconclusive ending; and I preached earlier of the freedom characters must be given. My problem is simple -what Charles wants is clear? It is indeed. But what the protagonist wants is not so clear; and I ara not at all sure where she is at the moment (ST-.348).
It may be axgued that in this passage, as a result of the blurring of the borderline between fiction and historical reality, several different contexts have been traced, not all of which can be accessed by all the communicators involved. This is particularly the case of the fictional character, the protagonist Charles. To start with, Charles is sitting in front of the stranger, and the scene is told by a heterodiegetic, third-person narrator. Then the narratorial voice shifts to an interchange "you-I", where he seems to address the external reader directly and which starts to weakly convey the implicature that it may be the writer hirnself the one that is speaking: "In my experience there is only one profession that gives that particular look, with its bizarre blend of the inquisitive and the magistral; of the ironic and the soliciting". Eventually this gives way to the direct address to the character, "Now could I use you? Now what could I do with you?", and the rest of the passage comes to reinforce or strengthen the assumption that it is the author himself speaking. The extemal reader may have realised that, in the form of dialogic communication 6 , he is being addressed by the author, even though the former does not prevent himself from keeping on asking the character directly, "But rather, what the devil am I going to do with you?", as a result of the blurring of the borderline between fiction and reality. S/he may reach such a conclusión on the basis of the previous knowledge which the reading of the work, particularly certain passages from chapter thirteen onwards, may have broughthim/her, and by now s/he might have expected shifts like these.
One of the important elements of the meaning of the passage is the delibérate play on omniscience that the narrator sets out to interpret. Thus, at the beginning, the narrator seems to want the audience to believe that he is only witnessing somebody else who in turn is scrutinising the protagonist: "I see this with particular clarity on the face, only too familiar to me, of the bearded man who stares at Charles". Butthe ambiguity is also cleared up by the narrator himself, who deüberately breaks any possible source of misunderstanding, by pointing at the context chosen: "And I will keep up the pretence no longer". The fact that he views this attitude in terms ofpretense does not exelude the fact that he is trying to distance himself from this situation, and is thus echoing it. At the same time, he also distances or echoes himself, from the moment that he undertakes to give an answer to questions which had previously remained rhetorical, unanswered: "Now could I use you? Now what could I do with you?".
There is an important element of self-contradiction in the passage, which helps the author to enhance the parody of Victorian fiction, from which the author-narrator seeks to distance himself as well: "But the conventions of Victorian fiction allow, allowed no place for the open, the inconclusive ending; and I preached earlier of the freedom characters must be given". In this case, parody may be understood as a mention of a previous expression, as Sperber and Wilson (1981) had predicted, even though this aspect has not been expanded by the theory any further.
The translation of the passage follows with the maintenance of similar discourse markers for the initial "now", which are conveyed as "Vamos a ver", and which retake the colloquial tone with which the narrator addresses the character. In contrast, the translation of the tbird "Now", in "Now the question that I am asking. as I stare at Charles. (...)", seems tohave followed a much narrower semantic constraint on relevance (Blakemore, 1987), "sin embargo", which might have been intended by the translator to further enhance the contextual shift marked by the identification of the narrator with the author himself. It might have been intended with the aim of guiding the reader towards the intended interpretation, so that no extra processing effort is demanded on him/her.
The translation of the irony in the passage shows the following features. To start with, there is a shift from subjeetto object, as well as inthematic structure, between the sentence, "There was a very clear suggestion in the sharp look sideways that Charles should keep his eyes to himself", and its translation: "En la rápida mirada de soslayo que le lanzó el desconocido se leía claramente la advertencia de que Charles no metiera los ojos donde no le importaba", where "a very clear suggestion" has been rhematised, so that the encyclopaedicentriesof "advertencia" arecloser to the subject concerned. Another possible translational option seems that of shifting from cause to result. Thus, "consoled himself that at least the person shunned intimacy as much as he did", has become "mientras se consolaba pensando que por lo menos aquel sujeto rehuía las farniliaridades tanto como él". At the same time, the Spanish text has introduced the discourse marker "así que", in "Así que se los dirigió entonces rápidamente hacia la ventanilla,...", which translates "He hastily directed his gaze outside his window,...". With all this, the narrator sounds even more omniscient in the Spanish text, the characters seem to have shrunk even smaller.
As the passage goes on, there seems to be a further effort on the part of the translator to make certain things more explicit than they appear in the original, with the eventual result of an easier perception of irony: for instance, "the eyes" have become "aquellos ojos", where the metonymic component of that part of the body which identifies the observer is more precisely determiaed. In the context, these shades of meaning are important, as the narrator further builds on the significance attached to that look. Such minute details, where the logical form has been thoroughly enriched, contrasts, however, with the allusion to the profession of the writer, which is only weakly implicated in the context, and is ultimately left to the external addressee's criterion to derive: "In my experience there is only one profession that gives that particular look, with its bizarre blend of the inquisitive and the magistral; of the ironic and the soliciting". In fact, it may be argued that it is only with the further reading of the passage that this assumption may be eventually confirmed. The Spanish versión opts for a negative construction, "Según mi experiencia, no existe más que una profesión que mire de ese modo tan particular, combinando lo inquisitivo y lo magistral, lo irónico y lo inoportuno". Again, the passage will introduce further assumptions, to enríen the logical form of what has been only weakly implicated: the following utterances will deal with "that look" further: "It is precisely, it has always seemed to me, the look of an omnipotent god -if there were such an absurd thing-should be shown to have". On top of that, this encyclopaedic entry, which if in the Spanish versión has always been conveyed as "la mirada", in contrast to the English original, where entries such as "look", "gaze" have introduced slightly more variation, is what introduces the topic of the writer as a god, which has been discussed earlier in the novel, and from which Fowles has sought to distance himself. Henee, it is perhaps striking that it is this invage that allows the external reader to recognise that s/he is being addressed by the author himself. In the context, only does Fowles construct this image to then break it and distance himself from it, in a further ironic twist of the twentieth century that witnesses the mindless realist from above, in the sentence: "It is precisely, it has always seemed to me, the look of an omnipotent god -if there were such an absurd thing-should be shown to have". This is conveyed in Spanish as "Es precisamente, siempre me lo ha parecido, la mirada que podría atribuirse a un dios omnipotente... si existiera algo tan absurdo", where the ironic remark of the narrator questioning the propositional content of his former statement is left to the very end, thus enhancing the speaker's disbelief.
At the same time, the homodiegetic narrator shifts from the description of the passenger to his own self-identification with him, which will continué in the following paragraph: '7 see this with particular clarity on the lace, only too familiar to me, of the bearded man who stares at Charles. And / will keep up thepretence no longer. II Now the question that / am asking, as 7 stare at Charles, ...". This is conveyed in Spanish as "Es una mirada que observo con toda claridad en el rostro, tan femiliar para mí, del hombre de la barba, que ahora está contemplando a Charles. Y ya basta de disimulos. // Sin embargo, la pregunta que me hago mientras miro a Charles..." The shift from the heterodiegetic narrator to his identification with Fowles himself has been subtly made, in such a way that there have been no changes in the grammatical subject, and yet there has been an apparent shift in the reference to the speaker, thus weakly confinning again that the narrator has in fect been the author himself speaking. This has also been maintainedin the translation. Itmay be argued that the strength of the implicatures that point at Fowles himself as the character observing Charles has been increasing gradually. The identification of the man being described as the author, John Fowles, may only be accessible to some readers, if they have perhaps been acquainted with a portrait of the author. On top of that, the reader may also rely on the previously acquired assumptions on the sort of allusions to the author found throughout the novel. But it is only towards the end that these impücatures or assumptions are confirmed: "But rather, what the devil am / going to do with you? I have already thought of ending Charles's career here and now; of leaving him for eternity on his way to London. But the conventions of Victorian fiction allow, allowed no place for the open, the inconclusive ending; and I preached earlier of the freedom characters must be given". This is conveyed in the Spanish translation as "Pero los convencionaHsmos de la novela victoriana no permitían, es decir, no permiten, el desenlace vago e indeterminado; además, antes he predicado ya que a los personajes hay que concederles libertad".
Thus, it may be concluded that the irony of the passage has tended to be reflected in the target versión with a similar degree of strength of assumptions, and with rather approximate contextual effects to that of the original. At the same time, there are no contextual shifts in the passage, that is, the different contexts that may have been traced in the original have been maintained in the translation: the nineteenth century character; his observer, who seems to have been in the same context of situation; and the progressive identification of the observer with the author, with the corresponding time shift.
If these conclusions can be generalised somehow to the importance of the analysis of the context in translation, it may be noted that if it is taken to be a cognitive entity, as relevance states, it has to be assessed in terms of the sort of assumptions that are made manifest, and the degrees of accessibility which the different participante may have to it. It is also the case that the assumptions are usually made manifest with a certain degree of strength, and the contextual implications to be derived from them may allow the external addressee a greater or lesser degree of freedom. This also amounts to considering that the comparative definition of relevance as a balance between processing efforts and contextual effects may be a guiding criterion in the assessment of translations (Guillen Galve, 1995-96). 7

Communicative clues and contextual sources
As shown above, in his study of the implications of the relevance theoretical model of communication for translation, Gutt develops the notion of communicative clues, which are those stylistic features "that guide the audience to the interpretation intended by the communicator" (1991:127). In a particular application, Navarro Errasti makes a synthesis of the sources for communicative clues, which may be the following: "semantíc representations, syntactic properties, phonetic properties, semantic constrainte, formulaic expressions, onomatopoeia, the stylistic valúes of words and acoustic poetic properties" (1993:79). If a basic relationship between communication and the context is assumed, as relevance theory does, and if in at least intra-lingual communication, as Gutt himself claims, "communicative clues are a sub-set of the textual properties that are significant for the intended meaning", and "there would not be any difference in essence between a textual property and a communicative clue" (2000b: 153), it might perhaps be argued whether the contextual sources traced by Yus (2000aYus ( , 1997 could also be regarded as communicative clues, and see how this could be applied for inter-linguistic communication. Gutt claims that in the case of inter-lingual or cross-lingual communication the situation may change because each language has its own set of linguistic properties. However, he also notes that the central aspect would be the following: in case properties might have similar effects, despite being different in each language, then they could be regarded as communicative clues: "...One can very often find some means B in language Y that achieves the same or at least a similar effect as property A did in language X, assuming identical contexts" (2000b: 15 3). If one possible approach to irony is based on the incompatibility of one or more of the contextual sources, as Yus claims that "the accessible information conveyed by one or more of these sources has to be incompatible with the proposition expressed by the utterance in order to aid the hearer in the identification of an ironic interpretation" (2000a:27), then the following may be concluded. The translator might analyse the sort of incompatibiüties between the different contextual sources on which irony relies, and then test whether the translated versión could also rely on similar sources, or else study those incompatibilities which, though perhaps relying on other different contextual sources, might achieve similar effects. This hypothesis will be tested next with the analysis of the following fragment from The French Lieutenant's Woman and its Spanish translation: That evening Charles found himself seated between Mrs. Tranter and Ernestina in the Assembly Rooms. (...) Charles and his ladies were in the doomed building for a concert. It was not, ofcourse -it being Lent-a secular concert. The programme was unrelievedly religious. Even that shocked the narrower-minded in Lyme, who professed, at least inpublic, a respect for Lent equal to that ofthe most orthodox Muslimfor Ramadan. There were accordingly some empty seats before the fern-fringed dais at one end of the main room, where the concerts were held.
Our broader-minded three had come early, like most ofthe rest ofthe audience; for these concerts were really enjoyed -in true eighteenth-century style-as much as for the company as for the music. It gave the ladies an excellent opportunity to assess and comment on their neighbours' finery; and ofcourse to show off their own. Even Ernestina, with all her contempt for theprovinces, fell a victim to this vanity. (...) (ST: 111-12, my italics).
Basically, it could be argued thaí the passage presente the following hints that may lead to an ironic interpretation. To start with, even the description of the precise location of Charles seems to contradict the narrator's general claim of the freedom that characters seem to be endowed with: thus," Charles found himself seated", no matter whether he wanted to or not. However, this aspect cannot be reflected in the Spanish text, since the form "se encontraba" cannot be read as a reflexive form. In any case, some of the reasons which may lead the external reader to conclude that the passage may nave been ironically intended can be linked to the following aspects: on the one hand, the narrator builds up the setting by reinforcing the assumptions just put forward: thus, "It was not, of course -it being Lent-a secular concert. The programme was unrelievedly religious". Such premises are emphasised in the Spanish versión, not only through tliis reinforcement, but also through its expression in a rather colloquial form, which somehow breaks with the solemnity of the scenery being described: "Naturalmente, no era un concierto de música prolana, pues se encontraban en Cuaresma. El programa era sacro de cabo a rabo". At the same time, what the explanation supplied by the narrator does, despite having been supposedly designed to allow the readership to grasp the sort of religious feelings entertained by those inhabitants, is to contradict their social expectations by comparing those "narrow-minded" with "the most orthodox Muslim", which would not surely nave been accepted by the people from Lyme, precisely. As for the contrast that the narrator seems to have intended to draw between the three protagonists and the rest of the town, it is based on a description of these characters which contradicts the assumptions just built about them: that is, whereas at first they are set in contrast to the rest of the people as "our broader-minded", which is conveyed in a likewise colloquial form in the Spanish versión as "de manga más ancha", yet at the same time they are soon after compared and equated with the rest: "like most of the rest of the audience", which is reflected in the Spanish versión as "como la mayoría del público". At the same time, another way in which the ironic meaning may be enhanced is the introduction of comments that at least partially contradict what has been just said, and are presented as confirmations of the formen this may be the case of "It gave the ladies an excellent opportunity to assess and comment on their neighbours' finery; and of course to show qff their own".
Thus, it may be concluded that the translator may be trained in the recognition of irony through the incompatibility of contextual sources such as those, which also provide hirn/her with communicative clues that help the audience to identify the ironic meaning intended. Such communicative clues or contextual sources may also be the basis to be followed in the expression of irony in the target language. However, two different aspects may also be noted: first, the identification of such contextual sources can only made aposteriori; that is, it is only through the analysis of the source text that the identificaüon of these contextual sources may be made, and then adjusted to the target text. Second, there is a further aspect why the tracing of these contextual assumptions is interesting, eventhough the conclusions reached must perhaps be handed with care: irony will be the more enjoyable the less it announces itself as such. This reinforces the idea, essential in the relevance account of poetic language, that it may often ultimately be the addressee's decisión to interpret those assumptions that have been made only weakly manifest by the speaker. Nevertheless, all this should be handled with care, and it may be argued that the notion of relevance itself as a balance between processing efforts and cognitive as well as contextual effects is particularly explanatory and appücable here: as commented above, the analysis of the contextual clues and their possible incompatibilities by means of which irony is conveyed may be guiding principies for the translator to decide what to make explicit and what to leave implicit, so that the TT readership may gain access to and enjoy similar contextual effects to those of the ST audience without being required any further processing efforts.

Conclusions
The present paper has set out to trace some of the recurrent challenges that the translator may face as an interlingual communicator in the translation of irony, on the basis of existing contemporary analyses of the problem, and to attempt to suggest certain proposals based on the relevance model of communication as an ostensive-inferential process, whereby the translator is assumed to trace the communicative intentions made manifest in the source text and to convey them in the target language so that they can be processed by the audience, who seeks to obtain communicative effects similar to those of the source readers, without being demanded any extra processing effort.
The analysis carried out has reached the following conclusions or proposals for the problems that the translation of irony may present the translator: -The translation of irony must rely on an approach to communication which views it as an ostensive-inferential process, in which the speaker aims at optimal relevance and at modifying the cognitive environment of the addressee. What is peculiar about translation is that not only an addresser and an addressee take part, but also a third participant, who is a cultural mediator and is both an addresser (a reader of the source text) and an addressee (a writer of the target text). As such, s/he sets out to infer the informative and communicative intentions of the speakers, which may be either explicit or implicit, and convey them in a different cognitive environment, so that the target text audience may reach similar effects to those of the original audience without being demanded any extra processing efforts.
-As a process of ostensive-inferential communication, translation entails a relaüonship between what is encoded and what may be inferred from it. The process of encoding and decoding is always subservient to the inferential recognition of the communicative intention. ín the case of translation, it turns out that so as to reach similar conclusions regarding this inferential process, the message may be codified in a different way. In the case of irony, there is always an extra layer of meaning which usually remains implicit and which has to be inferred from the context.
-So as to be optimally relevant, the translation is geared to interpretively resemble the original, which in the case of translation means that the total amount of implicatures and explicatures found either in the source or in the target text should be maintained. Even though the translator may make explicit certain items of information that may have been presupposed in the original cognitive environment, and which could not be possible accessible to the target audience, these should be differentiated from information that is absent in the original. What is also important is that a similar level of explicitness is maintained in the target text, since an under-explicitation of the ironic message may lead to misunderstandings, whereas its over-explicitation may spoil the whole effect.
-Every process of the inferential recognition of the ironic meaning intended by the speaker always takes place in a certain context, which is approached as a cognitive entity. Especially in the case of figurative language and in the case of irony, the assumptions entertained may come with weak degrees of strength, which means that they are the more enjoyable the less they announce themselves, and that the addressee is supposed to take a greater degree of freedom in the reach of the meaning intended. In any case, the relevance keystone concept as a balance between processing efforts and cognitive and contextual effects is particularly suitable here, as noted above: on the one hand, ironic utterances which cali too much attention to themselves may be annoying because no effort is required, and the ironic content is over-explicit; but on the other hand, an excessive under-explicitation leading to an uneasy identification of irony may require too much effort and even result annoying for the addressees. Therefore, this is another sense in which the criterion of optimal aecessibility to irony (Yus, 1997(Yus, -98, 2000a) is particularly useful both in the identification of irony and in its translation, as it points to a relevance balance between efforts and effects.
-Moreover, the contextual sources and the incompatibiüties shown by them may be a useful criterion to trace the ironic meaning, and may be the communicative clues that may guide the addressee in the process of recognition of the communicative intention. At the same time, the analysis of these contextual sources and their incompatibilities may help the translator in the conveyance of the ironically intended meaning in the target language. Notes 1. Even though I shall refer to the 1981 Sperber and Wilson's paper, "Irony and the Use-Mention Distinction", it may be remembered that this is an English versión of a paper which appeared in the 1978 monographic issue of Poétique devoted to irony. The title of the French versión was "Les ironies comme mentions" {Poétique 36: 399-412).
2. With regard to this, Barbe also remarks correctly that even if this is indeed the case, it is also true that such psychological explanations were never within the scope of traditional theories: "Perhaps Sperber and Wilson are correct when they say that traditional definitions do not consider the whole psychological picture. However, that was not the purpose of these definitions. Traditional