THE IMPACT OF DIRECT SPEECH FRAMING EXPRESSIONS ON THE NARRATIVE : A CONTRASTIVE CASE STUDY OF

This paper discusses an application of Relevance Theory methodology to an analysis of a literary text: a short story of Gabriel García Márquez “Buen viaje, señor Presidente” and its English translation. “Close reading” technique carried out on rather linguistic than literary basis allows for adding yet another layer of interpretation to this complex story. The analysis concentrates on the representation of direct speech and particularly on the impact of direct speech framing clauses on the reading of dialogic turns. Specifically, it is argued that the explicit mention of the addressee by indirect object pronouns (which are optional in direct speech framing turns) in Spanish makes the tension between the two protagonists even more palpable, therefore apparently courteous turns can be interpreted as defiant or otherwise antagonistic. In English similar role is played by the contrast between the absence of quotative inversion with subject pronouns and its presence when speakers are identified by full nominals. The parallel effect in both linguistic versions is traced to the distinction between linguistic items carrying mainly conceptual meaning (nominals) and carrying mainly procedural meaning (pronouns) and to the different ways these two kind of elements are processed in comprehension. The paper also provides some arguments for leaving aside literary considerations and treating a literary text as an act of ostensive communication.


Introduction
The aim of this is to present results of applying some elements of close reading technique to literary dialogues and in particular to dialogue framing expressions in Buen viaje, señor Presidente, a short story by Gabriel García Márquez published in Doce cuentos peregrinos, a volume that collects short stories written by the author at various times.Our close reading focused on specific semantic and syntactic features of framing expressions: the speech verb and the syntactic construction it is used in.Specifically, in the Spanish original we have analyzed the use of indirect object pronouns and how their presence affects the reading of direct speech turns.Since these pronouns do not appear in the English translation we attempted to verify if their absence is somehow compensated in it and what elements can be close read as their substitutes.
The paper is structured as follows.Section 2 contains a short overview of some issues concerning direct speech framing expressions.Section 3 is short summary of the plot.Section 4 presents the features of framing expressions in the original narrative (4.1) and the results of our earlier analysis of the Spanish original, conceived as an interdisciplinary study that owed more to literary studies than to linguistic (Linde-Usiekniewicz, Nalewajko, 2013) in 4.2, while (4.3) offers some refinements to the analysis in terms of linguistic theory.Section 5 discusses a possible analysis of the English text and section 6 contains conclusions and some additional arguments in favor of linguistic methodologies being used as tools for close reading of literary texts.
The decision to perform close reading on a piece of literary text in purely linguistic terms may at first glance appear controversial, as it does not take into account all other possible levels of the analysis and interpretation of the text in question.Yet in our initial study of the Spanish original we took care to check the interpretation offered by our analytic procedures against literary analyses of the text.Our claim thus is that close reading did not substantially alter the way the story should be interpreted; just the contrary, it helped to make explicit one more element that contributes to the way the narrative is built.It is therefore not surprising that other pointers to the same feature of the story can be traced to other elements of the narrative.
Another potentially controversial methodological decision has been to attempt to conduct our analysis of a literary text in terms of Relevance Theory, which could be and generally is seen as a theory of ordinary communication.Yet Relevance Theory is successfully applied to translation studies, both in theoretical terms (Gutt, 2000(Gutt, , 2010) ) and as illustrative examples taken from literary prose and poetry (Gutt, 2000, p. 380-381, 385-386).If the Relevance Theory can account for literary translation being apt or faithful, it should be even more applicable to the study of the original.Moreover, in a highly theoretical and linguistic study of procedural meaning Escandell-Vidal and Leonetti (2011) analyze a narrative excerpt (Escandell-Vidal, Leonetti 2011, p. 90) One valid point of criticism of our approach can be made on the grounds of absence of mutually accessible cognitive environment (Sperber, Wilson, 1986, p. 38--46).Obviously, in real-life communication the speaker and her 1 audience are capable of producing a rich set of assumptions about their mutual cognitive environment.By contrast, though some assumptions about the cognitive environment of the author can be made accessible to the readers (and accessing these assumptions actually constitutes a large body of literary studies) the reverse is impossible.Even if a particular writer has some specific reading public in mind, the real-life public may be quite different from the intended one.Yet linguists have successfully discussed pragmatics in isolation from real life (Linde-Usiekniewicz, in press).On the other hand, in real life communication the set of assumptions available in the cognitive environment of the audience and the speaker may vary from relatively large and specific to minimal.People are writing books, including non-fiction, press articles and other non-literary texts and they make some assumptions about the cognitive environment of readers they have never met, and of potential readers.And real-life readers manage to successfully read and understand those texts with few specific assumptions about the authors' mind-set.Outside literary theory and in terms of Relevance Theory it can be assumed that in the narrative prose, the cognitive environment of the reader is construed as they read the narrative: each representation contained in the text contributes to it.
The real difficulty in combining close reading and Relevance Theory lies elsewhere: in Relevance Theory it is said that the comprehension procedure should stop once the audience's "expectations of relevance are satisfied" (Wilson, Sperber 2004, p. 613) Yet close reading involves in-depth processing of some part of text, more fine grained (i.e. with more processing effort) that would be done otherwise.Normally, a satisfying cognitive effect of reading literary prose is reached without such processing.On the other hand, when applied to literary text the notion that "every ostensive stimulus presupposes its own relevance to an audience" is much more valid than in normal comprehension.Close reading takes it as its basic premise, and assumes that "it is the most relevant one compatible with communicator's abilities and preferences" (Wilson, Sperber 2004, p. 611).
2 Direct Speech Framing Expressions: An Overview Speech framing expressions are utterances that introduce or accompany speech representations in narratives.Their linguistic form depend on many factors, including the way speech is introduced (as direct, indirect, or free direct discourse).In particular, when accompanying direct speech turns, they are separated from these turns by typographical means.As it is widely known, in some languages (e.g.English) the typography consists in wrapping the direct speech turns in quotation marks (single or double), in others (e.g.Polish, Russian and Spanish), generally dashes are used, as can be seen in the following quote taken from the Spanish original of the short story and its reproductions in English and Polish.Speech passages within narratives have been studied both within the field of literary studies and from linguistic perspective.Within the theory of literature various features of speech representation in terms of direct speech (as opposed to indirect speech or free indirect speech) have been established, either with reference to single direct speech turns or to conversations represented as direct speech in narratives.While it is generally agreed that within the range of possible literary resources for representing speech direct speech turns are the most 'life-like', it is also recognized that though all the features perceived in them characterize the speech of the protagonist, these features are highly conventionalized (Mc Hale, 2011, $15) and the complete independence of the protagonist's voice is just an illusion.Theoreticians differ as to the degree to which the narrator controls the direct speech turns.For example Fludernik (1993, p. 453, quoted in McHale, 2011) argues for the complete illusoriness of the independence of such turns from the surrounding narrative.By contrast, "the Textinterferenz approach (cf.Schmid, 2005, quoted in Mc Halle, 2011), treats speech representation as a matter of interference or interaction between two texts, the narrator's text and the character's text."(Mc Hale, 2011, $20).On the other hand Głowiński (1997) recognizes both the partial independence of the dialogic turns and the fact that they are controlled by narrative commentary.Moreover, he points out that direct speech passages interrupt the flow of the narrative and framing expressions insert them back to it.In addition, framing expressions may inform the reader about the way of speaking, the speaker's attitude and other phenomena accompanying the speech.(Głowiński, 1997, p. 44-48;Suñer, 2000, p. 572;Sawicki, 2010, p. 155-156).
As already mentioned, both represented speech and the lexical verbs occurring in framing expressions have been object of multiple and exhaustive literary studies. 3However, in literary analyses of dialogue the dialogic turns have chiefly been studied within pragmatic frameworks, including speech act theory (cf.Austin, 1962;Searle, 1969) cooperation principle and implicatures (cf.Grice, 1975); and various methodologies of discourse analysis, including cognitive linguistics.Linguistic approaches to dialogue have similarly focused either on linguistic (pragmatic or cognitive) analysis of dialogic turns or on verbs appearing in framing expressions.
By contrast, the syntax of dialogue framing expressions have been researched mainly outside cognitive linguistics and pragmatics.Within generative linguistics the seminal work has been Collins and Branigan (1997) and Collins (1997).Quotative inversion has been studied cross-linguistically, e.g.Suñer (2000) or as an instance of more general inversion phenomena (Alexiadou, Anagnostopoulou, 2007;Baylin 2004;Branigan 2001).Within dependency syntax, and particularly the MTM model Mel'čuk (1988, p. 339-356;2004, p. 252-253) has proposed an important analysis of the use of emotion verbs in framing expressions in Russian, e.g.
lit. " 'But how dare you!", became-angry Ivan.' " 'How dare you!", said Ivan angrily.' He rightly observes that in the Surface-Syntactic representation of such con- The very volume of such research makes it impossible to discuss within this paper.
structions the quote has to precede the framing clause (direct speech introducing clause in his terminology) and the subject has to precede the verb.In his account Deep-Syntactic representation of such constructions feature a silent verb 'to say', and the emotion verb is the attribute of the silent verb.He also rightly points out that attitude verbs are excluded from that syntax and the emotion in question is taken to manifest itself through the speech presented in the quote (Mel'čuk, 1988, p. 345).Emotion verbs are used in direct speech framing expressions in other Slavonic languages, e.g.Polish (Linde- Usiekniewicz, 2012, p. 206-214) and in Romance (cf.Jamrozik, 1992;Jackiewicz, 2005;Lamiroy, Charolles 2008 for French andSuñer, 2000 for Spanish).Linde-Usiekniewicz (2012, p. 206-214) further restricts the meaning of such expressions: she argues that the emotion verbs in direct speech framing clauses encode the fact that the speaker " 'react[s] by being overtaken by the emotion X and showing this emotion through saying Y" ', thus the emotion has necessarily to be a true one, and triggered by the addressee (or their behavior) and not by anyone else.The reaction needs to follow immediately the trigger and not be a delayed one.Interestingly, while subject-verb inversion and the presence or absence of surface subject has been mentioned in several of the works cited so far (e.g.Sawicki, 2008), the presence or absence of indirect object has only been mentioned in generative works on quotative inversion, and only as a part of an argument in favor of one type of syntactic analysis over another (e.g.Suñer, 2000); by contrast, as far as we know there are no studies on the impact of their presence or absence on the semantics of the framing expressions.
3 Plot Summary of Buen viaje, señor Presidente Our analysis of the impact of the mention of the addressee in the framing expressions (Linde-Usiekniewicz, Nalewajko, 2013) did reveal an unexpected twist in the narrative.Yet, it would defeat our purpose to disclose it prior to presenting our analysis and its results.Nevertheless, introducing some idea about the contents of the story is necessary in order to provide some background against which dialogues and their framing is analyzed.That is why we decided to include a plot summary as it appeared in an Andrew Blackman's blog (Blackman, 2010): The dying ex-President of a Caribbean nation is in Geneva, seeing doctors about a mysterious ailment.A man from his home nation recognises him and invites him to his house, with the initial intention of making money out of him by selling him funeral services (the man is an ambulance driver and makes extra money working for funeral parlours, selling their products to dying patients).But as they have dinner and talk, the relationship becomes more complex, and they end up giving more than they take.In our analysis we will be concentrating on the part presenting the initial encounter between the two protagonists.
for rather scarce use of dialogue in his prose; he states that Spanish written dialogue is far-removed from spoken dialogue and real-life dialogue (or rather its mimesis) would not work well in novels (Mendoza, García Márquez 1994).However, the short story we studied is a striking exception: it contains 140 dialogic turns.This would make it an interesting object of study in itself, yet another interesting feature of this story is the way dialogues are framed.The most frequent speech verb used is decir 'to say', the least semantically specific verb available.This choice is particularly thought-provoking when seen against the function generally associated with framing expressions, which is either to describe the mode of speaking, or to provide syntactic support for identifying the speaker or an adjunct describing other circumstances accompanying speech, e.g.speaker's motivation, their attitude, or way of speaking (Sawicki, 2008, p. 155-156;Suñer, 2000, p. 572).Though the wordform dijo 'said' does appear in framing expressions featuring some of the above, as in the following turns: (3) -Esas flores no son de Dios, señor -le dijo, disgustada-.Son del ayuntamiento (p.17) "Those flowers don't belong to God, Monsieur," she said in vexation.[lit.
-Dígale a los que le pagan que no se hagan ilusiones -dijo el Presidente, sin perder la sonrisa ni el encanto de la voz-.Mi salud es perfecta.
-No tanto como el suyo, señor.(p.19-20) The man following him had to stop short to avoid a collision, and his startled eyes looked at him just a few inches away."Señor Presidente," he murmured.[lit.Ø murmured] "Tell the people who pay you not to get their hopes up," said the President, without losing his smile or the charm of his voice."My health is perfect.""Nobody knows that better than me," said the man, crushed by the weight of dignity that had fallen upon him."I work at the hospital."His diction and cadence, and even his timidity, were raw Caribbean."Don't tell me you're a doctor," said the President.[lit.to-him said the President] "I wish I could, Senor.[lit. said the man] I'm an ambulance driver." "I'm sorry," said the President, convinced of his error."That's a hard job.""Not as hard as yours, Señor."(p.8) in many instances it appears on its own, as in: ( In such turns the speaker is identified in the narrative part preceding direct speech.The absence subject pronoun indicates both co-referentiality with the subject of previous sentences and the fact that the subject is the discursive topic of the entire sequence of sentences (RAE, 2009, p. 645), or in other terms, 'old topic' (Givón, 1983).
Pronominal subjects appear in order to identify the speaker.The exchanges quoted in ( 8) and ( 9) take place in Homero's home, with Homero himself, the President and Homero's wife, Lázara, talking.Thus él 'he' in (8) points to the President in this context, while ella 'she' in (9) can refer uniquely to Lázara.
-Sáyago y otros -dijo él-.Todos como yo: usurpando un honor que no merecíamos con un oficio que no sabíamos hacer.Algunos persiguen sólo el poder, pero la mayoría busca todavía menos: el empleo.(p.33) "Sayago," said Homero."Sayago and others," he said."All of us usurping an honor we did not deserve with an office we did not know how to fill.Some pursue only power, but most are looking for even less: a job.Though the verb decir 'to say' normally requires not only a direct object, but an indirect object as well, similarly to English to tell, and in Spanish generally objects cannot be dropped when specific (RAE, 2009, p. 677-678), the indirect object is not obligatory in direct speech framing expressions, as could be seen in previous examples.

Previous Analysis
Our analysis of the impact of introducing the addressee into direct speech framing expressions ran along the following lines.First of all, we observed that typographically direct speech is introduced in the Spanish original in the narrative in two distinct way: wrapped in French quotes as in: (14) [. . .] El médico lo escuchó en suspenso y con el puntero inmóvil en la pantalla.
«Por eso nos despistó durante tanto tiempo», dijo.«Pero ahora sabemos que está aquí» Luego se puso el índice en la sien, y precisó: [. . .] (p.16) [. . .] The doctor listened to him without moving, the pointer motionless on the screen."That is why it eluded us for so long," he said.[lit.Ø said] "But now we know it is here."Then he placed his forefinger on his own temple and stated with precision: (p.4-5)5 and more often, separated by dashes.Moreover, when dashes are used, direct speech together with the framing expression constitute a separate paragraph.Since segmentation into paragraphs "organize the narrative texts into intermediate structural units" (Ji, 2008(Ji, , p. 1719) ) we also assumed that individual paragraphs should be thus taken to constitute a meaningful units into which the narrative is segmented.In accordance to Herman (2009: 96), who argues for applying Langackerian notions of conceptualization and construal (cf.Langacker, 1987Langacker, , 1999Langacker, , 2001) ) instead of that of focalization (cf.Genette, 1980Genette, , 1988)), we took each paragraph to correspond to the maximal scope of conceptualization (Langacker, 1999, p. 49).Thus direct speech framing expressions featuring indirect object were understood to introduce the addressee into the maximal scope of conceptualization.However, cognitive grammar offered no clue to why addressees are only introduced by pronouns and never by a full nominal, in contrast to speakers, who are identified by either, as for Langacker both definite nominals and pronouns serve as grounding elements for the predication (Langacker, 2001, p. 165-167).By contrast, Relevance Theory (Blakemore, 1987;Wilson and Sperber, 1993;Bezuidenhout, 2004;Cram and Hedley, 2005;Hedley, 2005;Wilson, 2011) provides a valid distinction between the two: nominals predominantly encode concepts (Wilson, 2011, p. 6-7), such as 'the President', 'the man' 'Homero', and pronouns represent predominantly 'procedural meaning', which roughly corresponds to instructions how to process the representations arising from the utterances.Specifically, for the purpose of our analysis anaphoric pronouns were understood as instructing the reader to seek the referent in the previous discourse. 6In the narrative studied this referent is usually identified by the nominal outside the paragraph that constitutes the maximal scope of conceptualization.As a result, the appearance of pronominal objects was seen as forcing the reader to quit their mental comfort zone determined by each paragraph and remind themselves of the identity of the addressee, and to come back to the paragraph featuring direct speech in order to follow the narrative.This mental toing and froing imposed on the reader was taken to make the readers feel as if they were spectators of some kind of tennis match, trying to follow the ball with their gaze. 7The added effort of keeping track of who is talking to whom created an extra tension in the reader, and this tension has pervaded the dialogic exchange between the protagonists, even when there is no indirect object in the framing expressions.
Seen in this light, almost every line spoken by Homero can be construed as either defiant or threatening.This can be perfectly seen in the initial conversation quoted in (4).When the President says (falsely) "My health is perfect."Homero apparently agrees "Nobody knows it better than me", but adds "I work in the hospital".This utterance evidently violates either Grice's (1975) quantity maxim, as it does not contribute to the conversation, or the relevance maxim.However, coupled with the fact that the reader knows that the President is lying it may be interpreted as actually constituting a challenge to the lie.Working in the hospital here might be taken to imply possession of some insider's knowledge and in consequence to indicate that Homero is capable of calling the President's bluff.Next we can observe yet another series of violations or transgressions, this time concerning rules of linguistic politeness.Here both protagonists commit their respective offences.After the narrator's voice describes Homero's voice and behavior as "raw", i.e. un-educated, the President actually insults Homero with his "Don't tell me you're a doctor."However, after Homero explains he is an ambulance driver, the President backs out with a phatic "That's a hard job", which is a polite turn only when spoken by a social superior (Laver, 1975;Leech 1983).Homero violates the rules of polite phatic exchange with his response, which can be read as his challenging the President's right to address him as his social inferior.Something similar occurs in the next turns.When asked where he is from Homero answers "The Carribean" which again violates Gricean quantity principle, since (the reader knows that) the President already knows it from Homero's speech.Here the reply can be interpreted as refusing to cooperate, and by the same token, rejecting the President's right to interrogate him.This reading is further reinforced by "From the same as you".In the same vein, defiance can be seen in Homero initial refusal to have lunch with the President ("I never eat lunch") and apparently solicitous Homero's remarks about the President drinking coffee which the latter says had been forbidden to him.
-¿Se dio cuenta?-dijo el Presidente-.Pero hoy fue sólo una excepción en un día excepcional (p.21) He looked at Homero with a roguish eye and changed his tone."In fact, I'm not allowed to eat anything.""You're not allowed to have coffee either," said Homero, "but you drink it anyway.""You found that out?" said the President."But today was just an exception on an exceptional day."(p.10) This remark not only reinforces the image of Homero as possessing special knowledge about the President, but also reminds the reader that Homero have been following the President before their encounter.Interestingly, the President ends by giving an excuse to his behavior, thus apparently recognizing Homero's right to challenge him.

Some Refinements to the Analysis
In our original paper (Linde-Usiekniewicz, Nalewajko, 2013) we concluded that the possibility of interpreting the exchanges as antagonistic underlay their apparent courteous character and provided yet another layer of tension in the narrative.We also mentioned without going into details, that the overwhelming effect of tension was also due to other elements of the narrative, including changes in focalization and a mismatch between the possibly discourteous direct speech exchanges and the narrator's commentaries.Two instances of such apparent mismatch can be seen in ( 4).There when the President indirectly accuses his pursuer of being in the pay of his supposed enemies, he does it "without losing his smile or the charm of his voice".Yet this observable calm should be read as a bit of playacting: either meant to confound his political enemies or, as is revealed at the end of the story, to gain Homero's trust in order to actually swindle him.In the next paragraph (Homero's reply) the feeling of respect that is said to overcome Homero when first talking to the President is described as burden and not an uplifting honor (cf.Nalewajko, 2012 for the role of metaphors in Gabriel García Márquez's prose).Another clue to the narrator's attitude to his protagonists can be seen, among others, in a scene preceding the first encounter: the President steals a flower (cf.3), which indicates that he is not such noble character as may seem at first glance.
More importantly, our analysis can be refined in conceptual and theoretical terms.The framework we originally used combined elements of Langacker's Cognitive Grammar and Relevance Theory.Cognitive Grammar approach allowed us to explain explicit mention of the addressee as their introduction into the maximal scope of predication, while Relevance Theory seemed to provide analytic tools for describing the impact of the addressee being represented by a pronoun, and not by a full nominal.The leap from cognitive methodology to Relevance Theory was facilitated by Dancygier and Vandelanotte's (2009, p. 353) mention of cognitive effort involved in tracing who is talking.In addition, when taken pre-theoretically, Langacker's (2001) analysis of how various elements of discourse are processed and conceptualized can be seen as strongly resembling some models of utterance processing established within Relevance Theory.Specifically, multiple mention of discourse expectations (Langacker, 2001, p. 152) brings to mind the notion of assumptions being raised and presuppositions being foregrounded and backgrounded, and more exactly that of utterances being incrementally processed (cf.Sperber and Wilson 1986, p. 170 and its further application to processing of sentence stress in Breheny, 1998;Wedwood, 2005;andSax 2011, 2012).The existence of some conceptual parallels between the two approaches was noted, but not elaborated upon, in our original paper.
Yet, the original initial steps of our approach, carried out within the framework of Cognitive Grammar i.e. identifying a paragraph featuring direct speech and the framing expression with maximal scope of conceptualization and analyzing the indirect object in framing expression as its introduction into the maximal scope of conceptualization can be reformulated in terms of Relevance Theory alone, or rather within a framework devised on the basis of it, and presented in Linde-Usiekniewicz (in press).
One such major departure from the mainstream Relevance Theory proposed in this work is the introduction of the notion of immediate cognitive environment, as opposed to extended cognitive environment.The immediate cognitive environment is said to contain only those assumptions that are necessary to process a single utterance-type, i.e. an utterance abstracted from its context.The extension of an utterance-type is restricted to the scope of corresponding grammatical sentence either simple or complex.It is further claimed that the original Relevance Theory's statement "when you communicate, your intention is to alter the cognitive environment of your addressees" (Sperber, Wilson 1986, p. 46) can be taken to refer not directly to the extended cognitive environment of the audience, but first of all, to the immediate environment, with actual changes to the extended cognitive environment occurring at a second step, for example when the sense of a larger chunk of discourse is apprehended.This distinction can be illustrated by the contrast between (16a) and ( 16b).In the original quote (16a) the knowledge that the lawyer in question is female is a part of produced change to the immediate cognitive environment of the audience, while in (16b) it is not, though it might become a part of the extended cognitive environment at a later stage or possibly might already be a part of it, as a bundle of assumptions about the speaker's legal advisor.
The notion of immediate cognitive environment can be invoked in the analysis of direct speech framing expressions discussed here because whenever the framing expression follows the direct speech turn, such sequence constitutes a single, albeit complex sentence, in contrast to sequences where framing expressions precede direct speech turns (Linde-Usiekniewicz, 2012, p. 186-197).Thus each paragraph containing a quote and a framing expression can be seen as containing the entire set of changes to be brought about in the immediate cognitive environment of the reader.Since the appearance of the pronoun constitutes the instruction to process a representation (Escandell-Vidal, Leonetti 2011, p. 84) and this representation already present in the extended cognitive environment, the reader is required to process each utterance on its own (i.e.concentrate on the content of the quote and the framing clause, thus staying within their immediate cognitive environment) and at the same time to process it against their extended cognitive environment, established by previous elements of the narrative.Since the indirect object pronouns cannot be analyzed as "relevant to syntactic computation" only, as they are not obligatory, as it was already mentioned, they need to be taken as "interpretive",9 and correspond to what is "usually called 'procedural' in relevance-theoretic terms (Escandell-Vidal, Leonetti 2011, p. 84).
Reformulating our analysis exclusively in terms of Relevance Theory has yet another advantage, besides making it more elegant and less susceptible to criticism on the grounds of methodological eclecticism.It enables us to account for the very presence of the addressee in framing expressions and for their being introduced as a pronoun in a uniform way, while in our previous analysis we needed to split the two phenomena.Similarly, our observations about Gricean implicatures of some of the turns can be reformulated in terms of implicatures as approached within Relevance Theory (Sperber, Wilson 1986, p. 193-202).

Lost or Saved in Translation
As can be easily seen in the translations provided, in English framing expressions corresponding to Spanish ones with pronominal indirect objects the addressee is not mentioned.This holds for the entire story.The same occurs in Polish translation, as can be seen in ( 3).Should our analysis of the original role of the presence of indirect objects in them be correct, several things could have happened in the translated texts.One would be a complete disappearance of the underlying tension.That would mean that our analysis of the original was correct, and the framing expressions in Spanish were crucial to the overall effect of conflict between the protagonists.The second would be an attenuation of the tension.This could be taken to indicate that other elements of the story, i.e. the direct speech turns themselves, or the surrounding narrative passages, including multiple references to the protagonists eyeing each other, still create tension, but possibly obviously so. 10 In other terms, absence of one contributing factor would attenuate the effect.The third would be that the translation maintains the same level of tension and the absence of indirect objects is compensated by some other linguistic element (cf.Newmark, 1988, p. 90).
In order to examine which was the case, we decided to close read the English and the Polish version as if they were independent narratives, since our objective was not to do translation criticism.We concluded that in the English version the tension between the protagonists was much more apparent than in the Polish one. 11Since a detailed check of the two translations proved that all elements of narrative passages that could point to the conflict were faithfully rendered in them, we concluded that the difference should be associated with the way framing expressions were translated, and possibly with the way some elements of direct speech were rendered.What happened in the Polish translation will be discussed elsewhere (Linde-Usiekniewicz, Nalewajko, in preparation).Here we will analyze the English version in an attempt to find how the English compensated for the absence of indirect objects.

Some Noteworthy Features of the English Translation
The English translation has several features that are worth looking into, particularly as they do not appear in the Polish one (and for obvious reasons are not present in the original).One is the title, in which the original Buen viaje 'lit.(have a) good trip' is rendered as a French expression Bon voyage, while the vocative señor Presidente is translated as Mister President.The second is appearance of polite vocative form señor in Homero's turns: (17) "I wish I could, Señor.I'm an ambulance driver."(p.8) (18) "The same as you, Señor," the man said, and offered his hand."My name is Homero Rey." (p. 9) (19) "Of course, Señor," said Homero with amusement."Size forty-one."(p.14) 10 We are grateful to Agnieszka Pantuchowicz (pc) for pointing out that the difference in social class of the two protagonists is yet another factor contributing to the conflict.This issue is further elaborated in Linde-Usiekniewicz, Nalewajko (in preparation).
11 In order to confirm that our impression of the two translated versions was not biased by our knowledge of the original we asked several individuals to read the first few pages of the story (the passages that were subject to analysis) and to comment on how they saw the attitude of the protagonists toward each other: The antagonism was noted in the English version only.
This identifies the protagonists as Spanish-speakers, 12 but also shows that on one level at least Homero does recognize the President as his social superior.On the other level, these form of address appears once in the early part of the conversation, when Homero is still trying to dispel the President's initial suspicions, and twice in turns otherwise identified as defiant.These forms allow the reader to identify the exchanges as non-familiar (in the original both the President and Homero address each other in the polite 3rd person form).Yet translating Señor as Sir would either indicate a greater social distance between the two characters, or a more official setting of the conversations, or, on the other hand, since Sir would not be appropriate in a similar English conversation, could be read as overtly sarcastic, which is not how the original conversation is presented. 13 Interestingly, only one instance of the vocative Señor occurs within a turn that could be read as obsequious: in (17) Homero reacts to the President's challenging and insulting "Don't tell me you're a doctor."In the remaining two instances the vocative accompanies utterances that we had analyzed as defiant: refusing to cooperate and capping presidential joke.

Direct Speech Framing Expressions
These three instances of the vocative can hardly be seen as contributing alone to the overall tension.Our hypothesis is that the effect of tension is at least partly due to the form of the framing expressions themselves.As can be seen when comparing the translations with structural glosses, what distinguishes English expressions from Spanish ones is not only the absence of indirect object pronoun, but the obligatory presence of subject pronouns.Since pronouns, be they subjects or objects, have procedural meaning associated with them cross-linguistically, the reader of the English text is instructed to access the representation of the referent of the subject pronoun in their extended cognitive environment established by the narrative.His task, however, is easier than that of the reader of the original, as subject-pronouns are mainly co-referential with the subject of the sentence(s) preceding direct speech (as can be seen in the English translations of ( 5)-( 7)).In those (and other) passages subjects that are dropped in Spanish, but retained in English correspond to continued topics (Givón, 1983) and are easier to keep track of.Moreover, they appear transparent to the reader since, by virtue of being syntactically obligatory, they are relevant to syntactic computation (Escandell-Vidal, Leonetii 2011, p. 84) However, the presence of pronominal subjects in English framing expressions adds something to the text that was absent from the original: they contrast with framing expressions with nominals, for which the quotative inversion is used almost exclusively, even when the speech verb is coordinated with 12 The hypothesis that one of the functions of leaving Spanish forms of address is to indicate the language spoken is borne out by rendering original señor as Monsieur in (3).The flower vendor is supposed to speak in French to the President.
13 Another interesting feature of the translation is that when Homero introduces himself, his name is translated into English: Homero relaxed."It gets better," he said."Homero Rey de la Casa-l'm Homer King of His House." which establishes him as the President's symbolic equal: a king facing a President.We are indebted for this insight to Martin Veselka (pc).He also mentioned the fact that the protagonist's first name should be seen as symbolic, i.e. alluding to the author of The Iliad and The Odyssey.
an action verb, e.g.:14 (20) "That's my town," said Homero, and he pointed to himself in the group.
"This is me."' (p.11) ( 21) "That's right," said the President, and he stood up, more charming than ever."It appears you even know my shoe size."(p.14)

The Analysis
Our contention is that in the English version the effect of tension produced in the original by the use of indirect object pronouns that is only partly reproduced by the presence of subject pronouns is further reinforced by contrast between framing expressions featuring pronouns, i.e. without quotative inversion, and framing expressions featuring nominals, i.e. with quotative inversion.This contrast does not appear in the original, since pronominal subjects are generally dropped, yet it induces the same kind of mental toing and froing (between processing the utterance against the immediate cognitive environment and against the extended cognitive environment).How this metaphorical mental going back and forth is produced is best explained again in terms of Relevance Theory, and specifically by following the assumption that utterances are processed incrementally by their audience (Sperber, Wilson 1986, p. 170;Breheny, 1998;Wedwood, 2005;andSax 2011, 2012).In the case of direct speech representation in a narrative, where the quote precedes the framing expression, i.e. in the case of 'big quotative inversion' (Linde-Usiekniewicz, 2012, p. 198) the reader is explicitly informed that somebody is talking and builds some hypotheses or assumptions concerning who is talking, how, for what reason or purpose, or under what conditions.Some of these assumptions are based on the assumptions already present in the reader's cognitive environment.On the other hand the reader's knowledge of syntax enables them to produce some expectations of what may directly follow the quote within the framing expression: in English it is either the subject or the verb of the framing clause-again in contrast to Spanish, where the verb is much more likely to follow the quote (Suñer, 2000, p. 572) or will obligatorily follow if the framing expression and the quote really constitute the same sentence (Linde- Usiekniewicz, 2012, p. 186-197).Thus at the stage when the quote itself has been processed, the reader of the English text is faced with an alternative.If the framing expression is headed by the pronoun the reader is induced to do two things at almost the same time: to process the pronoun, i.e. to access their extended cognitive environment for the appropriate representation, and at the same time to build a hypothesis about the lexical verb that would follow, and possibly its internal syntactic arguments and adjuncts.If the framing expression is headed by the verb, there is no impulse to process the utterance against the extended cognitive environment, just the contrary, as the identification of the speaker constitutes the rhematic part of the entire sentence (Sawicki, 2008, p. 155-156;Linde-Usiekniewicz, 2012, p. 199), the readers are made to build anticipatory hypotheses about who is talking.In contrast to the original, which included pronouns that triggered the special type of processing discussed here only in some instances (six times with dijo and four times with preguntó 'asked' out of 60 turns in the entire section of the narrative) in the English version the reader is almost constantly made to follow either of the two procedures (only 13 turns out of 60 have no framing expression following them).
Our observations concerning the form of direct speech framing expressions in the English translation of García Márquez short story should not be taken to mean that we believe that in English the presence of structures featuring subject pronouns and, by the same token no quotative inversion, and structures featuring quotative inversion would always produce the sensation of tension in the reader.As can be seen from the survey of the literature, both concerning syntactic and communicative aspects of quotative inversion, and inversion in general (Kreyer, 2006), there are many factors that can influence its appearance or its absence in written text, including text genre (Sams, 2009), and even editorial directives (Zwicky, 2009).Just the contrary, our contention is that in the story we were analyzing the effect is produced partly by the volume of such expressions, and partly because being the way they are they do not counteract other indicators of antagonism present in the story, including the possibility of reading the exchanges, and particularly Homero's turns, as defiant, and other pointers to the battle of wills presented in the story.Some of the evidence in favor of our hypothesis comes from the parallel analysis of Polish translation, which shows that the way dialogues are framed may be a major factor dispelling or weakening the sensation of conflict between the two protagonists (Linde-Usiekniewicz, Nalewajko in preparation).

Conclusions
In our analysis we consciously avoided reference to the communicative intention, which again is one of the basic concept of Relevance Theory (Sperber, Wilson 1986).Yet it is an important issue that has to be addressed.Another issue we have not addressed, when contrasting the original with a translation was translation strategy adopted.As far as the putative intentionality of introducing indirect object pronouns in Spanish is concerned, it is reasonably safe to assume that they were put there with an ostensive purpose.A writer of García Márquez's caliber has to be believed to be in control of the effect of his prose.The effect of tension we believe is produced in the reader is a part of the writer's general attempt not only to depict sensations of his protagonist, but also to produce similar sensations in his reader (Nalewajko, 2012).In his interview with Mendoza when commenting on Graham Greene's influence on him he says that using Greene's method one can reduce the mystery of the tropics to the smell of a ripe guava (Mendoza, Marquez 1994, p. 22).However, in the English version the quotative inversion of nominal subject might have been the result of trying to achieve "optimal resemblance" (Gutt, 2000, p. 377) to the original in formal terms: nominal subjects were postposed in the original and there was no syntactic necessity to alter it, while indirect object pronouns were dropped and subject pronouns were introduced for grammatical reasons.Nevertheless either consciously or unconsciously the translator has produced a text that resembles the original " [. . . ] in respects that make it adequately relevant to the audience-that is, that offer[s] adequate contextual effects; [. . .] it yields the intended interpretation without putting the audience to unnecessary processing effort."(Gutt, 2000, p. 377).In other terms, the presence of subject pronouns coupled together with quotative inversion can be seen as having the same effect on the English language audience as the indirect object pronouns have on the Spanish language one.
In spite of their being an instance of interlingual interpretive use (Gutt, 2000, p. 378) our contention is that the translations can and should be analyzed independently of the original since they exist fairly independently as far as their readership is concerned (Evan-Zohar, 2000).They are accessed by their readers through the target language only: thus it is the linguistic elements of the translated text that are processed by the readers.This raises yet another question, concerning the general methodological acceptability of applying linguistic tools to literary analysis.As already mentioned, writers do rely on their readers' linguistic skills to get their narratives across and linguistic analysis, however flat and reductionist it may seem, corresponds to an important layer of text understanding.As Kwame Anthony Appiah aptly puts it: "[. . .] we do not have to believe that Jane Austen tells us that "it is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife" in order to express her own ironic attitude to the relations of marriage, gender and property, but we are plainly meant to rely on our understanding of the fact that an utterance of this sentence would convey that ironic attitude outside the fiction."(Appiah, 2000, p. 424). 2

(
16a) I called my lawyer, in one of those midtown firms, and she said [. . .] 8 (16b) I called my lawyer, in one of those midtown firms, and was told [. . .] 5) El Presidente se lo agradeció.-No todos reconocen como usted dignidad del exilio -dijo.(p.21) The President thanked him."Not everyone recognizes as you do the dignity of exile," he said.[lit.Ø said] (p. 9-10) (6) El Presidente se anticipó al reproche.-Yo, por supuesto, ni siquiera me fijaba en usted -dijo.(p.22) The President anticipated his reproach."I, of course, did not even notice you," he said.[lit.Ø said] (p.11) (7) El Presidente miró la hora en el relojito de bolsillo, y tomó las dos tabletas de la noche.Luego escrutó el fondo de la taza: no había cambiado nada, pero esta vez no se estremeció.-Algunos de mis antiguos partidarios han sido Presidentes después que yo -dijo (p.33) The President checked his small pocket watch and took his two evening pills.Then he peered into the bottom of his cup: nothing had changed, but this time he did not shudder."Some of my old supporters have been Presidents after me," he said.[lit.Ø said] (p.22)