Article body

1. Introduction

Over the last two decades, the use of corpus material and methodologies has notably increased in the always evolving field of translation studies, opening new avenues of analysis for the study of translated texts. This new area of research has been labeled corpus translation studies (henceforth CTS) and has adopted and developed corpus linguistic methodologies to analyze translating for descriptive and applied purposes (Laviosa 2013: 228). Baker (1993; 1995; 1996; 1999; 2000; 2004) may be considered a pioneer in this respect. She promoted the use of corpus methodologies to uncover the nature of translation as “a mediated communicative event” (Baker 1993: 243). Laviosa (1998, 2002; 2011; 2013; 2015), Bowker (2001), Granger, Lerot et al. (2003), Scarpa (2006), Anderman and Rogers (2008), Beeby, Rodríguez Inés et al. (2009), Xiao (2010), Kruger, Wallmach et al. (2011), Oakes and Ji (2012) or Sánchez Nieto (2014), to name just a few, followed in her footsteps, which has resulted in a scholarly corpus that occupies a new research niche.

This increasing use of corpus material and corpus methodologies in translation studies has reached literary translation studies too. The conflation of CTS and disciplines such as corpus stylistics has made possible the analysis of literary translations using corpus methods. For example, the translated works of Miguel de Cervantes (Ji 2012), James Joyce (Patton and Can 2012) or Tennessee Williams (Ruano San Segundo 2014) into Chinese, Turkish and Spanish respectively have been scrutinized from such a perspective. These analyses, though, are still small in number. Given that the use of corpora in literary translation studies is only a relatively recent development, scholars are constantly learning about, and refining, new methods and areas of application. In this regard, this article is meant to contribute another example to this branch of CTS. More specifically, I will analyze speech verbs in Charles Dickens’s Hard Times and their translation into Spanish in four different versions –Ribera’s,[1] Ros’s,[2] Pozanco’s[3] and Melendo’s[4]– using a corpus-based methodology.[5] Apart from their basic function of introducing speech, these verbs can also contribute to characterization. For example:

Both blustered and muttered support the content of what is being said, thus contributing to the depiction of Mr. Bounderby and Louisa. As Culpeper (2001: 215) stresses in his model of characterization, “the way one speaks can trigger information about […] personality,” which in the case of Dickens and his well-known penchant for caricature results in certain personalities gravitating towards speech acts uttered in very particular ways. Speech verbs fulfill an important function in this respect in his novels (Ruano San Segundo forthcoming). The use of blustered in Hard Times, on the one hand, helps to project Mr. Bounderby’s bad temper as well as his ineffectiveness, since the four occurrences of the verb are used to introduce his words exclusively (see section 4); muttered, on the other hand, reinforces Louisa’s haughtiness and helps to single her out, since it is attached to her voice exclusively too. The use of these “workshop-vivid choices” (Lambert 1981: 15), therefore, affects the way readers form impressions of characters in their minds, and is thus another tool in Dickens’s armory of techniques for characterization.

Rendering this selection of such specific speech verbs is fundamental if Dickens’s style and his individualization of the fictional voices is to be preserved. Thanks to a corpus-based methodology it is possible to measure whether and to what extent translators maintain the stylistic nuances of such speech verbs: with the help of a concordancing software, verbs can be systematically retrieved from the English novel, which allows the emergence of patterns which would escape the naked eye (or trained intuition); the use of an aligned parallel corpus, in turn, allows a detailed analysis of how the original communicative acts are encoded in Spanish. As will be seen, subtle yet important stylistic differences in Pozanco’s, Ribera’s, Ros’s and Melendo’s approaches to the English text can be detected. These results will illustrate the potential of corpus-based approaches in literary translation studies.

2. Methodology

Before discussing the methodology as such, the scope of the study should be defined. The analysis will be limited to direct speech, the most widely used speech-representation strategy among nineteenth-century English novelists (Busse 2010). Furthermore, say, speech verb par excellence, will be left out of the analysis, since, being neutral, it cannot play a characterizing role by itself. This does not mean that it may not be important stylistically speaking. As pointed out by Caldas-Coulthard (1988: 167), a narrator “can gloss utterances with the reporting verb ‘say’ plus either an adverb, an adjective, or a prepositional clause which will mark either manner or attitude,” thus being able to match the accuracy and the range of details supplied by more specific choices of verbs. By itself, however, said does not characterize speech, which is why it has not been searched for and analyzed here.

As far as the search for verbs was concerned, a computer-assisted methodology was used.[7] The English text was processed with the software WordSmith Tools version 6 © (Scott 2013).[8] From a procedural standpoint, direct speech has some advantages which make the retrieval of verbs possible with such a tool: the more or less stable structural pattern (namely, subject + speech verb + projected clause) allows a systematic location of speech verbs, as described below. With very few exceptions, Dickens always follows the same two patterns when glossing his characters’ words in direct speech: placing the projected clause after the projecting clause,[9] with an inversion of the structure subject-verb (example 2), and embedding the projecting clause in the projected clause, also with the aforementioned inversion (example 3):

These arrangements share two features that prove to be crucial in the search of verbs: (i) that Dickens places no element between the closing inverted comma in the projected clause and the speech verb and (ii) that speech verbs are almost always in the past simple tense. Bearing in mind that most ways of saying are regular (Wierzbicka 1987; Levin 1993) – with the exception of said and some other examples that will be referred to below –, their past simple occurrences will share a common characteristic: the suffix -ed. Hence, by searching for those tokens with an -ed ending which appear after an inverted comma (’ *ed in the concordance tool in WordSmith), speech verbs can be effectively isolated, as displayed in Figure 1, which shows twenty examples of the 332 hits retrieved:[10]

Figure 1

Sample of speech verbs retrieved in Hard Times after concordance search ’ *ed

Sample of speech verbs retrieved in Hard Times after concordance search ’ *ed

-> See the list of figures

Thanks to this concordance search, every example like (2) and (3) will be found. It is true, however, that when Dickens uses a personal pronoun instead of the name of the character, the aforementioned pattern is altered, for there is no inversion of the structure subject-verb in the projecting clause:

To overcome this “obstacle,” an alternative concordance will suffice (’ he *ed). Thus, apart from the search ’ *ed, two more searches need to be carried out, in order to locate verbs after the third-person singular male pronoun (concordance ’ he *ed) and the third-person singular female pronoun (concordance ’ she *ed), as in the following examples:

Obviously, no occurrences of irregular verbs are retrieved with such concordance searches. However, this exclusion does not affect the stylistic analysis of Dickens’s use of speech verbs, since most of the irregular verbs are not usually significant from a stylistic point of view.[11] In fact, in the case of Hard Times, examples of only four verbs (began, put in, quoth and went on) were found on close reading. None of these contribute to characterization, since they do not characterize speech:

Finally, it is also worth noting that there are examples which do not conform to the aforementioned patterns, as can be seen in examples (8) and (9):

Such cases are, however, much less frequent than those retrieved after the concordance searches detailed above. In fact, barely thirty examples were found on close reading, a small number when compared with the 332 examples retrieved with the concordance searches detailed below (see Table 2).

Finally, as far as the search of the Spanish verbs is concerned, a manual examination following the 332 English examples retrieved suffices to identify translators’ choices. An English-Spanish aligned corpus containing Hard Times and the four translations under analysis here has been used for this manual examination, which makes it possible to systematically compare how each translator renders each verb in Spanish, as shown in table 1:[12]

Table 1

Excerpt from aligned corpus containing Hard Times and the four Spanish versions under analysis

Excerpt from aligned corpus containing Hard Times and the four Spanish versions under analysis

-> See the list of tables

3. Results

The concordance searches of Hard Times retrieve a total 332 hits. These were then sorted alphabetically, yielding forty-two different verbs, as seen in Table 2:

Table 2

Speech verbs in Hard Times

Speech verbs in Hard Times

-> See the list of tables

The inventory of verbs used in Spanish by Melendo, Ros, Ribera and Pozanco is shown in table 3:

Table 3

Spanish speech verbs in the four Spanish translations

Spanish speech verbs in the four Spanish translations

-> See the list of tables

The tag “NO TEXT” in Ribera’s and Pozanco’s inventories refers to those examples which have not been translated into Spanish because an excerpt from the original novel –which contains a speech verb– has been omitted. This is the case, for instance, of Ribera’s translation of (10). As can be seen, he omits an excerpt of the English text (my emphasis) which contains the speech verb added:

The tag “NO MARK,” on the other hand, refers to those examples where only the projected clause is preserved, thus turning direct speech instances into free direct speech instances, which results in the loss of the speech verb. For example:

Finally, as can be seen in the previous table, there is a mismatch between the number of verbs in the different inventories. As far as variety is concerned, on the one hand, Ribera uses only 25 different verbs, while Pozanco employs a greater range than even Dickens (48 examples). Regarding frequency of use, on the other, translators’s inventories also present striking differences. Pozanco, for instance, uses the Spanish neutral dijo 50 times, more than any other translator. It is true that phrases and clauses may be attached to this verb to match the accuracy of more specific choices. However, translators often resort simply to a bare dijo when specific verbs are used in the original text, which affects the potential of those verbs from a characterizing point of view. These and other aspects are discussed in section 5, but let us first ascertain the stylistic value of the English verbs so as to then focus on the most representative examples.

4. Characterizing role of speech verbs in Hard Times

The stylistic analysis of Dickens’s use of speech verbs that follows will be, of necessity, brief. Nevertheless, a quick look at the most specific choices suffices to discover that speech verbs play a characterizing role in Dickens’s tenth novel. To begin with, many examples are used to gloss certain characters’s words exclusively, thus individualizing their speeches. In the case of Louisa, she is the only character whose words are introduced by pouted, muttered and hesitated:

The use of pouted and muttered contributes to the projection of her childlike behavior. Hesitated, on the other hand, reflects her submission to her father – in fact, it is used to introduce a speech act addressed to him, as can be seen in (14).

Apart from such an individualization, some speech verbs are also repeatedly used to gloss certain characters’s words. This is the case of blustered, which is not only attached to Mr. Bounderby’s speech exclusively but also used on as many as four occasions to project his speech:

If we also bear in mind that roared, bawled and ejaculated introduce only Mr. Bounderby’s speech too, it can be safely concluded that the way readers form an impression of Bounderby can be very much influenced by the choices used by Dickens to gloss his words:

Whimpered is also a good example of a specific verb used repeatedly to gloss the words of a character. Although not exclusively –whimpered is used to introduce Tom’s words once, as will be discussed in (38)–, Dickens uses this verb three times to characterize Mrs. Gradgrind’s speech, which contributes to the projection of her faintness:

This verb renders her weakness and contrasts her, to some extent, to her husband, whose words –and no one else’s– are introduced by groaned, which contributes to the projection of his tendency to complain:

With Tom, finally, an individualizing use of speech verbs is also observed. He is the only character whose words are glossed by growled and grumbled, which contribute to single out his speech and project his sulky character:

In the light of these examples, it can hardly be denied that Dickens’s use of speech verbs is far from serendipitous. Maintaining this stylistic role when translating them into another language is, therefore, fundamental, so as to preserve their characterizing value. Nevertheless, in none of the four versions analyzed here is such value entirely preserved. As will be seen, translators not only do not use a character-bounded selection of verbs, but sometimes they also do not preserve the specific manner of speaking projected by these verbs, which results in a stylistic loss from a characterizing point of view.

5. Speech verbs in the Spanish translations

The meaning of direct speech descriptive verbs as those analyzed here “has to do with the manner of utterances rather than matter” (Caldas-Coulthard 1987: 162). When dealing with them, therefore, translators must bear in mind that their sound component is as central as their communicative value (Faber and Sánchez 1990). The recurrence of this sound component is different in English and Spanish. In the case of English, for instance, the use of sound verbs – i.e., those related with the emission of sounds (Martínez Vázquez 2003) – with a communicative meaning is quite widespread. This is best reflected in the information supplied by dictionaries. As Martínez Vázquez claims (2003: 105), “most English sound verbs show a communicative meaning which is the result of […] metonymy. Spanish verbs rarely show this extended meaning in dictionaries.” Such a lexicographical difference between English and Spanish, though, is normally neutralized in fiction. As Rojo and Valenzuela (2001: 477) assert,

it seems to be a conventional feature of English dialogues to use profusely the general verb to say, leaving to the reader the inference of the specific role that the utterance plays in the interaction frame. On the contrary, in Spanish dialogues, the continued repetition of the general verb decir results in a lack of stylistic naturalness, and translators prefer to look for variants, by incorporating to the verb information surrounding the communicative event.

Needless to say, this description does not characterize Dickens’s use of say. However, Rojo and Valenzuela’s words emphasize a significant point, namely, the Spanish tendency to provide the reader with details surrounding the communicative event. Such a tendency and the fact that they found “that Spanish specific verbs of saying outnumber English ones” (Rojo and Valenzuela 2001: 473) in their corpus reveals the Spanish translators’s tendency to preserve details related to the manner of speaking when translating speech verbs. It stands to reason, then, that Dickensian translators will render direct speech descriptive verbs in Spanish maintaining the wide range of details supplied by Dickens’s careful choices. For example:

Both rugió and bramó preserve the banker’s vehemence and bad mood reflected in the use of roared. They are, therefore, perfectly valid options. However, such translational accuracy is not always maintained in Melendo’s, Ros’s, Ribera’s and Pozanco’s versions. In fact, when one looks at how they translate the examples discussed in section 4, it is discovered that in almost every case there is at least one version which does not keep the characterizing value of the English verb. This results, as will be seen next, in a stylistic loss which affects characters’s portrayals. Let us start with Louisa and Dickens’s exclusive use of pouted to gloss her words. Pozanco, for example, resorts to terció:

The use of terciar does not maintain the childlike attitude projected by pouted. Quite on the contrary, the communicative sense of the Spanish verb alludes to a somewhat mediation on the part of the speaker, as can be seen in the definition of the verb in Diccionario de la lengua española (henceforth DRAE):[13] “terciar. 9. intr. Interponerse y mediar para componer algún ajuste, disputa o discordia” (to interpose and mediate to resolve a dispute or argument; translated by the author). The use of a satellite (Rojo and Valenzuela, 2001: 471) marking Louisa’s attitude, thus matching the details supplied by pouted in the English novel, would have preserved this nuance. This is a well-established practice when translating specific reporting verbs (Rojo and Valenzuela, 2001: 473) and it is, in fact, the strategy devised by Ribera, Ros and Melendo in their versions:

  1. –Precisamente por eso… –dijo Luisa con aire mohíno.

    Dickens 1854/1972: 20, translated by Ribera

  2. –¡Por eso precisamente! –exclamó Luisa, con cara enfurruñada.

    Dickens 1854/1992: 106, translated by Ros

  3. –¡Por eso mismo! –dijo Louisa, torciendo el gesto.

    Dickens 1854/2005: 67, translated by Melendo

Despite being longer constructions, they maintain the characterizing value of the original verbs, which is the key component to be preserved beyond their communicative function. Pozanco resorts to this strategy too, as can be seen in his translation of one of the four occurrences of blustered to introduce Bounderby’s words:

Even though he could have used a verb such as bramar which would have concentrated the meaning in a specific verb (as Ros[14] and Melendo[15] do), this strategy is certainly better than the use of a verb which does not maintain the menacing and ineffective attitude projected by blustered, as in Ribera’s use of exclamó to render the verb in Spanish:

  1. –¡Cómo! – exclamó Bounderby–. ¿Ocurre algo? ¿Por qué tiene aspecto gruñón nuestro Tomás?

    Dickens 1854/1972: 19, translated by Ribera

The verb exclamar maintains just the pitch, but not the attitude,[16] which is the key aspect of Dickens’s selection of blustered to gloss Bounderby’s words. This loss of information is not detected in Ribera’s translation of the remaining examples of blustered, though. In two of them, in fact, he resorts to a satellite too:

Both sopló el huracanado Bounderby and preguntó el señor Bounderby, con voz tempestuosa maintain his attitude. It is true, though, that the regular co-occurrence of blustered + Mr. Bounderby found in the English text is lost when such satellites are used. However, there is no denying that this strategy allows the projection of his callous attitude too. These constructions explain, to a great extent, the frequent use of dijo in the Spanish inventory shown in table 3. However, as mentioned in section 3, translators sometimes resort simply to a bare dijo to render a specific verb which projects character traits. To continue with the example of Bounderby, let us have a look at Ribera’s translation of ejaculated, used to introduce the banker’s words exclusively. Ribera’s use of dijo neutralizes Bounderby’s vehemence:

Estalló or exclamó, without a doubt, better preserve the brusqueness of Bounderby’s words, as can be observed in Melendo’s and Pozanco’s versions respectively:

  1. –Nueve aceites, Merrylegs, no dar una en el clavo, aros, banderolas y saltos mortales, ¡ja! –estalló Bounderby con su mejor carcajada–. Singular compañía, también, para un hombre que se ha hecho a sí mismo.

    Dickens 1854/2005: 84, translated by Melendo

  2. –¡Bálsamo de nueve aceites, pies con bola, ripios, banderas y un perro que se llama Patalegre! –exclamó Bounderby con su más estentórea carcajada–. ¡Un hombre como yo en semejante compañía!

    Dickens 1854/1995: 54, translated by Pozanco

Lázaro Ros’s choice, however, may not project the banker’s impetuousness appropriately, since he chooses suspiró:

  1. –¡Nueve aceites, Patas Alegres; no dar pie con bola, cintas, banderas y morcillas! ¡Vaya, no está mal el trato con esta gente para un hombre que ha subido por su propio esfuerzo! – suspiró Bounderby, y largó luego la más típica de sus carcajadas.

    Dickens 1854/1992: 122, translated by Ros

Suspiró is much closer to the English verb sighed than to ejaculated. It is true that suspiró indicates vehemence, but this verb may also refer to pain or pity.[17] The use of a verb such as estalló, exclamó or even dijo and a satellite would have been more accurate to account for the banker’s vigor.

The use of dijo to render specific verbs which mark attitude is actually more frequent than might be expected, which results in a significant loss of details that affects the stylistic value of Dickensian speech verbs in the translated versions. Let us move on to the Gradgrinds. In the case of Mr. Gradgrind, for example, Pozanco resorts to dijo to translate groaned, used by Dickens to introduce his words exclusively:

This choice neutralizes Gradgrind’s attitude, subtly projected by the use of such a speech verb. Ribera, Melendo and Ros, conversely, maintain his grumpiness thanks to the use of gruñó and gimió, a much more appropriate selection of verbs:

  1. –Pero mírele –gruñó el señor Gradgrind –. ¿Qué coche querrá…?

    Dickens 1854/1972: 274, translated by Ribera

  2. –Pero, con ese aspecto –gimió el señor Gradgrind –, ¿querrá el coche…?

    Dickens 1854/2005: 382, translated by Melendo

  3. –Fijaos en su estado –gimió el señor Gradgrind–. ¿Creéis que puede ir así…?

    Dickens 1854/1992: 431, translated by Ros

In the case of Mrs. Gradgrind and Dickens’s repeated use of whimpered to gloss her words, a significant stylistic loss has also been detected in two translations. Melendo, on the one hand, uses dijo to render the verb in Spanish, which neutralizes her feebleness and her tendency to see herself as being victimized:

The exact same problem is seen in Ribera’s version. Although he does not use the neutral dijo, he chooses a verb (prosiguió) which barely guides the reader through her words, thus neutralizing the weakness projected by whimpered:

  1. –¡Oh! –dijo la señora Gradgrind–, ¡has terminado ya este asunto! […]. Voy a darte un beso de enhorabuena, Luisa; mas no me toques el hombro derecho: siento en él un dolor que me va de arriba abajo. Cásate, que ahora –prosiguió la señora Gradgrind, ajustando su chal a raíz de esta afectuosa ceremonia – tendré que devanarme los sesos, de la mañana a la noche, para saber cómo llamarle a él.

    Dickens 1854/1972: 96, translated by Ribera

Last but not least, let us look at those examples tagged as “NO TEXT” and “NO MARK” which were referred to in section 3. The omission of verbs either because a larger excerpt of text has been omitted or because the translator gets rid of the projecting clause where the verb is may of course result in a significant loss from a characterizing point of view too. Let us take the example of Tom. As commented on in section 4, he is the only character whose words are introduced by growled, which helps to project his resentful character. Ribera, however, skips an excerpt which includes the turn where Tom’s words are introduced by this verb (my emphasis):

Albeit minute, this omission results in the loss of a characterizing trait which contributes to the depiction of Tom Gradgrind in the novel. Such a loss is not perceived in Ros’s, Pozanco’s and Melendo’s versions. They translate the excerpt where the speech verb is and translate growled either as gruñó or espetó, which project the sulkiness reflected in the original text as far how the speech act is articulated:

  1. –El mequetrefe gruñó, revolviéndose intranquilo en su oscuro rincón:
    – ¡Buena estáis vos! ¡Venir aquí con imputaciones de esa clase…! Bien merecido tendríais que os echase a la calle por no saber comportaros debidamente

    Dickens 1854/1992: 390, translated by Ros

  2. –¡Mucha gramática parda tiene usted! –le espetó Tom, que se decidió a salir de su rincón–. ¡Atreverse a venir aquí con semejantes acusaciones! ¡Si se la pusiera en la puerta por no saber conducirse no se haría sino lo debido!

    Dickens 1854/1995: 377, translated by Pozanco

  3. –¡Buena pieza está usted hecha –gruñó el lobezno, revolviéndose intranquilo en su oscuro rincón – al venir aquí con esas preciosas imputaciones! ¡Se la tenía que despachar por no saber cómo comportarse, y le estaría muy bien!

    Dickens 1854/2005: 342, translated by Melendo

The same loss detected in those cases where no text is found in the Spanish version can be applied to those labelled as “NO MARK.” To continue with the example of Tom, Dickens also uses whimpered to introduce his words once. This verb, employed normally to gloss Mrs. Gradgrind’s speech acts, has the same characterizing role when attached to Tom: it is indeed used when he plays the victim. More specifically, it is used when his guilt is about to be exposed after it is discovered that he robbed the bank and threw the blame on Stephen Blackpool:

Ros, however, opts for omitting the projecting clause, which includes the speech verb, thus turning an instance of direct speech into an instance of free direct speech, where the narrator’s interference and the characterizing potential of the verb are nullified:

  1. –Me imagino que habrá que hacer eso. De todos modos, en ninguna parte seré más desdichado de lo que he sido aquí durante todo el tiempo a que alcanzan mis recuerdos. ¡Que conste eso!

    Dickens 1854/1992: 430, translated by Ros

Without a doubt, it seems much more accurate to translate the projecting clause and to use a verb which projects Tom’s moaning through the use of a very particular way of speaking, as Ribera, Pozanco and Melendo do:

  1. – Sé que no tengo otro remedio. Por lo demás, no sé que pueda llevar en otro sitio una vida más miserable que la que llevo en éste, desde que tengo uso de razón –dijo el mequetrefe lloriqueando.

    Dickens 1854/1972: 274, translated by Ribera

  2. – Supongo que sí. Dudo que pueda sentirme peor en cualquier otra parte de lo que aquí me siento –gimoteó el canalla– desde que tengo memoria. Así que… algo es algo.

    Dickens 1854/1995: 428, translated by Pozanco

  3. – Supongo que sí. En ninguna parte será más desgraciado –gimoteó el lobezno– de lo que he sido aquí hasta donde me alcanza la memoria. Ésa es la verdad.

    Dickens 1854/2005: 382, translated by Melendo

As can be observed, the omission of speech verbs affects the depiction of characters too. The number of examples under the labels “NO TEXT” and “NO MARK” is different in each version. In Melendo’s, for instance, only three cases of “NO MARK” have been spotted (and none of “NO TEXT”), which makes a big difference from the ten cases of “NO MARK” and the fifty-one cases of “NO TEXT” identified in Ribera’s, for example. Such a difference has a direct impact on the preservation of Dickens’s style in the translated texts.

All in all, the translation of speech verbs in Hard Times into Spanish is substantially different in the four versions analyzed here. Rather than focusing on the translators’s inventories of speech verbs, which differ greatly too, the analysis has concentrated on how they render those specific choices which entail a characterizing role within the story. As has been demonstrated, none of the four translators entirely preserve such a role. To a greater or lesser degree, stylistic losses were detected, which can affect the way readers form impressions of characters in their minds. It is worthy of note, however, that such losses do not seem to stem from a translational difficulty, but from translators overlooking the characterizing value of speech verbs in Dickens, since it is very difficult to glimpse their potential due to their dispersed distribution in the text.

6. Conclusion

Thanks to the combination of approaches from different areas of research such as corpus linguistics, corpus stylistics or corpus translation studies, the use of corpus methodologies to analyze translation and translating has recently reached the field of literary translation studies too. The potential of corpus methodologies in literary translation studies is, in fact, proving as effective as it is for literary stylistics at large. The use of innovative models combining quantitative and qualitative analyses of corpus materials has opened new avenues of study for translated literary works, bringing new insights into the nature of literary translation. The speech verbs under analysis here are a good example. Scattered throughout the novel, it is very difficult to appreciate their role in terms of characterization, let alone to carry out a systematic translation analysis assessing whether and to what extent such a role is maintained in four different texts. With the help of a computer-assisted methodology, though, it has been possible to gauge the varying degrees of correlation between the different Spanish versions. The use of such a methodology has shown not only the varied strategies used by translators to render speech verbs in Spanish, but also that translators do not entirely preserve the characterizing nuances of these linguistic units. In sum, this analysis has demonstrated the potential of corpus methodologies in literary translation studies, contributing another example to the still emerging CLTS (corpus literary translation studies).