39
IN SEARCH OF TRANSLATIONAL NORMS
THE CASE OF SHIFTS IN LEXICAL REPETITION IN ARABIC-ENGLISH TRANSLATIONS[1]
… as hard as one might try, it is impossible to reproduce networks of lexical cohesion in a target text which are identical to those of the source text.
(Baker, 1992:206; italics added)
1. Introduction
The objective of this chapter is twofold: (a) to try to arrive at some preliminary generalizations concerning the norms which govern the type and direction of translation shifts in the area of lexical repetition; and (b) to illustrate the above attempt by analyzing and discussing an English translation of an Arabic literary text. I shall therefore be specifically addressing the issue of how lexical chains in Arabic texts are rendered in English; the basic argument postulated being that every translation is bound to result in shifts both in the type and size of lexical repetition.
The high level of tolerance of lexical repetition in the Arabic culture is reflected in the texture of its texts by the important rhetorical and textual roles assigned to this phenomenon. By the same token, lexical repetition seems to be almost always functional in the Arabic literary polysystem. Conversely, English rhetoric discourages lexical repetition and tolerates it only when motivated and used as a figure of speech (cf. Johnstone, 1991: 4 and Hatim, 1997: 32). This textual-cultural discrepancy between Arabic and English is expected to have a bearing on both the translation processes and their products whenever the two languages are involved.
Various studies of texts translated from different source languages have tentatively revealed that avoiding lexical repetition seems to be a common translational norm (see, Ben-Ari 1998:2). Besides being worthwhile in themselves, translation studies like these can contribute to translation theory by providing a testing-ground which may lead to the validation, refutation, or –as is mostly the case-, to the modification of these theories. It would, therefore, be illuminating to find out if this postulated repetition-avoidance policy, just mentioned above, applies to Arabic-English translations and, if so, how it is actually implemented by means of different translation shifts. The present study would also attempt to arrive at viable explanations of the norms which determine these shifts
1.1. The nature and role of translational norms
Norms, in general, represent conventional, social standards –or models- of acceptability of behavior which are shared by the members of a certain culture. Translational norms, in particular, embody the general values and expectations of a given community at a given time regarding the correctness and appropriateness of both the process and the product of translation (cf. Toury, 1980: 51). Thus, it becomes evident that it is not only the two language systems involved in translation that are exclusively, or even- mainly, responsible for forming and formulating translated texts. Rather, it is the dominant conventional norms, especially those of the target pole, which intersubjectively play a pivotal role in moulding a model for regulating translation phenomena. This can help explain the high level of regularity of translational phenomena which are observed, described, and explained in the translated texts in a given TL and culture.
The credit goes to Gideon Toury and Theo Hermans for bringing the ‘concept’ of norms to the forefront in translation studies today. A translated text is no longer considered as a product of ‘transcoding linguistic signs’ but rather as ‘retextualizing’ the SL-text (see Shäffner 1999: 3), so as to make it ‘acceptable’ within the norms framework of the TT-culture. The teleological nature of translation in the TL and its culture is hence emphasized. Besides, every translation can be seen as involving a polar opposition of two sets of norms: those of the source text and culture and those of the goal. Consequently, translations are bound to exhibit traces of both poles. Translation scholars therefore differentiate between two types of basic norms: ‘adequacy’ norms, those of the source, and ‘acceptability’ norms, those of the target.
Another distinction made is that between ‘preliminary’ norms and ‘operational’ ones. The former refer to the translation ‘policy’ in a given community whereas the latter govern the translator’s decision-making process which operates during translating. The regularity of prevalence of different translational norms has also been found to vary in translated texts. Consequently, some translation scholars, like Toury (1995:67), also distinguish between: ‘basic’ or ‘primary’ norms, viz. those which are so common as to be almost mandatory for a certain translation phenomenon; ‘secondary’ norms which are favorable and common but not mandatory; and ‘tolerated’ or ‘permitted’ which are not common but operative. More types of norms have been suggested by other scholars, but what all taxonomies share is that the study of norms, notwithstanding nomenclature, has almost become an indispensable prerequisite for the study of translated texts.
Finally, it would also be necessary perhaps to point out that conflicting and opposing sets of norms may be observed to operate simultaneously, and side by side, sometimes. Complete coherence in the application of translational norms is therefore not to be taken for granted. In fact, instances of deliberate departures from expected norms can often be detected in translated texts. Such departures are, however, usually motivated by stylistic considerations and need to be observed and examined both by translators and translation scholars (cf. Hatim and Mason, 1997: 54). It is essential also to keep in mind that any departure from the norms, be those of the ST or the TT or others, is bound to result in translation ‘shifts’; as discussed below.
1.2. Translation ‘Shifts’ or ‘Transpositions’
In a pioneering article entitled “Translation Shifts” published in 1965, J.C. Catford presents us with one of the earliest systematic discussions and taxonomies of the linguistic phenomena of shifts in translated texts. In his article, Catford defines shifts as “departures from formal correspondence in the process of going from the SL (source language) to TL (target language)” (Catford 1965, in Venuti ed., 2000: 141). He then classifies shifts into two main types: (a) level shifts and (b) category shifts. To the former belong all the shifts which occur in translation between the two linguistic levels of grammar and lexis, such as when one language resorts to lexicalize what is grammaticalized in another; as is done sometimes in Arabic when rendering the English aspectual components of the present progressive or perfect verb structures. Four sub-types of category shifts are also listed; the most common of which is perhaps ‘class shifts’. These occur when translation equivalents in two languages belong to two different grammatical classes; such as, for example, when the ‘adjective’ in the English phrase ‘medical student’ is rendered in Arabic by a ‘noun’.
Vinay and Darbelnet’s book, Sylistique comparee du francais et de l’anglais, represents yet another, even earlier, attempt than Catford’s to describe and classify shifts in translation. The book was first published in French in 1958, but was not translated into English until 1995 when it appeared under the title of Comparative Stylistics of French and English: A methodology for translation. In their book, Vinay and Darbelnet use the term ‘transposition’ instead of ‘shift’ and define it as a translation procedure which “involves replacing one word class with another without changing the meaning of the message” (p.36). It thus becomes obvious from this definition that Vinay and Darbelnet’s concept of ‘transposition’ is not as comprehensive as that of Catford’s ‘shift’, since it is limited to only one of the latter’s subsystems; viz., that of ‘class shifts’ just mentioned above.
The broader definition of ‘shift’ above, as well as its limited one, restricts its scope to ‘formal’ changes due to differences in the linguistic systems of the two languages involved in the translation process. Yet, it would be safe to maintain that a translated text does not only, not even –mainly, exhibit micro-linguistic changes. The translation process also entails various macro-linguistic shifts; those in lexical repetition being one example thereof. These latter type of ‘shifts’ are not strictly due to differences between the SL and TL ‘systems’ as such, but rather to the different textual and discoursal norms of the two cultures which these languages belong to. As a result, the study of translation shifts has taken on a broader scope lately so as to incorporate linguistic as well as textual/cultural changes in translated texts.
Besides, shifts –even in their microlinguistic sense- are no longer to be seen as departures from the SL alone. Rather, translated texts are assumed to exhibit deviations, viz. non-correspondence, when compared both to the SL as well as to the TL. The teleological nature of a translated text can best be appreciated by the fact that the very existence of such a text is actually realized, first and foremost, in the target language and culture in which it is now written and by whose readers it is read (cf. Toury, 1995: 24). Even so, however, neither the linguistic nor the textual correspondence can be full between the translated text and the TL language and culture, since each and every translated text seems to carry along with it some ‘finger-prints’ from its ST. This binary nature of translated texts is aptly portrayed by Toury when stating that such texts which are produced in a certain target language both “occupy a certain position, or fill in a certain slot, in the culture that uses that language” as well as “constituting a representation in that language/culture of another, preexisting text in some other language, belonging to some other culture” (in Schäffner ed.,1999: 20). Translated texts have consequently come to be widely considered as representatives of a special text-type, among many others, within the TL polysystem.
Deviations from the TT and those from the ST are still unable, on their own, to cater for and explain all the linguistic and textual shifts in translated texts. Descriptive Translation Studies has revealed that many translated texts, regardless of their STs and TTs, clearly share some common linguistic and textual features among them (see Blum-Kulka 1986, for example). Consequently, it would be safe to conclude from such recurrent observations that such features shared by ‘unrelated’ translated texts are attributable to the translation process itself, viz. that they reflect some sort of translation ‘universals’. One such ‘universal’ candidate, for example, is the high level of explicitness observed in translated texts as a result of the translators’ common strategy of explicating information stated implicitly in the source text; the phenomenon which has led to postulating the so-called ‘Translation Explicitation Hypothesis’ (see Baker, p.212; Blum-Kulka, p. 300; and Toury 1980: 60; among others). It would thus be tenable, based on the above discussion, to maintain that shifts in translated texts are of three major types: (a) SL-, or adequacy-induced, (b) TL- or acceptability-induced, and (c) translation process-induced.
Before concluding this section on the various types of translation shifts, it also seems worthwhile to distinguish between ‘obligatory’ and ‘optional’ shifts. The first are due to some differences between the linguistic systems of the SL and the TL, as when the translator has no alternative but to shift from one word-class to another (as was discussed under ‘class shifts’ above). Such shifts, because of their regularity, are also considered part and parcel of the study of norms in translation studies although they are rule-governed. Optional shifts, on the other hand, are still considered more significant since they indicate and better reflect the choices translators actually make during the translation process. Consequently, comparative translation studies, like the present paper, usually focus on non-obligatory shifts.
2. Research corpus and methodology
The research corpus consists of an original Arabic text and its English translation. The Arabic text, by a Jordanian writer, is a short story of about five pages which has been translated into English by a reputed American literary translator.[2] A literary text has been chosen for analysis since lexical repetition in such texts is usually highly motivated and has more functional value than in other text-types (see Lotfipour 1997: 190).
For the sake of analysis, the source text (ST) was thoroughly examined for all chains of lexical repetition. A lexical repetition chain (henceforth, LRC) is made up of at least two occurrences of the same[3] word repeated in a given text, regardless of whether the repetitions occur within the boundary of one sentence or in separate sentences (see, al-Khafaji 2004). Only open-system ‘lexical’, and not ‘functional’, words can make up such chains. Consequently, every single lexical word in the analyzed text had to be manually checked against all other such words in the text so as to detect all instances of lexical chains. Once this laborious and time-consuming task was over and all LRCs were sorted out and recorded, a comprehensive reference table was then worked out. The table, a sample of which is reproduced below, lists all the constituents of each chain starting with the initial item, in the order they actually occur in the ST text. The full version of the following table comprises 97 LRCs as the total number of chains found in the analyzed Arabic text. The longest 4 LRCs consist of 11 lexical words each while the number of lexical constituents in the other LRCs of the text ranges between 2 to 10 words. LRC in the sample Table below, for example, comprises four constituents only. The following text portion is reproduced in order to show the textual environment in which these four constituents of this particular LRC actually appear in the ST (underlining is mine):