Exploring the Retranslation Hypothesis in The Holy Qur’an’s English Translations

1. Introduction

Employing various types of equivalents has long been represented as a solution for coping with some translation problems. Equivalents can show the proximity level of the source text (ST) and the target text (TT).

Selecting specific types of equivalents is highly affected by the kind of procedure employed by translators (Golchinnezhad & Afrouz, 2021a, 2021b; Latifi Shirejini & Afrouz, 2021a, 2021b; Afrouz, 2021c, 2021d; Parvaz & Afrouz, 2021). While some procedures are called source-oriented (i.e., the TT produced via resorting to them is closer to the ST), some others may be target-oriented (i.e., the resulted TT would sound like an original text in the target language).

Different factors can potentially affect the choice of translation procedures. The time period can be one possible factor (Tobias, 2006). Latter translations of work may have specific characteristics which would not be shared by previous ones. Based on Berman’s Retranslation Hypothesis (RH), “later translations tend to be closer to the source text” (Chesterman, 2004: 8); in other words, “earlier translations are closer to the target texts than the later ones” (Afrouz, 2021b: 2). In this regard, it seems reasonable to expect later translations’ quality to be higher than earlier ones (Afrouz & Mollanazar, 2018). However, it may not always appear so. Even the major principles of the RH are not always confirmed. As Deane (2011) asserted, the RH should be “tested in more depth” (p. 9). The current paper is an attempt to test the RH in the context of sacred texts (SCs).

Asserting that sacred texts play a specific “role in society”, Van Poucke (2017: 97) maintains that the essence of such texts “often hampers attempts to modernize the translation, as the slightest misinterpretation of the narrative may result in involuntary blasphemy”. As an implication of his finding, one may expect the Retranslation Hypothesis to be confirmed in dealing with sacred text translations. The present study is conducted to examine the Retranslation Hypothesis in the context of literary-religious texts by concentrating on the Holy Qur’an’s cultural references.

Sacred texts and literary texts are deeply rooted in the “culture” of a nation (Afrouz, 2019: 3, 2022a: 60); therefore, culture-specific terms (CSTs) or terms specifically related to the source language culture are among the most challenging translation problems  (Ordudari, 2008a, 2008b; Thawabteh, 2017; Afrouz & Mollanazar 2016; Putrawan, 2018; Afrouz, 2017, 2020, 2021a, 2022b; Setyawan, 2019; Bywood, 2019; Slavova & Borysenko, 2021; Huber & Kairys, 2021).

It can be claimed that translators may encounter various degrees of difficulty in rendering culture-bound items. Tobias (2006: 28), confirming that their rendering will be more complex “the further apart the two languages and cultures are”, elucidates that this complexity also relies much on “the points in time when the ST was written and when it was translated”. In the present paper, this ‘time’ factor is also considered.

The current paper focuses on translations of the Holy Qur’an to study the potential effect of ‘time’ on the procedures employed in rendering CSTs. The two main questions are:

1) Does the passage of time affect translators' employment of certain (source- or target-oriented) procedures?

2) Is Re-translation Hypothesis confirmed or declined as far as English translations of The Holy Qur’an are concerned)?

2. Literature Review

We focused on some studies conducted on the issue of ‘Time’ in the translation of literary texts, and the RH in The Holy Qur’an.

2.1. Studies Concentrating on the ‘Time’ Factor

Shabanirad (2004) has explored the techniques used in rendering culture-bound terms in literary texts. Seeking “to investigate the passage of time and its effect on the strangeness of translation during two different periods” (Shabanirad 2004: 65), she has limited her study to the comparison and analysis of 20 sentences extracted from 3 English literary texts and two Persian translations for each work. One translation was selected from the pre-revolution, and another one was selected from the post-revolution of the Islamic Republic of Iran. Finding that translators of pre-revolution employed “borrowing” and “notes” more frequently than the post-revolution translators, Shabanirad (2004: 87) concludes that “the passage of time has an important effect on adopting procedures” in dealing with CSTs. The more frequent use of ‘borrowing’, as an extreme source-oriented procedure, by the former translators implies that their findings do not support the Re-translation Hypothesis.

Karshenas and Ordudari (2016), working on translation techniques in a time span, have focused on metaphors in English translations of Sa’di’s the Rose Garden. They adopted Newmark’s (1988: 108-111) framework consisting of the following procedures: recreating the same image in the target text; substituting the source-language image with a standard image in the target cultural system; rendering the metaphorical expression by simile; keeping the image; rendering the metaphorical expression by simile plus sense; converting the metaphor into sense; omission; and addition. Translations were chosen from Rehatsek’s (1888) and Newman’s (2004) books with over a one-century time span. They have also found that the ‘Time’ factor plays a pivotal role in adopting certain procedures. The findings of their study revealed that Rehatsek (1888) resorting to literal translation had most adopted Newmark’s (1988) first procedure; while Newman (2004) had more inclination towards transferring the sense of the source text to the target audience and providing them with a “more understandable” or target-oriented translation (p. 104). Their findings have a very significant implication for the RH. The translation produced by the 19^th-century translator had been closer to the ST than the one by the 21^st-century translator. In other words, their finding stands in stark contrast to the Hypothesis.

2.2. The Re-translation Hypothesis in Religious Texts

Only two studies, as far as the researcher has explored, are focused on testing the validity of RH in religious texts: the first one is an MA thesis conducted in 2015 and the other is an article published in 2018.

El Damanhoury (2015) explored the Japanese translations of the cultural references in the Holy Quran. The researcher’s main objective was to provide some “translation  techniques  that  can  be  used  in  the  translation  of  CRs  from  Arabic  to Japanese” (p. 2).  Furthermore, she has mentioned testing the applicability of the RH as another aim of her study. The works of three Japanese translators were analyzed by the researcher: Toshihiko Izutsu ((1957-8) 1964), Umar (Ryōichi) Mita (1972), and Hassan (Kou) Nakata (2014). Her findings indicated that “Izutsu used foreignizing techniques at approximately 60%, Mita at 62% and Nakata at 70%” of the cases; in other words, “the first translation by Izutsu (1957-8) was the most domesticating, followed by Mita (1972) and lastly Nakata (2014)”; therefore, as she has underlined, the results of her study had been “compatible with” the RH (p. 56-59).

Oyali (2018: 84) has explored the validity of the RH “in representations of certain Biblical concepts in the translations of the Bible into Igbo” and found out that “most of the borrowings in the first translation are de-borrowed in the retranslations”. He finally came to the conclusion that as far as Bible translations are concerned, the RH cannot be validated. Interestingly, Oyali (2018: 97) has realized “that the opposite of the RH is true in this scenario: later translations are more target culture-oriented than earlier translations”.

As it seems, the results of the two studies are not in the same line: one approves the RH and the other disapproves it. This proves, as the researchers have also recommended, the need to test the hypothesis in other language pairs and larger corpora. Therefore, since no one, as far as the researcher knows, has yet validated the RH in the Holy Quran’s English translations, and since the results yet achieved are seemingly contradictory, the present study is conducted to fill the gap and clarify the issue.

3. Methods

The study is corpus-based quantitative research. The corpus includes approximately 3320 Arabic CSTs accompanied by their English equivalents.

3.1. Corpus

The Holy Qur'an, as the “word” of God (Allah) and the greatest ever religious-literary text in Muslim countries (Hassan, 2016: 118; Abdullah and Edris, 2021: 41), “is a distinctive text characterized by its enthralling language” (Najjar, 2020: 1). It was selected as the material of the study since it contained various types of culture-specific terms, and its renditions have been readily available.

The whole 114 chapters (= Surahs) of the book were studied in this respect. The original CSTs and their Persian translations were extracted from the two exegetical texts of the Holy Qur'an entitled “Tafsir Nemuneh” (Makarem Shirazi, 1994) and “Tafsir Noor” (Qara’ati, 1995).

The English equivalents were extracted from the five original translations (being referred to in the reference section).

3.2. Procedure

The Holy Qur’an and its English translations were studied and compared based on a model presented by Eirlys E. Davies (2003).

The general procedure of conducting the study was as follows: identifying Qur’anic CSTs and their English equivalents; specifying the procedures adopted by each translator in dealing with the CSTs; categorizing translators of the same century; tabulating the data and analyzing them; exploring the general textual behavior of translators in each category; providing justifications for such behaviors; and examining the status of Re-translation Hypothesis in the context of religious texts.

Davies (2003) classifies procedures of translating CSTs under 7 major categories:

1. Localization: attempting “to anchor a reference firmly in the culture of the target audience” (Davies, 2003: 84), translators who show a tendency towards Localization, replace CSTs with terms that are more familiar to the target readers. For instance, the CST ‘الأَزْلاَمُ’ /azlam/ refers to a specific game of chance played by ancient Arabs: “یَا أَیُّهَا الَّذِینَ آمَنُواْ إِنَّمَا الْخَمْرُ وَالْمَیْسِرُ وَالأَنصَابُ وَالأَزْلاَمُ رِجْسٌ مِّنْ عَمَلِ الشَّیْطَانِ فَاجْتَنِبُوهُ لَعَلَّکُمْ تُفْلِحُونَ {المائدة/90}”. Irving (1985), the first American translator of the Holy Qur’an, has localized the term as ‘raffles’ for the TT audience: “You who believe, liquor and gambling, idols and raffles, are only the filthy work of Satan; turn aside from it so that you may prosper” (p. 76). Moreover, the term ‘إِمَامٍ مُبِینٍ’ /imamon mubin/ in “فَانتَقَمْنَا مِنْهُمْ وَإِنَّهُمَا لَبِإِمَامٍ مُّبِینٍ {الحجر/79}”, which literally signifies “a clear Book of evidence and guidance”, is rendered by the MG (2012) as ‘a clear ledger’: “So, we sought revenge from them. And they are in a clear ledger.” (p. 169). In fact, the cultural term ‘ledger’ is preferred by the translators to the general term ‘book’. This preference is made based on selecting the strategy of ‘localization’.
2. Globalization: trying to produce a text which is more “accessible to audiences from a wider range of cultural backgrounds” (Davies, 2003: 83), translators who use Globalization, aim at replacing CSTs with terms that are more general or neutral. As an instance, the term ‘الْقِتَالُ’ /qital/ in “کُتِبَ عَلَیْکُمُ الْقِتَالُ وَهُوَ کُرْهٌ لَّکُمْ وَعَسَى أَن تَکْرَهُواْ شَیْئًا وَهُوَ خَیْرٌ لَّکُمْ وَعَسَى أَن تُحِبُّواْ شَیْئًا وَهُوَ شَرٌّ لَّکُمْ وَاللّهُ یَعْلَمُ وَأَنتُمْ لاَ تَعْلَمُونَ {البقرة/216}”, which denotes ‘fighting in the cause of Allah’, is generalized as ‘fighting’ by most translators: “Prescribed for you is fighting, though it be hateful to you. Yet it may happen that you will hate a thing which is better for you, and it may happen that you will love a thing which is worse for you; God knows, and you know not” (Arberry, 1955: 31). Furthermore, the term ‘حُورٍ عِینٍ’ /ḩūren a‘in/ in “کَذَلِکَ وَزَوَّجْنَاهُم بِحُورٍ عِینٍ {الدخان/54}”, which refers to the ‘fair damsels of Paradise, having large black eyes’, is generally rendered by Yuksel et al. (2007) as ‘wonderful companions’: “So it is, and we coupled them with wonderful companions” (p. 317).
3. Transformation: employing this procedure is likely to cause some change in meaning. As an instance, the term ‘الذِّکْر’ /dhikr/ in “إِنَّا نَحْنُ نَزَّلْنَا الذِّکْرَ وَإِنَّا لَهُ لَحَافِظُونَ {الحجر/9}”, which denotes ‘the reminder’, is translated via the use of transformation procedure by Sale (1734) as ‘the Koran’: “We have surely sent down the Koran, and we will certainly preserve the same from corruption” (p. 193). Another translator has just simply rendered it as “the Reminder”. Here, the Sale’s transformation procedure did not result in a change of meaning. As another example where employing transformation caused change in meaning, we can consider Irving (1985) who rendered the term ‘الْجِنَّةِ’ /jenna/ in “مِنَ الْجِنَّةِ وَ النَّاسِ {الناس/6}” which refers to ‘Jinn’, as ‘sprite’: “whether among sprites or Mankind” (p. 332). The term ‘sprite’ (in stories) refers to a small creature with magic powers, especially one that is graceful or likes playing tricks[1]). Irving’s proposed equivalent is not precise since it refers to an imaginary or fictional creature, while, as Muslims truly believe, Jinn is a real creature.
4. Addition: striving to keep the ST term but supplementing it with necessary informative notes is described by Davies as an Addition. In some cases, the note is added as a footnote, within the main text, or being inserted within brackets. As an instance, we can consider the term ‘یَوْمُ الْخُرُوجِ’ /yamol xorūj/ in “یَوْمَ یَسْمَعُونَ الصَّیْحَةَ بِالْحَقِّ ذَلِکَ یَوْمُ الْخُرُوجِ {ق/42}”, which means ‘the day of coming forth’. It is rendered by Pickthall (1930) as ‘the day of coming forth (from the graves)’ via utilizing ‘Addition’: “The day when they will hear the (Awful) Cry in truth. That is the day of coming forth (from the graves)” (p. 168). The translator deemed it necessary to add an informative note (i.e., from the graves) within brackets. As another example, the term ‘یَوْمٍ عَظِیمٍ’ /yaumen azim/ (literally, ‘a great day’) was translated by Saffarzadeh (2001), the first female translator of the Holy Qur’an, as ‘the Great Day of Resurrection’. Here the ‘informative note’ (i.e., the Resurrection) was inserted within the main text. One privilege of such a procedure is that the TT audience’s attention would not be disturbed.
5. Omission: deleting CSTs and providing no substitutes for them in the TL is the procedure described by Davies as Omission. As an example, the term ‘مَّرْضِیَّة’ /marẓiāh/ in “ارْجِعِی إِلَى رَبِّکِ رَاضِیَةً مَّرْضِیَّةً {الفجر/28}”, which denotes ‘pleasing’, was omitted by Muhammad Ali (1917) and Pickthall (1930).
6. Preservation: attempting to “maintain the source text term in the translation” (Davies, 2003: 72) is described as Preservation. Needless to say, it is a source language-oriented procedure. As an instance, the term ‘الْمَجُوس’ /majus/ in “إِنَّ الَّذِینَ آمَنُوا وَالَّذِینَ هَادُوا وَالصَّابِئِینَ وَالنَّصَارَى وَالْمَجُوسَ وَالَّذِینَ أَشْرَکُوا إِنَّ اللَّهَ یَفْصِلُ بَیْنَهُمْ یَوْمَ الْقِیَامَةِ إِنَّ اللَّهَ عَلَى کُلِّ شَیْءٍ شَهِیدٌ {الحج/17}” which refers to ‘the Magians’, was merely ‘transliterated’ or ‘preserved’ by Yuksel et al. (2007) as ‘the Majoos’: “Those who acknowledge, the Jews, the Converts, the Nazarenes, the Zoroastrians, and those who have set up partners; God will separate between them on the day of Resurrection. For God is witness over all things.” (p. 229). The rest of the translators translated it as “the Magians”.
7. Creation: trying to create “culture-specific references” that are not found in the source text is the procedure described by Davies (2003: 88) as Creation. The CST ‘السَّاجِدونَ’ /sajedun/ in “التَّائِبُونَ الْعَابِدُونَ الْحَامِدُونَ السَّائِحُونَ الرَّاکِعُونَ السَّاجِدونَ الآمِرُونَ بِالْمَعْرُوفِ وَالنَّاهُونَ عَنِ الْمُنکَرِ وَالْحَافِظُونَ لِحُدُودِ اللّهِ وَبَشِّرِ الْمُؤْمِنِینَ {التوبة/112}”, which denotes ‘those who prostrate themselves’, is rendered by Nikayin (2000) as ‘prostrators’: “So do repenters, worshippers and praisers, Activists, kneelers, and prostrators, Advocators of justice and forbidders; Of evil, and observers of God’s laws; So give the faithful this good news.” (p. 332). The highly creative Persian poet first added the suffix ‘-or’ to the verb ‘prostrate’ and then pluralized his invented term. Up until now, as far as the researcher knows, no such a lexical item exists in the target vocabulary.

One more procedure was used by the translators which had been left unmentioned in Davies’ model. Following Newmark (1988), we can call it ‘Couplet’. It occurs when two or more procedures are employed in harness. It can occur, for instance, when the translator first applies one of the substitution procedures and then complements it with notes outside the main text. As an example, Mohammad Ali (1917) has first substituted the Qur’anic PN ‘طَالُوتَ’ /talut/ in “وَقَالَ لَهُمْ نَبِیُّهُمْ إِنَّ اللّهَ قَدْ بَعَثَ لَکُمْ طَالُوتَ مَلِکًا قَالُوَاْ أَنَّى یَکُونُ لَهُ الْمُلْکُ عَلَیْنَا وَنَحْنُ أَحَقُّ بِالْمُلْکِ مِنْهُ وَلَمْ یُؤْتَ سَعَةً مِّنَ الْمَالِ قَالَ إِنَّ اللّهَ اصْطَفَاهُ عَلَیْکُمْ وَزَادَهُ بَسْطَةً فِی الْعِلْمِ وَالْجِسْمِ وَاللّهُ یُؤْتِی مُلْکَهُ مَن یَشَاء وَاللّهُ وَاسِعٌ عَلِیمٌ {البقرة/247}” with its Biblical counterpart ‘Saul’ and then has provided his readers with a footnote: “And their prophet said to them: Surely Allah has raised Saul* to be a king over you. They said: How can he have kingdom over us while we have a greater right to kingdom than he, and he has not been granted abundance of wealth? He said: Surely Allah has chosen him in preference to you, and has increased him abundantly in knowledge and physique, and Allah grants His kingdom to whom He pleases, and Allah is Ample-giving, Knowing.” (p. 115). This procedure can also occur in a different way. As an instance, Sale (1734) renders the term ‘أُمَّ الْقُرَى’ /omolqora/ in “وَکَذَلِکَ أَوْحَیْنَا إِلَیْکَ قُرْآنًا عَرَبِیًّا لِّتُنذِرَ أُمَّ الْقُرَى وَمَنْ حَوْلَهَا وَتُنذِرَ یَوْمَ الْجَمْعِ لَا رَیْبَ فِیهِ فَرِیقٌ فِی الْجَنَّةِ وَفَرِیقٌ فِی السَّعِیرِ {الشورى/7}”, which denotes ‘the mother of cities’, as ‘the metropolis of Mecca’: “Thus have we revealed unto thee an Arabic Koran, that thou mayest warn the metropolis of Mecca, and the Arabs who dwell round about it; and mayest threaten them with the day of the general assembly, of which there is no doubt: one part shall then be placed in paradise, and another part in hell.” (p. 362). He has first substituted the term with ‘Mecca’, which the CST actually refers to, and then attempts to denote the reason why it is called ‘أُمَّ’ /om/ (i.e., mother) by adding the word ‘metropolis’.

Figure 1 illustrates the continuum of source- and target-oriented procedures. Note that only the first letter of Davies’ procedures is referred to:

Figure 1. The continuum of source- and target-oriented procedures

The frequency of the abovementioned procedures chosen by each category of translators (see Table 1) are identified via the Excel software. Finally, the frequencies are compared and the Re-translation Hypothesis is examined.

4. Results and Discussion

Translation Studies consider translation products both diachronically and synchronically. When several translations of one single ST rendered by different translators at diverse times are compared with each other, they will better explain the common challenges and dilemmas under which translators have resorted to specific methods. As Gentzler (1993: 85) puts it, “[t]he addition of a historical horizon, albeit a purely literary one, is an important one for the development of translation studies, for it provides not only a basis of comparison but also implies a diachronic evolution of language”.

As Table 1 reveals, the translators belong to three periods of time—while only one of them is from the 18^th century, six and seven of them belong to the 20^th and 21^st centuries, respectively.

Table 1. Translators Categorized on the Basis of Time

A separate table was provided for presenting the data related to the frequency of the procedures selected by translators of the three centuries (i.e., 18^th, 20^th, and 21^st). However, since George Sale was the only translator of the 18^th century, no separate table was allotted to present the frequency of the procedures he had adopted. Furthermore, due to the space limitation, translators of the 20^th century are abbreviated in Table 2 as follows: Maulana Muhammad Ali (MA); M. Marmaduke Pickthall (MP); Yusuf Ali (YA); Arthur John Arberry (AJA); Mohammedali Habib Shakir (HS); Thomas Ballantyne Irving (TBI).

Table 2. The 20th-Century Translators of the Holy Qur’an

As is demonstrated in Table 2, each latter translation, compared to its earlier translation, shows a 2% decrease in the tendency towards choosing source-oriented procedures. Therefore, we can safely claim that the RH is disproved for the 20^th-century English translations of The Holy Qur’an.

To have a more vivid picture of the general tendency of each individual translator, the data demonstrated in Table 2 are formulated into Figure 2.

Figure 2. Translator’s general tendency in 20th century

As Figure 2 reveals, from 1917 to 1985 we can observe an increase in the general tendency of translators towards adopting target-oriented procedures, and consequently, a somehow stair-like progressive decrease in selecting source-oriented procedures. This stands in sharp contrast to the main principle of the Re-translation Hypothesis.

In Table 3, translators of the 21^st century are abbreviated in the following manner: Fazlollah Nikayin (FN); Nicolas Starkovsky (NS); Edip Yuksel et al. (EY); The Monotheist Group (MG).

Table 3. The 21st-Century Translators of the Holy Qur’an

As far as the source- or target-oriented procedures are concerned, each translator’s broad tendency can be better realized through Figure 3.

Figure 3. Translator’s general tendency in 21st century

As can be observed in Figure 3, in the translations presented in the 21^st century, there is a decrease in adopting source-oriented procedures (and consequently, a gradual increase in choosing target-oriented procedures). Therefore, in general, no justifiable evidence was detected for confirming the Re-translation Hypothesis in the 21^st-century English translations of The Holy Qur’an.

To have a bird's-eye view, we need also to consider the collective behavior of the translators who published their translations in the same century. Figure 4 displays the average frequency of the procedures in each category.

Figure 4. Average frequency of the procedures

Ignoring the slight difference between the average frequency of the procedures adopted by 20^th- and 21^st- century translators, we can observe a decrease in adopting source-oriented procedures (and consequently, an increase in choosing target-oriented procedures).

Table 4 shows the distribution of procedures being preferred by each group of translators:

Table 4. Percentage of procedures adopted by each group of translators

Interestingly, had we ignored the slight difference (i.e., 1%) between the total percentages of source- and target-oriented procedures adopted by translators of the 20^th and 21^st centuries, we would have concluded that translators of various centuries had an identical collective behavior—they had averagely adopted source- and target-oriented procedures in 8% and 92% of the items, respectively.

5. Concluding Remarks

As Table 4 reveals, the procedure of Addition is employed by the translator of the 18^th century twice and four times as much as the 21^st- and 20^th- century translators did, respectively. It seems that the earlier translators were more inclined to provide clarifying notes and explanations for their TT readers than the recent ones. It conceivably may have its roots in the modern technology and the availability of information and the facility of obtaining information via mass media, the Internet, e-books, etc.

Furthermore, comparing the use of the Globalization procedure by 18^th-century translator with translators of the 20^th and 21^st centuries, we observed an average of 10% increase which denotes stronger inclinations of the latter translators toward selecting general-neutral equivalents.

The finding of the study stood in stark contrast to Van Poucke’s (2017) findings which implied the confirmation of the Retranslation Hypothesis as far as sacred texts were concerned. Similarly, the results of the current study contrasted with the findings of El Damanhoury’s (2015) study.

On the other hand, the findings confirmed Oyali’s (2018) findings in totally disproving the retranslation hypothesis when it comes to dealing with religious or sacred texts.

As regards the collective behavior of translators of the three centuries, we can safely come to the conclusion that the Re-translation Hypothesis could not be confirmed. In other words, the latter translations were not proved to be closer to the source text than the former ones.

Prospective researchers are highly recommended to conduct a confirmatory study on other Qur’anic translations or other sacred texts to see if the findings of the current study are verified.

[1] Oxford Advanced Genie (Software Dictionary)