1. Introduction
There have been several studies which deal, among other things, with the linguistics of negation in Arabic, where negation is divided into explicit and implicit negation (
Anees 1975;
Al-Makhzumi 2016;
Alsalem 2012;
Muslah 2015). While explicit Arabic negation employs negative particles such as
lam [did not],
laa [do not], and
lan [will not], implicit Arabic negation uses grammatical devices such as interrogatives and conditionals. However, there are only very few studies that have dealt with Arabic and English negation from a translational perspective (e.g.,
Dendane and Dendane 2012;
Al-Ghazalli 2013). In particular, there are no studies, to my knowledge, that raise the question of English non-generic explicit negation by
too and
-ly adverbs and
little and
few determiners, which is supposed to be a problematic area because Arabic lacks this type of explicit negation.
The present study, therefore, aims to fill in this gap in English–Arabic translation studies by addressing itself to the translation of non-generic adverbial and determiner negation, as an area of contrast between English and Arabic that may cause serious problems. While English uses both explicit generic negation (negation by not and similar negative particles) and explicit non-generic adverbial/determiner negation, Arabic lacks the latter type as it exclusively employs explicit generic negation by negative particles, such as laa, laysa, lam, maa, lan, etc., which are all tense sensitive (laa and laysa in present, lam and maa in past (with lam [followed by imperfective verb form being used much more commonly in Modern Standard Arabic than maa (followed by perfective verb form)], and lan in future). It seems necessary, in such case, to examine English non-generic adverbial/determiner negation from a translational perspective in order to explore the translation strategies that translators resort to, when modulating this type of negation, in an attempt to capture its meaning and pragmatics. Such investigation is expected to have theoretical as well as practical implications for people working in English–Arabic translation studies, and may also prompt future investigations between English and other languages that lack this type of negation.
The study is structured as follows.
Section 2 reviews the related literature from both a linguistic and a translational perspective.
Section 3 presents the research questions.
Section 4 describes the textual materials used.
Section 5 offers a detailed analysis of the data.
Section 6 discusses the findings. Finally,
Section 7 concludes the study.
2. Review of Literature
Negation, whose universality is unanimously confirmed in the existing literature on human language (
Dahl 1979;
Payne 1985;
Horn and Kato 2000;
Horn 2001, among others), is generally considered an operator that reverses the truth value of a proposition. It is a unique property of human language: “Negative utterances are a core feature of every system of human communication and of no system of animal communication” (
Horn and Kato 2000, p. 1). Linguists (
Klima 1964;
Clark 1976;
Horn 2001, among others) usually divide negation into two types: explicit negation and implicit negation. On the one hand, explicit negation employs explicit negative particles such as
not (as in
John did not go shopping), negative affixes such as -
il in
This act is illegal, or other negative adverbs, for example,
hardly (as in
John hardly knows anything about mathematics) or determiners, for example,
few (as in
John has only few friends). On the other hand, implicit negation is implied semantically (as in
John prevented his daughter from joining the club, which semantically entails
John did not allow his daughter to join the club) or implicated pragmatically (as in the rhetorical question
Should we keep silent after all these heinous crimes, which conversationally implies
We should not keep silent after all these heinous crimes).
Similarly, Arabic negation is divided into explicit and implicit negation. While English verbal negation is uniform in nature, as it only employs the negative particle
not in such negation (viz.
John does not go to school,
John did not go to school, and
John will not go to school), its Arabic explicit counterpart is highly diversified (for details about Arabic negation, see
Anees 1975;
Al-Makhzumi 2016;
Alsalem 2012;
Muslah 2015), viz. لا يذهب جون إلى المدرسة
laa yaðhabu joon ‘ilaa al-madrasati [not (present) go John to the school], لم يذهب جون إلى المدرسة
lam yaðhab joon ‘ilaa al-madrasati [not (past) go John to the school], لن يذهب جون إلى المدرسة lan
yaðhaba joon ‘ilaa al-madrasati [not (future) go John to the school].
Al-Makhzumi (
2016, p. 265), for example, defines negation as “a linguistic category which is opposed to affirmation and intended to disprove or deny the truth value of a proposition”. Explicit Arabic negation employs negative particles such as
lam and
maa [did not] (e.g., لم يكتب/ما كتب سالم الرسالة
lam yaktub/maa kataba Saalimun ‘ar-risaalata [Negative particle (NEG) write/NEG wrote Salim the-letter] “Salim did not write the letter”,
laa/laysa [do not/is not] (e.g., لا يكتب سالم رسائل
laa yaktubu Saalimun rasaa’ila [NEG write Salim letters] “Salim does not write letters” and ليس الكتاب جيدا
laysa-l-kitaabu jayyidan [NEG-the-book good] “The book is not good”, and
lan [will not] (e.g., لن يكتب سالم رسالة
lan yaktuba Saalimun risaalatan) “Salim will not write a letter”. By contrast, implicit Arabic negation uses grammatical devices such as interrogatives (e.g., هل يتساوى الخير والشر
hal yatasaawa-l-xayru wa-š-šarru? [Question word (Q) equal-the-good and-the-evil] “Are the good and evil equal?” and conditionals (e.g., لو كنت موجوداً لأخبرته
law kunto mawjuudan la-’xbartu-hu [if was(I) present so-told-him(I)); both examples involve implicit negation, viz. the answer to the question must be in the negative (No, they are not equal), and the latter statement implies that the speaker had not been there. Nevertheless, this fact does not cause any serious problems for the translator into Arabic because the one-to-many correspondence between the negative particles is obvious, being tense oriented in Arabic.
Notably, while Arabic translation correspondents are usually accessible for generic negation, e.g.,
John does not adhere to punctuality/لا يلتزم جون بالمواعيد
laa yaltazimu joon bi-l-mawaaʻiidi [not (present) adhere John to-the-punctuality], affixal negation, for example,
John is unhappy/جون غير سعيد
joon ɤayru saʻiidin [joon not happy], and implicit negation (Arabic being as highly lexicalized as English), for example,
John declined the offer/رفض جون العرض
rafaḍa joon al-ʻarḍa [declined john the-offer], which semantically entails
John did not accept the offer and لم يقبل جون العرض
lam yaqbal joon al-ʻarḍa [not (past) accept John the-offer], respectively. However, the translator is required to modulate (
Vinay and Darbelnet [1958] 1995) or look for translation equivalents (
Koller 1979) in the case of adverbial/determiner negation by usually retrieving negation (whether explicit or implicit) in Arabic. For example, the adverbial negation by
too in
John was too ambitious needs to be modulated by recovering generic negation in Arabic as in لم يكن جون واقعياً في طموحاته
lam yakun joon waaqiʻiyyan fii ṭumuuḥaati-hi [not (past) be John realistic in ambitions-his] or implicit negation as in تجاوز جون الحدود في طموحاته
tajaawaza joon al-ḥuduuda fii ṭumuuḥaati-hi [exceeded John the-limits in ambitions-his]. Similarly, the determiner negation in
John has little interest in politics needs to be modulated into generic negation as in ليس لدى جون إلا القليل من الاهتمام في السياسة
laysa ladaa joon ‘illa-l-qaliilu min al-’ihtimaami fis-siyaasati [not (present) with John except little from interest in the-politics] or لا يهتم جون إلا قليلاً في السياسة
laa yahtammu joon ‘illaa qaliilan fis-siyaasati [not (present) be interested John except little in the-politics], or implicit negation as in ينأى جون بنفسه عن الاهتمام في السياسة
yan’aa joon bi-nafsihi ʻan-il-’ihtimaami fis-siyaasati [alienates John with-himself from the-interest in the-politics].
In terms of translation, there are only very few studies on the translation of negation.
Apostolatu and Apostolatu (
2012) deal with literary translation of English negation into Romanian. They show that some negative markers are sometimes unjustifiably omitted, which is usually caused by the differences between the two languages involving negative polarity, scope of negation, and double negation.
Dendane and Dendane (
2012) refer to the one-to-many correspondence between the English particle
not and the many counterparts in both standard and vernacular Arabic, which causes serious problems to machine rather than human translation.
Li (
2017) points out the difficulty Chinese English foreign language (EFL) learners face when expressing adverbial negation by
too due to its Chinese counterpart, which functions as an intensifier. Hence, Chinese learners often erroneously employ the negative adverb
too instead of the intensifier “very” or “so” (e.g., “The party was too good” may be used to mean “The party was very/so good”).
A similar mishap may occur in English-into-Arabic translation.
Farghal and Almanna (
2015a, p. 27) briefly examine negation while discussing syntactic features in translation. Whereas they state that English generic negation by
not is not problematic when rendering it into Arabic, despite the existing one-to-many correspondences, it is argued that the negation embedded in
too can pose a challenge because it requires a translation strategy that recovers negation in Arabic, whether explicitly or implicitly. To demonstrate this point, they give the following example from a published translation in which the negation is missed by replacing the negative adverb
too with the Arabic intensifiers
jiddan [very] (a list of Arabic phonetic symbols (mainly International Phonetic Alphabet (IPA)) is provided in
Appendix A):
1. | I think you’ve been too busy to notice where I have been. |
| ‘aðunnu ‘anna-ka kunta mašɤuulan | jiddan li-tulaaḥiða | ‘ayna ‘anaa |
| think (I) that-you be-you busy | very to-notice | where I |
| “I think you were too busy to notice where I am”. |
| .أظن أنك كنت مشغولاً جداً لتلاحظ أين أنا |
Al-Ghazalli (
2013) discusses the translation of Arabic implicit negation where he unjustifiably argues for unpacking Quranic implicit negation in rhetorical questions. According to his analysis, the Quranic verse هل يستوي الأعمى والبصير
hal yastawi-l-’aʻmaa wa-l-baṣiir [Q equal the-blind and-the-sighted] is erroneously rendered as a generic rather than a rhetorical question by Quran translators as in Yousef Ali’s
Can the blind be held equal to the seeing (p. 135) and M. Pickthal’s
Are the blind and the seer equal (p. 133). Therefore, he claims, implicit negation should be made explicit as in
Are the blind and the one who sees equal? Definitely, this is not true (p. 139). Needless to say here, that mainstream translation theorists (
Nida 1964;
Catford 1965;
Newmark 1988;
Baker 1992;
Hatim 1997;
Dickins et al. 2002;
Pym 2010;
Munday 2012;
Farghal 2012;
Farghal and Almanna 2015b;
Farghal et al. 2015, among others) emphasize the translator’s ability to call up textual/functional material in the target language (TL) that effectively relays its counterpart in the source language (SL). One should note that, textually as well as functionally, Arabic rhetorical questions readily translate into English rhetorical questions (Ali’s and Pickthal’s above), thus remaining within the scope of implicit rather than explicit negation.
5. Analysis
The analysis examined the translation strategies employed by the translators in rendering adverbial and determiner negation in terms of frequency and percentage, which gave a clear picture about their utility when encountering such negation. First,
Section 5.1 presented the markers of adverbial and determiner negation and their distribution in the English corpus. Second,
Section 5.2 examined the translation strategies employed in rendering
too adverbial negation, using both explicit negation (
Section 5.2.1) and implicit negation (
Section 5.2.2). Third,
Section 5.3 looked at how the translators dealt with -ly adverbial negation in terms of translation strategy. Fourth,
Section 5.4.1 investigated how the translators had tackled
little determiner negation and the translation strategies adopted. Finally,
Section 5.4.2 considered
few determiner negation and how sensitive the translators had been to this type of negation.
5.1. Adverbial and Determiner Negative Markers
In terms of the type of negative marker, the English corpus is distributed as shown in
Table 1 below.
Table 1 shows that adverbial negation by
too is the most frequent in the corpus (37%) followed by determiner negation by
little (31.85%). Third comes adverbial
-ly negation including
hardly,
scarcely, and
barely which together account for (23%). Within
-ly negation,
hardly emerges as the most frequent (12.60%), followed by
scarcely (7%), then
barely (4%). The least frequent in the data is determiner negation by
few, which only accounts for (7.40%). These percentages may only give us a preliminary picture about the frequency of adverbial/determiner negation in English discourse. To affirm such frequencies, a large scale quantitative and qualitative corpus linguistics investigation needs to be carried out, which is far beyond the scope of the present study.
5.2. Translating Adverbial Negation by Too
5.2.1. Explicit Negation
Table 2 below presents the frequency and percentage of employing translation strategies in rendering
too adverbial negation.
Table 2 above shows that there are three main translation strategies adopted by Arabic translators when opting to render English
too negation by explicit Arabic negation: (1) Unpacking by coordination, (2) nominalization in simple or complex structures, and (3) indicating degree of attribute.
Firstly, unpacking
too negation by a coordinate Arabic structure featuring explicit negation is the most frequent translation strategy at 36.17% (17 cases) for rendering
too negation. It proves to be a workable strategy, as can be observed in the following example:
2. | Harry was too deeply asleep to hear her. (HP/4) |
| haarii | kaana | ɤaariqan | fii nawmi-hi fa-lam yasmaʻ a-haa |
| Harry | was | sinking | in sleep-his so-not hear-her |
| “Harry was deeply asleep, so he didn’t hear her” |
| .هاري كان غارقاً في نومه فلم يسمعها |
It is possible, also, to relay
too negation in such cases by maintaining the English complex structure (second strategy below), as is shown in the rephrasing of (2) below, with a shift of focus in the ordering of the two propositions:
3. | lam yasmaʻ-haa | haarii li’anna-hu | kaana ɤaariqan fii nawmi-hi |
| not hear-her | Harry because-he | was sinking in sleep-his |
| “Harry didn’t hear her because he was deeply asleep” |
| .لم يسمعها هاري لأنه كان غارقاً في نومه |
Secondly, nominalization in simple or complex Arabic structures comes second at 31.91% (15 cases). The translators have mostly succeeded in capturing
too negation using this strategy, as can be noted in the examples below:
4. | I was too late to save the girl. (HP/3) |
| lam ‘astatiʻ | ‘inqaað-l-fataati fi-l-waqti-l-munaasib |
| not be able | saving-the-girl-in-the-time-suitable |
| “I couldn’t save the girl at the right time”. |
| .لم أستطع إنقاذ الفتاة في الوقت المناسب |
5. | … but his Patronus was too feeble to drive the dementor away. (HP/2) |
| wa-laakinna | taʻwiiðata-hu | lam takun | bi-l-quwwati-l-kaafiyati |
| and-but | Patronus-his | not be | with-the-strength-the-enough |
| li-’ibʻaadi-l- ḥaaris |
| to-driving away-the-dementor |
| “… but his Patronus wasn’t strong enough to drive the dementor away”. |
| .ولكن تعويذته لم تكن بالقوة الكافية لإبعاد الحارس |
Thirdly, capturing
too negation by indicating the degree of the attribute in question accounts for 23.40% (12 cases) of the examples in this category. Semantically, it corresponds to awkwardly rephrasing
too negation by using the phrase
to the extent that with negation by
not in English, viz.
John was too short to touch the ceiling may awkwardly be rephrased as
John was short to the extent that he couldn’t touch the ceiling. In Arabic, this strategy proves very useful for rendering
too negation. The following example is illustrative:
6. | Professor Trelawney seemed too tipsy to have recognized Harry. (HP/5) |
| badat al-’ustaaðatu triiloonii maxmuuratun li-darajati ‘anna-ha lam taʻrif haarii |
| seemed the-professor Trelawney drunk to-extent that-she not know Harry |
| “Professor Trelawney was drunk to the extent that she didn’t recognize Harry”. |
| .بدت الأستاذة تريلوني مخمورة لدرجة أنها لم تعرف هاري |
Finally, let us look at one example where the translator has failed to capture the subtlety of too negation by reducing it to generic Arabic negation, viz. “… but Harry was too used to this to care” is rendered into ولكن هاري لم يهتم بهذا wa-laakina haari lam yahtam bi-haaðaa [and-but Harry NEG care in-this] “but Harry did care about this”. As can be seen, the translator has managed to recover explicit Arabic negation but failed to capture the meaning of too negation. To do this, one may suggest ولكن هاري لم يهتم لتعوّده على هذا wa-laakina haari lam yahtam li-taʻawwudi-hi ʻalaa haaðaa [and-but Harry NEG care because-being used to-him on this] “… but Harry did not care because he was used to this”.
5.2.2. Implicit Negation
Table 3 below displays the frequency and percentage of using implicit negation translation strategies when rendering
too adverbial negation.
Apart from the mixed bag of erroneous translations/under-translations (16 instances, 30.19%),
Table 3 indicates that there are three main implicit negation strategies for capturing the meaning of
too negation: (1) Employing the comparative form, (2) employing negative verbs, and (3) indicating degree of attribute.
The translation strategy using the Arabic comparative form (16 instances, 28–30%) emerges as very useful for handling
too negation. The comparative Arabic forms
‘afʻal min [more of an attribute than] and
‘akθar maṣdar (verbal noun)
min [more of verbal noun than] here capture the nuance that the force of X’s attribute goes beyond the capability of Act Y, for example,
‘aðkaa min ‘an yuxdaʻ [cleverer than that (he) be deceived] and
‘akθara ðakaa’an min ‘an yuxdaʻ [(has) more cleverness than that (he) be deceived] which both idiomatically translate into
Ali is too clever to deceive. Following are two illustrative examples from the corpus:
7. | … because it was too long to memorize. (FS) |
| … li’anna-ha | ‘aṭwalu min ‘an | tuḥfaða | ɤayban |
| … because-it | longer from that | be learned | by heart |
| “… because it was longer than it could be memorized”. |
| .لأنها أطول من تُحفظ غيباً … |
8. | My way of life here is pitched too high for his young head. (BF) |
| ‘inna ḥayaatii | hunaa | ‘akθara | ṣaxaban min ‘an yaḥtammalu-ha ra’sa-hu-l-yaafiʻ |
| verily life-my | here | More | noisy from that tolerated-it head-his-the-young |
| “Indeed, my life was more noisy than what his young head could tolerate”. |
| .إن حياتي هنا أكثر صخباً من أن يتحملها رأسه اليافع |
The second translation strategy utilizes negative verbs/verbals (11 instances, 20.75%) to relay the meaning of
too negation. This is a familiar strategy in English as well as in Arabic to express negation implicitly rather than explicitly. For example, “the act of denying doing something” implies “the act of not admitting doing it”. Consequently, this strategy constitutes an important option when translating negation in general and
too negation in particular. The two examples below are illustrative:
9. | Anna dies or becomes too ill to continue writing it. (FS) |
| ‘aana maatat | ‘aw balaɤat min-al-maraḍi | ḥaddan ḥaala bayna-haa |
| Anna died | or reached from-the-illness | degree prevented between-her |
| wa-bana-l-’istimraari | fi-l-kitaabati |
| and-between-the-continuing | in-the-writing |
| “Anna died or reached a degree of illness that prevented her from continuing writing”. |
| .آنا ماتت أو بلغت من المرض حداً حال بينها وبين الاستمرار في الكتابة |
10. | It just seemed too good to be true that he was going to be rescued from the Dursleys. (HP/5) |
| fa-qad badaa ‘anna ‘inqaaða-hu min ‘aalii duurislii ‘amrun yafuuqu ‘aḥlaama-hu |
| so-? seemed that rescuing-him from family Duresleys matter exceeds dreams-his |
| “It seemed that rescuing him from the Duresleys was something that exceeds his dreams”. |
| .فقد بدا أن إنقاذه من آلي دوريسلي أمر يفوق أحلامه |
The implicit negation in (9) and (10) is achieved by the use of the negative verbs ḥaala [prevented] and yafuuqu [exceed], which both imply propositions employing explicit negation in Arabic.
The third translation strategy employs the degree of the relevant attribute (10 instances, 18.87%) as a marker of implicit negation by using the degree formulas
min +
maṣdar [verbal noun], for example,
from the-smallness [i.e., too small] and
verb/adjective +
ʻalaa [on], for example, (
grow)
old on [i.e., too old]. Observe the two examples below:
11. | Fines for violations were too small to be effective, … (ES) |
| fa-l-ɤaraamaat-u ʻan-il-muxaalafaati | kaanat min-aṣ-ṣiɤari biḥayθu | faqadat faaʻiliyyata-ha |
| so-the-fines from-the-violations | were from-the-smallness so | lost(they) effect-their |
| “The fines for violations were so small that they lost their effect” |
| .فالغرامات عن المخالفات من الصغر بحيث فقدت فاعليتها |
12. | “No, now I am too old to learn anything.” (BF) |
| laa wa-‘anaa al’aana kabirtu ʻalaa taʻallumi ‘ayyi šay’in |
| no and-I now grew old on learning any thing |
| “No, I have grown (too) old to learn anything”. |
| .لا وأنا الآن كبرت على تعلم أي شيء |
Finally, we have the mixed bag, which includes mistranslations/under-translations that account for 16/53 instances (a full 30.19%) in the cases of implicit negation. The two examples below are illustrative:
13. | but it’ll take too long to explain now. (HP/3) |
| laakinna-haa | qiṣṣatun | ṭawiilatun |
| but-it | story | long |
| “but it was a long story”. |
| .لكنها قصة طويلة |
14. | It is still far too cold to undress at night. (BF) |
| laayazaalu-ṭ-ṭaqsu | baaridan | jiddan | li-xalʻi-θ-θiyaabi | fi-l-layli |
| still-the-weather | cold | very | to-take off-the-clothes | in-the-night |
| “The weather is still very [too] cold to take off clothes at night” |
| .لايزال الطقس بارداً جداً لخلع ثيابي في الليل |
5.3. Translating -ly Negation
Table 4 presents the translation strategies which Arabic translators employ when rendering English
-ly negation.
The sample of negative adverbs (which is extracted from BF, FS, GC, and ES, to the exclusion of HP) includes 34 instances of hardly, 19 of scarcely, and 11 of barely, coming to a total of 64 instances. These adverbs share the fact that they communicate a negative orientation when used in English sentences. That is why they are often interchangeable, albeit they may be sensitive to normality conditions (i.e., one may sound natural in one context, while another may not). For example, John was barely 17 when he joined college is natural, whereas John was scarcely 17 when he joined college is not. In terms of translation, the focus is on relaying the negative orientation which is shared by all of them.
Table 4 above shows only two successful strategies the translators have employed in rendering
-ly negation: (1) Correct explicit negation (12 instances, 18.75%), and (2) correct implicit negation (12 instances, 18.75%). The remaining cases go for: (1) Inappropriate vernacular
bilkaad (20 instances, 31.25%), (2) under-translated explicit negation (16 instances, 25%), and (3) mistranslations (three instances) and omission (one instance), together 6.25%.
Let us start with cases where
-ly negation is accounted for in Arabic (12/28 instances, 42.85%) using correct explicit negation. The following examples are illustrative:
15. | … this little kid who could barely walk … (FS) |
| … haaða-ṭ-ṭiflu-ṣ-ṣaɤiiru | ‘allaðii laa yakaadu yamšii … |
| … this-the-kid-the-little | who not hardly walk … |
| “… this little kid who can hardly walk …” |
| … هذا الطفل الصغير الذي لا يكاد يمشي … |
16. | The economics of society … was hardly such as to provoke the curiosity of |
| a thoughtful man. (ES) |
| kaanat ‘iqtiṣaadiyyaat-ul-mujtamaʻi … | takaadu laa tuθiiru fuḍuula | rajulin mufakkirin |
| were economics-the-society … | hardly not provoke curiosity | man thoughtful |
| “The economics of society … could hardly provoke the curiosity of a thoughtful man”. |
| .كانت اقتصاديات المجتمع … تكاد لا تثير فضول رجل مفكِّر |
The translators of (15) and (16) have successfully employed a negated yakadu, viz. laa yakaadu [not hardly] and takaadu laa [hardly not], which exactly capture the meaning of the -ly negation in them. One should note that yakaadu is an Arabic defective verb which translates into the negative adverb hardly when it is negated in Arabic, while it translates into almost/nearly when it is not negated, for example, كاد أن يسقط في البركة kaada ‘an yasquṭa fi-l-birkati [almost (he) that fall in-the-pool] “He almost/nearly fell in the pool”. The negated yakadu proves so useful when rendering English -ly negation.
However, -
ly negation does not seem as straightforward as (15) and (16) may suggest. While capturing the notion of negation in general, almost 58% of the Arabic renderings (16/28 instances) fail to account for the nuance inherent in -
ly negation. Instead, this kind of negation is erroneously relayed as Arabic negation that corresponds to English negation by
not, thus amounting to serious under-translations. Following are some illustrative examples:
17. | The children of large families hardly ever learn to talk to themselves aloud, … (BF) |
| lam yataʻallam ‘awlaadu-l-ʻaa’ilaati-l-kabiirati | ‘an yataḥaddaθuu | ilaa ‘anfusihim |
| not learned children-the-families-the-big | that talk | to themselves |
| bi-ṣawtin masmuuʻin … |
| with-voice audible … |
| “The children of large families did not learn to talk to themselves in an audible voice …” |
| .لم يتعلم أولاد العائلات الكبيرة أن يتحدثوا إلى أنفسهم بصوت مسموع |
18. | The Freifrau scarcely heeded her. (BF) |
| lam tubaali-l-baaruunatu | bi-haa |
| not heed-the-Baroness | with-her |
| “The Baroness did not heed her”. |
| .لم تبالي البارونة بها |
By way of illustration, in (17) the translator obliterates the subtle nuance of the negation in hardly by opting for explicit Arabic negation by lam [not]followed by the main lexical verb yataʻallam [learn], which back-translates into English negation by not, viz. “The children of large families did not learn to talk to themselves …”. To capture the negation inherent in hardly, one may need to employ a negated yakaadu viz. لا يكاد أولاد العائلات الكبيرة يتعلمون التحدث إلى أنفسهم بصوت مسموع laa yakaadu ‘awlaadu-l-ʻaa’ilaati-l-kabiirati yataʻallamuuna-t-tataḥadduθa ‘ilaa ‘anfusihim bi-ṣawtin masmuuʻin … [not hardly children-the-families-the-large learn-the-talking to themselves with-voice audible] “The children of large families hardly learn to talk to themselves in an audible voice …”. Or, alternatively, one may use a paucity adverb like naadiran maa and qalamaa [rarely], which both inhere the nuance that “the circumstances in which those children live hardly allow them to talk to themselves”, viz. نادراً ما/قلما يتعلم أولاد العائلات الكبيرة التحدث إلى أنفسهم بصوت مسموع naadiran maa/qallamaa yataʻallamu ‘awlaadu-l-ʻaa’ilaati-l-kabiirati at-tataḥadduθa ‘ilaa ‘anfusihim bi-ṣawtin masmuuʻin … [rarely/hardly learn children-the-families-the-large the-talking to themselves with-voice audible] “Rarely/hardly (do) the children of the large families learn to talk themselves in an audible voice”.
The second strategy for rendering -
ly negation is the employment of the vernacular negative adverb
bilkaad, which is a malformed version of the standard
la yakaadu. The question is whether it is appropriate to use a vernacular form when it is possible to utilize the standard negated
yakaadu. What is surprising here is the absence of this vernacular form in
Harry Potter’s translation where the informal register may sanction it and the frequency of using it in the other works—it accounts for 31.25% of the -
ly data (20/64 instances). By way of illustration, the example in (31):
19. | I hardly know you, Augustus Waters. (FS) |
| ‘anaa bilkaad | aʻrifu-ka | yaa ‘uɤusṭus wuutarz |
| I hardly | know(I)-you | oh Augustus Waters |
| “I hardly know you Augustus Waters”. |
| .أنا بالكاد أعرفك يا أوغسطس ووترز |
can readily be rephrased naturally in standard Arabic using the negated
yakaadu as in (20) below.
20. | ‘anaa laa ‘akaadu | aʻrifu-ka | yaa ‘uɤusṭus wuutarz |
| I not hardly(I) | know(I)-you | oh Augustus Waters |
| “I hardly know you Augustus Waters”. |
| .أنا لا أعرفك أكاد يا أوغست ووترز |
Next, we have the strategy of implicit negation which accounts for 18.75% in the -
ly negation data (12/64 instances). They mainly employ
paucity or
difficulty expressions in an attempt to capture the negative nuance inherent in -
ly negation. Consider the two examples below:
21. | Auguste nowadays scarcely ever went out at all, … (BF) |
| fii miθli haaðihi-l-’ayyaami | kaana | min-an-naadiri bi-nnisbati li-’uuɤast |
| in like these-the-days | was | from-the-rarity as-regards to-Auguste |
| ‘an taðhaba xaariji-l-manzili |
| that go(she) outside-the-house |
| “In these days Auguste rarely leaves home”. |
| .في مثل هذه الأيام كان من النادر بالنسبة لأوغست أن يذهب خارج المنزل |
22. | It is asked incessantly, most of the time however hardly noticeably, … (BF) |
| su’aalun yus’alu | bi-stimraarin | raɤma ‘anna-hu yulaaḥaðu | bi-ṣuʻuubatin … |
| question be asked | with-continuity | despite that-it be noticed | with-difficulty … |
| “A question asked incessantly despite (the fact that) it is noticed with difficulty …” |
| … سؤال يسئل باستمرار رغم أنه يلاحظ بصعوبة |
In (21), the translator successfully employs a paucity expression min-al-naadiri [from the rarity] to capture -ly negation. One should note that a negated yakaadu can be readily used for that purpose, viz. في مثل هذه الأيام لا تكاد أوغست تخرج من المنزل fii miθli haaðihi-l-’ayyaami laa takaadu ‘uuɤast taxruju min-al-manzili [in like these days Auguste not hardly goes out from-the-house] “In these days Auguste hardly ever goes out”. In (22), a difficulty expression bi-ṣuʻuubatin [with difficulty] is utilized. The difficulty expression approximates rather than replicates -ly negation. A negated yakaadu would capture the meaning more closely, viz. سؤال يسئل باستمرار رغم أنه لا يكاد يلاحظ su’aalun yus’alu bi-stimraarin raɤma ‘anna-hu laa yakaadu yulaaḥaðu [question be asked with-continuity despite that-it not hardly be noticed] “A question (that is) asked incessantly, despite the fact that it is hardly noticed, …”.
To close this section, let us examine two mistranslations (out of three) which are found in the -
ly negation data (they all come from ES) below:
23. | This hardly seems like a particularly exciting subject for historical scrutiny. |
| yakaadu | haaðaa | ‘ašbaha | bi-mawḍuuʻin | muθiirin bi-wajhin xaṣṣin |
| hardly | this | like | with-subject | exciting with-face particular |
| lil-baḥθi-l-taariixiyyi |
| for-research-historical |
| “This almost seems like a particularly exciting subject for historical research”. |
| .يكاد هذا أشبه بموضوع مثير بوجه خاص للبحث التاريخي |
24. | This is particularly true when we begin at the stage of scarcely-better-than-subsistence … |
| wa-haaðaa | ṣaḥiiḥun | bi-wajhin xaṣṣin | ʻindamaa | nabda’u marḥalata |
| and-this | true | with-face particular | when | begin(we) stage |
| maa yakaadu yaziidu | ʻalaa | mujarradi-l-ʻayši … |
| which hardly more | than | mere-the-livelihood … |
| “This is particularly true when we begin the stage which is almost beyond mere livelihood …” |
| … وهذا صحيح بوجه خاص عندما نبدأ مرحلة ما يكاد يزيد على مجرد العيش |
To explain, the translator in (23) wrongly uses the affirmed rather than the negated yakaadu, which is an approximating rather a negating marker (i.e., here it communicates the message that “X is almost Y”). To capture -ly negation, the translation should read لا يكاد هذا يبدو صحيحاً لموضوع مثير بوجه خاص للبحث التاريخي laa yakaadu haaðaa yabduu šabiihan li-mawḍuuʻin muθiirin bi-wajhin xaṣṣin lil-baḥθi-l-taariixiyyi [not hardly this seems like with-subject exciting with-face particular with-the-research-historical] “This hardly seems like a particularly exciting subject for historical research …”. In (24), the translator also fails to employ a negated yakaadu, perhaps misguided by the presence of maa, which coincides, in form, with a negative particle, but is used as a relativizing marker in this sentence. To correct this mistake, the negative particle laa needs to be inserted before yakaadu in order for the Arabic translation to read وهذا صحيح بوجه خاص عندما نبدأ مرحلة ما لا يكاد يزيد على مجرد العيش wa-haaðaa ṣaḥiiḥun bi-wajhin xaṣṣin ʻindamaa nabda’u marḥalata maa laa yakaadu yaziidu ʻalaa mujarradi-l-ʻayši …, which corresponds to “This is particularly true when we begin the stage which hardly goes beyond mere livelihood …”.
5.4. Translating Determiner Negation
English determiner negation by little and few furnishes an utterance with a negative orientation just like too and -ly adverbial negation. They may also be used as adjectives to denote their lexical meaning by indicating smallness in size and number, respectively, which corresponds to ṣaɤiir [small/little] and qaliil [little/few] in Arabic. For example, there is not much beyond their semantics in there are little children playing in the garden and The next few years will be prosperous. However, little and few are often employed as negative determiners that contrast with their positive counterparts a little and a few. Compare “There is little time for discussion” لا يوجد إلا القليل من الوقت للنقاش laa yuujadu ‘illa-l-qaliilu min-al-waqti li-n-niqaaši [not exist except-the-little from-the-time for-the-discussion] with “There is a little time for discussion” يوجد بعض الوقت للنقاش yuujadu baʻḍu-l-waqti li-n-niqaaši [exist some-the-time for-the-discussion] and “There are few mistakes in the report” لا يوجد إلا القليل من الأخطاء في التقرير laa yuujadu ‘illa-l-qaliilu min-al-’axṭaa’i fi-t-taqriiri [not exist except-the-few from-the-mistakes in-the-report] with There are a few mistakes in the report يوجد بعض الأخطاء في التقرير yuujadu baʻḍu-l-’axṭaa’i fi-t-taqriiri [exist some-the-mistakes in-the-report]. While little and few color the utterances with a negative orientation, a little and a few color it with a positive orientation, hence the different Arabic renderings. The discussion in this section aims to show to what extent Arabic translators are aware of this subtle type of negation.
5.4.1. Determiner Negation by Little
Table 5 below displays the translation strategies that the translators have followed in dealing with
little determiner negation.
Out of the 99 extracted examples featuring little, 86 (86.87%) are found to involve a negative orientation that goes beyond its denotative (dictionary) meaning. The examination of little negation data shows that the translators’ attempt to handle this kind of negation involves four strategies: proper explicit negation (33 instances, 38.37%), under-translated explicit negation (18 instances, 20.93%), proper implicit negation (seven instances, 8.14%), erroneous implicit negation (25 instances, 29%), and mistranslation (one instance (1.16%).
To start with the first category, where determiner negation is rendered by explicit negation, which is the most frequent (33 instances, 38.37%), one can notice two main translation strategies. The first (23 instances/69.70%) usually employs explicit negation, with the exception particle
‘illaa or
siwaa [except] followed by a paucity-derived word, for example,
‘illaa qaliilan [except little/few] or
siwaa-l-qaliila [except-the-little/few]. Consider the following example:
25. | We all know that there’s very little time. (TH) |
| kullu-naa naʻrifu | ‘anna-hu | laysa ladaynaa siwaa qaliilun min-al-waqti |
| all-we know | that-it | not have-we except little from-the-time |
| “We all know that we don’t have except little time” (i.e., “we all know that we only have little time”). |
| .كلنا نعرف أنه ليس لدينا سوى قليل من الوقت |
The second translation strategy, which claims (10 instances, 30.30%), employs a negated antonym, for example,
lan … al-kaθiira [not … a lot/much], as is shown in the following example:
26. | And all with very little effort on your part, I assure you. (HP/5) |
| wa-lan yakuuna | ʻalay-ka ‘an tabðula-l-kaθiira min-al-waqti | ‘uṭam’inu-ka |
| and-not be | on-you that spend-the-much from-the-time | assure (I)-you |
| “And you won’t have to spend much time, I assure you”. |
| .ولن يكون عليك أن تبذل الكثير من الوقت أطمئنك |
One should note that Arabic can also employ
belittling expressions such as
laa yuðkar [not to be mentioned] or
laa yastaḥiq-uð-ðikr [not worth mentioning] to capture the meaning of
little-negation in examples like (26), which can be rephrased in (27):
27. | wa-lan yakuuna | ʻalay-ka baðlu juhdin yastaḥiqqu-ð-ðikra | ‘uṭam’inu-ka |
| and-not be | on-you making effort worth-the-mentioning | assure(I)-you |
| “And you won’t have to make (any) effort worth of mentioning, I assure you”. |
| .ولن يكون عليك بذل جهد يستحق الذكر أطمئنك |
The second translation strategy (18 instances, 20.93%) includes cases where the translator succeeds in recovering Arabic explicit negation but, unfortunately, misses the focus of determiner negation (i.e., he/she under-translates this subtle type of negation). By way of illustration, witness the following example:
28. | The truth was, I had very little idea how dangerous things were. (TH) |
| fi-l-ḥaqiiqati | lam | ‘akun ‘aʻrifu madaa xuṭuurati-l-’amri |
| in-the-fact | not | be know(I) extent dangerousness-the-situation |
| “In fact, I didn’t know how dangerous the situation was”. |
| .في الحقيقة لم أكن أعرف مدى خطورة الأمر |
As can be seen,
little-negation in (28) is rendered in Arabic to what corresponds to
not negation in English, thus missing the nuance of this type of negation. To capture this nuance, the Arabic rendering should employ an
exception expression along with explicit negation as in (29):
29. | fi-l-ḥaqiiqati | lam ‘akun ‘aʻrifu | ‘illa | ‘aqali-l-qaliili | ʻan |
| in-the-fact | not be know(I) | except | smallest-the-small | about |
| madaa xuṭuurati-l-’amri |
| extent dangerousness-the-situation |
| “In fact, I knew only little about how dangerous the situation was”. |
| .في الحقيقة لم أكن أعرف إلا أقل القليل عن مدى خطورة الأمر |
The next two strategies involve the translator’s attempt to render
little negation by implicit negation. The outcome is far from being impressive: Only seven cases (8.14%) may be considered successful in implementing this strategy, while 28 cases (32.50%) falter in this respect. Following are two examples where the first succeeds in relaying
little negation (30), while the second falters (31):
30. | “You know that Father punishes you very little”, said Sidonie coaxingly. (BF) |
| ‘anta taʻrifu yaa birnaard ‘anna ‘abii naadiran maa yuʻaaqibuka … |
| you know oh Bernard that father(my) rarely punish(you) |
| “You know Bernard that my father rarely punishes you …” |
| … أنت تعرف يا بيرناد أن أبي نادراً ما يعاقبك |
31. | Very little has changed with her health. (TH) |
| laqad | ṭara’a | taḥassunun ʻalaa ṣiḥḥati-haa |
| ? | happened | improvement on health(her) |
| “Her health has improved”. |
| .لقد طرأ تحسن على صحتها |
In (30), the translator employs the paucity expression
naadiran maa [rarely] to capture the meaning of
little negation. By contrast, the Arabic rendering in (31) embraces a positive orientation towards the referent’s health conditions, which runs counter to the negative orientation in the English utterance. To capture this orientation, the Arabic rendering may be rephrased using explicit negation along with a
belittling expression as in (32) below:
32. | lam yaṭra’ | taḥassunun | yastaḥiqqu-ð-ðikra | ʻalaa | ṣiḥḥati-haa |
| not happened | improvement | worth-the-mentioning | on | health-her |
| “There wasn’t any improvement worth mentioning about her health”. |
| .لم يطرأ تحسن يستحق الذكر على صحتها |
5.4.2. Determiner Negation by Few
Table 6 below presents the frequency and percentage of translation strategies which the translators have employed in rendering
few determiner negation.
Negation by
few is the least frequent in the corpus. Out of 93 extracted examples involving the employment of
few, only 20 are found to furnish the English utterance with a negative orientation. In the rest of the examples, the determiner
few reflects its dictionary meaning, which corresponds to
qaliil [few] in Arabic, without any coloration of negation. In such cases, the rendering of
few into Arabic is straightforward, as no negation is to be accounted for. The following example is illustrative:
33. | The next few weeks is real important for Mae Mobley. (TH) |
| kaanat-il-’asaabiiʻu-l-qaliilatu-t-taaliyatu | haamatan jiddan binnisbati | ‘ilaa | maw muublii |
| were-the-weeks-the-few-the-next | important very as regards | to | Mae Mobley |
| “The next few weeks were very important for Mae Mobley”. |
| .كانت الأسابيع القليلة التالية هامة جدا بالنسبة إلى ماو موبلي |
However, when
few is employed as a negative determiner, which is meant to express the producer’s unfavorable attitude towards the state of affairs in question, explicit Arabic negation may be needed. The data shows that explicit negation has been correctly employed in seven instances (35%), while implicit negation is erroneously employed in 12 instances (60%), and only once correctly (5%). The two examples below are representative of the success and failure in rendering
few negation:
34. | They stopped at no inns, and exchanged very few words. (BF) |
| lam yatawaqqafaa | ʻinda | ‘ayyi ḥaanatin | wa-lam yatabaadalaa | ‘illa-l-qaliilu |
| not stopped(dual) | at | any inn | and-not exchanged | except-the-few |
| min-al-kalimaati |
| from-the-words |
| “They didn’t stop at any inns and didn’t exchange but few words”. |
| .لم يتوقفا عند أي حانة ولم يتبادلا إلا القليل من الكلمات |
35. | And there are as yet few business models for journalism originating on the Internet. (GC) |
| wa-hunaaka | ḥatta-l’aan | ʻadadun qaliilun | min namaaðiji-l-’aʻmaali-ṣ-ṣaḥafiyyati |
| and-there | till-now | number small | from models-the-businesses-the-journalism |
| an-naaši’ati | ʻala-l-intarnit |
| the-new | on-the-Internet |
| “There are as yet a small number of new business models on the Internet”. |
| .وهناك حتى الآن عدد قليل من نماذج الأعمال الصحفية الناشئة على الإنترنت |
As is clear in (34) above, the translator has duly accessed explicit negation along with an
exception expression in Arabic to account for
few negation. By contrast, the translator has failed to invest explicit negation and, consequently, opts erroneously for an affirmative utterance in (35). The competent reader can readily feel the missing negative orientation in the rendering due to the translator’s failure to capture the pragmatics of
few negation. To remedy this situation, explicit Arabic negation along with an
exception expression may be accessed to furnish a negative orientation (36 below):
36. | wa-laa | yuujadu | ḥatta-l’aan | ‘illaa ʻadadun qaliilun min namaaðiji-l-’aʻmaali-ṣiḥḥati-haa |
| an-laa | exist | till-now | except number small from models-the-businesses |
| ṣ-ṣa ḥafiyyati | an-naaši’ati | ʻala-l-intarnit |
| the-journalism | the new | the-journalism |
| “There aren’t as yet but few new business models on the Internet”. |
| .ولا يوجد حتى الآن إلا عدد قليل من نماذج الأعمال الصحفية الناشئة على الإنترنت |
6. Discussion
In response to the first research question concerning how Arabic translators tackle too adverbial negation (which is formally missing in Arabic) and what translation strategies they employ, the study shows that they resort to explicit negation and implicit negation as two general strategies. This clearly proves that the recovery of Arabic explicit negation is an effective translation strategy in dealing with too negation. Within explicit negation, three strategies are employed: unpacking by coordination, nominalization, and indicating degree of attribute.
Unpacking by coordination turns out to be a very effective translation strategy for rendering too negation. The translators have successfully managed to explicate the English negation encapsulated in the negative marker too by using a consequential coordinate clause involving explicit negation. In this way, for example, the too negation in Harry was too deeply asleep to hear her is successfully relayed into an Arabic rendering that back-translates into Harry was deeply asleep, so he didn’t hear her, which is an alternative English textualization that employs generic explicit negation. Arabic translators; therefore, need to be fully aware of this workable strategy when dealing with too negation.
Resorting to explicit negation by nominalizing the English verb in a simple or complex Arabic structure also proves to be a workable translation strategy for rendering too negation. For example, the translator has managed to recover Arabic generic negation by nominalizing the English verb save into the Arabic verbal noun ‘inqaaði “saving”, in I was too late to save the girl, into a simple Arabic structure which literally back-translates into *I couldn’t saving the girl. Similarly, the translator has successfully nominalized the verb drive away, in but his Patronus was too feeble to drive the dementor away, in the English infinitive clause, into ‘ibʻaadi (driving away) in an Arabic complex structure (which back-translates into a workable English textualization, viz. but his Patronus wasn’t strong enough to drive the dementor away).
Indicating the degree of the attribute in question, for its part, presents itself as a very useful translation strategy when rendering too negation by explicit Arabic negation. It corresponds to an awkward English textualization that may paraphrase too negation. For example, the Arabic idiomatic rendering of Professor Trelawney seemed too tipsy to have recognized Harry back-translates into the awkward English paraphrase Professor Trelawney was drunk to the extent that she didn’t recognize Harry.
There are few cases when the translator’s recovery of Arabic generic negation does not convey the nuance of too negation properly. For example, the too negation in but Harry was too used to this to care is rendered into Arabic explicit negation that back-translates into but Harry didn’t care about this, thus doing away with the shade of meaning inherent in too negation. To capture this shade of meaning, the translator could have indicated the degree of the attribute in question by offering an Arabic rendering that back-translates into the English paraphrase but Harry was used to this to the extent that he didn’t care or, alternatively, the translator could have unpacked too negation by coordination by offering what back-translates into Harry was used to this, so he didn’t care.
As for implicit negation, it is not as successfully employed as explicit negation. However, this does not mean that implicit negation is not a viable option; it just suggests that it needs to be utilized more carefully by calling up three strategies: use of comparative form, use of negative, and indication of degree of attribute, which prove to be effective in rendering too negation.
The employment of the comparative form idiomatically renders many cases of too negation. For instance, the English too negation in because it was too long to memorize is correctly relayed into Arabic to what back-translates into an English workable textualization, viz. because it was longer than it could be memorized. Similarly, the use of negative verbs/verbals may appropriately capture too negation. The English too negation in Anna dies or becomes too ill to continue writing it, for example, lends itself to translating by employing a negative verb whose semantics takes care of negation, viz. the Arabic rendering back-translates into Anna died or reached a degree of illness that prevented her from continuing writing. Likewise, employing set formulas to indicate the degree of the relevant attribute succeeds in capturing too negation implicitly. Notice how the Arabic rendering, whose back-translation is the fines for violations were so small that they lost their effect, proves to be an idiomatic translation of the too negation in Fines for violations were too small to be effective.
Despite the successful employment of these strategies, the translators have failed to use implicit negation correctly in about one third of the cases. This makes the strategy of implicit negation more challenging than that of explicit negation. Therefore, the translator’s first strategy should be to consider explicit negation, and he/she needs to exercise utmost care when opting for implicit negation. For example, the too negation should not be confused with the negation-free intensifier jiddan “very”. In this way, the too negation in It is still far too cold to undress at night should not be rendered in Arabic as لايزال الطقس بارداً جداً لخلع الثياب في الليل laayazaalu-ṭ-ṭaqsu baaridan jiddan li-xalʻi-θ-θiyaabi fi-l-layli “*The weather is still very cold to take off clothes at night.” To use Arabic implicit negation properly, the translator could have employed an Arabic negative verb whose semantics takes care of too negation, viz. كان الجو البارد يمنعنا من خلع ملابسنا في الليل kaan-al-jawwu-l-baaridu yamnaʻu-naa min xalʻi malaabisi-naa fi-l-layli “The cold weather prevented us from taking off our clothes at night.”
To address the second research question concerning the translation strategies in rendering -ly negation, the data shows that the successful use of explicit negation and implicit negation accounts for only (37.50%) in the corpus. The rest of the cases goes for the vernacular bilkaad (31.70%), which is ungrammatical, and mistranslations/under-translations/omission (31.75%). This clearly shows how problematic rendering -ly negation into Arabic is. To render this type of negation properly, translators into Arabic need to be made aware of the negated Arabic verb yakaadu when employing explicit negation, as well as paucity/difficulty expressions when opting for implicit negation. Only then can the pragmatics of -ly negation be captured in Arabic.
Arabic negation by explicit negative particles emerges as the most common translation strategy for rendering the -ly negative adverbs, which clearly indicates the translators’ awareness of their negative orientation. However, the coding of this orientation in Arabic seems to be a challenging task. In fact, more than half involves under-translating this adverbial negation by rendering it into what corresponds to negation by not in English. In this way, the subtle nuance of this type of negation is lost in translation.
To employ explicit negation properly when dealing with -ly negation, the translator needs to access the Arabic negated verb yakaadu, which does not seem to be an easy task. Apparently, Arabic translators more often than not fall in the trap of under-translation by offering English generic negation by not, viz. The Barn didn’t heed her (back-translation of Arabic translation) for Frefrau scarcely heeded her. In this way, they fail to call up the appropriate negated yakaadu, viz. لم تكد البارونة تبالي بها lam takad al-baaruunatu tubaali bi-haa “The Baroness scarcely/hardly heeded her”. Alternatively, they are erroneously attracted by the vernacular bilkaad, which is a negative adverb commonly used in most Arabic vernaculars for this type of negation. This option reflects the translator’s deficient knowledge of Standard Arabic, which has its standard version (the negated yakaadu verb) for the vernacular bilkaad. Arabic translators; therefore, need to be cautioned against rendering ly-negation into the generic not negation as well as the use of vernacular bilkaad.
Arabic implicit negation may also be considered when translating -ly negation. When implementing this translation strategy, the translator needs to call up Arabic paucity expressions such as naadiran maa/qallamaa “rarely” or the difficulty expression bi-ṣuʻuubatin which approximates than replicates -ly negation implicitly. Though implicit negation is used much less frequently than explicit negation where several errors are made, it proves to be very appropriate in some cases, especially when the translator utilizes paucity expressions.
To respond to the third research question regarding the rendering of little determiner negation, results show that the success rate is less than 50%, which clearly indicates how problematic little negation is when relaying it into Arabic.
The most effective translation strategy when employing explicit negation is to use exception expressions combined with paucity words, which can properly capture this kind of subtle negation. Arabic translators; however, need to be cautioned against falling in the trap of under-translating little negation into what corresponds to English not negation, viz. The truth was, I had very little idea how dangerous things were may wrongly be translated into an Arabic rendering that back-translates into In fact I didn’t know how dangerous the situation was, a mishap which belies several cases of little negation. To capture little negation here, the translator needs to use explicit negation along with an exception expression, viz. في الحقيقة لم أكن أعرف إلا أقل القليل عن مدى خطورة الأمر fi-l-ḥaqiiqati lam ‘akun ‘aʻrifu ‘illa ‘aqalli-l-qaliili ʻan madaa xuṭuurati-l-’amri, which back-translates into In fact I knew only very little about how dangerous the situation was.
Negated antonyms may also be used to relay little negation explicitly, though at a lesser degree of success, viz. And all with very little effort on your part, I assure you relatively corresponds to the Arabic rendering that back-translates into And you won’t have to spend much effort, I assure you. However, the employment of explicit negation with a belittling expression such laa yastaḥiqu-ið-ðikr “not worth mentioning” can be a more effective strategy for capturing the shade of meaning in little negation than using explicit negation with a negated antonym, viz. ولن يكون عليك أن تبذل الكثير من الوقت أطمئنك wa-lan yakuuna ʻalay-ka ‘an tabðula-l-kaθiira min-al-waqti ‘uṭam’inu-ka, which back-translates into And you won’t have to make a lot of effort, I assure you. Consequently, the Arabic translator needs to be alerted to the importance of using exception expressions as well as belittling expressions when using explicit negation in rendering little negation.
As for implicit negation, which is employed in only few cases, it resorts to expressions inherently marked for paucity such as naadiran maa/qallamaa “rarely” independently of negative particles. The bulk of cases; however, indicates that the translators are only little aware of the negative orientation furnished by little negation. In fact, the erroneous translations furnish a positive rather negative orientation of little negation. For example, the rendering of Very little has changed with her health in لقد طرأ تحسن على صحتها laqad ṭara’a taḥassunun ʻalaa ṣiḥḥati-haa into Her health has improved has completely deprived the Arabic translation of the negative orientation. It is of utmost importance; therefore, to alert Arabic translators to the need to recover explicit negation along with paucity or belittling expressions to capture little negation. In this way, the above example should be rendered as لم يطرأ تحسن يذكر على صحتها lam yaṭra’ taḥassunun yuðkaruʻalaa ṣiḥḥati-haa, which back-translates into There wasn’t any improvement worth mentioning with her health.
Finally, in response to the fourth research question, the results show that few determiner negation proves to be so problematic in Arabic translation. This finding clearly points to the challenging subtlety of few negation which needs to be brought to the consciousness of Arabic translators who are supposed to be aware of the difference between few as a negative operator and few as a mere determiner. Apparently, the translators are not aware of the need to recover explicit Arabic negation along with an exception expression to relay the negative attitude encapsulated in this type of subtle negation. The attempt to employ implicit negation has failed except in one instance, which clearly indicates that explicit rather than implicit negation is the appropriate translation strategy to use when rendering few negation. Translators into Arabic; therefore, need to be sensitized to the nature of few negation and the appropriate strategies that may be used to render its pragmatics.