ISSN 1799-2591

Theory and Practice in Language Studies, Vol. 14, No. 1, pp. 71-78, January 2024
DOI: [Link]

Investigating Problems Related to the Translation

of Idiomatic Expressions in the Arabic Novels
Using Neural Machine Translation
Abdullah S. Aldelaa
Language Center, The Hashemite University, Zarqa, Jordan

Mai Abdel karim Mohammad Malkawi

Language Center, The Hashemite University, Zarqa, Jordan

Abstract—NMT systems encounter many challenges and problems when rendering Arabic idiomatic
expressions into English. The current research tackled some examples selected from Arabic corpora and their
English translations generated by the selected NMT systems and the English versions of the selected corpora.
The research presents a qualitative analysis of the selected samples from three novels (The Season of
Migration to the North, the Fall of the Imam and the Girls of Riyadh). The research is mainly focusing on the
problems resulted when rendering Arabic idiomatic expressions into English.

Index Terms—NMT, idiomatic expressions, Arabic novels, translation problems

I. INTRODUCTION
An idiomatic expression is defined according to the Longman Dictionary of Contemporary English (1998) as “a
sequence of words which has a different meaning as a group from the meaning it would have if you understand each
word separately”. Idioms, therefore, cannot be broken up into their constituent elements because they are fixed
expressions (Cowie & Mackin, 1975; cited in Balfaqeeh, 2009). Idiomatic expressions play a key role in human
communication. They are emotive, as well as full of cultural connotations, that facilitate the expression of meaning at
the linguistic, as well as cultural levels. The literal meaning of fixed expressions does not make sense on the TL if it is
taken as individual words and it is hard to understand these fixed expressions unless you have learned or heard them
before.
The structure of the idioms differs from the structure of normal direct language. There are two kinds of structures
related to Arabic idiomatic expressions. The first structure does not have specific structural characteristics that serve as
the identifying features. Normal sentences or phrases that have meaning, which is directly understood from their
constituents, are included in this common category, for example, ‘Ali plays football’. The second structure cannot be
directly identified; only through structural, as well as semantic characteristics (Al-Anbar, 2001), for example ‫ﺥﻑﻱ‬ ‫ (ﻉﺍﺩﺏ‬,
‫ ُﺡﻥﻱﻥ‬He came back with Hunayn’s shoes, which means that his efforts came to naught). Idiomatic expressions such as
)‫ﻥﺽﺏ ﻡﺍء ﻭﺝﻩﻩ‬ (and its literal meaning is ‘his face water ran out’, which means that he is no longer ashamed of himself.
Such characteristic often poses difficulties when idiomatic expressions are translated from the SL into the TL (Cowie &
Mackin, 1975; Al-Qassimi, 1979; Husam Al-Din, 1985; Al-Hamzawi, 2000; Balfaqeeh, 2009).

II. LITERATURE REVIEW

Studies have investigated various types of idiomatic expressions in MSA (Husam Al-Din, 1985; Abu-Saad, 1987;
Kharma & Hajjaj, 1989; Quwaydar, 2000; Faied, 2003; Ghazala, 2003). Idioms are divided into two basic types:
complex idioms and simple idioms (Husam Al-Din, 1985). Complex idioms comprise more than two words while
simple idioms comprise two words or less. He sub-categorized the complex idioms into: verbal, nominal, and fixed, e.g.,
the comparative like ‫ﺍﻙﺭﻡ ﻡﻥﺡﺍﺕﻡ‬, which is literally translated into ‘more generous than Hatem’, doubled (a rhyming
phrase, e.g., ‫ﺏﻱﺍﻙ‬‫)ﺡﻱﺍﻙ ﷲ ﻭ‬, and follow-up (in which the second word qualifies the first, for example, ‫ ﻡﻠﺢ ﺍﺝﺍﺝ‬, literally
‘salt water’). Simple idioms are classified into six types: collocation, additive, titled expressions which combine ‫ﺃﺏ‬
father or ‫ ﺃﻡ‬mother with a noun, for example, ‫( ﺃﻡ ﻉﻮﻑ‬literally, mother of Awf meaning the locust), titled expressions
combine ‫ﺏﻥﺖ‬daughter or ‫ ﺍﺏﻥ‬son with a noun, for example, ‫( ﺍﺏﻥ ﺃﺭﺽ‬literally translated into ‘son of earth’) meaning ‘a
stranger’, dual expressions, for example, (‫ ﺍﻟﺜﻘﻼﻥ‬peoples and Jinn, referring to mortals and devils), expressions based on
metonymy, for example, ‫ ﻓﻼﻥ ﺍﻳﺔ‬outstanding, which refers to someone who is skillful in his specialization (Husam Al-
Din, 1985, p. 223).
Idioms are classified based on the syntactic structures into: genitive idiomatic expressions such as ‫ﺻﻑﺭ ﺍﻟﻱﺪﻳﻥ‬empty-
handed, adjectival idiomatic expressions e.g. ‫ ﺍﻟﺱﻮﻕ ﺍﻟﺡﺭﺓ‬duty free, verbal idiomatic expressions ‫ ﻳﻤﺱﺢ ﺍﻟﺠﻮﺥ‬he
compliments somebody (an apple polisher), nominal idiomatic expressions, e.g., ‫ ﺍﻟﺥﺭﻭﺝ ﻡﻥﻉﻥﻕ ﺍﻟﺯﺝﺍﺝﺔ‬bottleneck, and

© 2024 ACADEMY PUBLICATION

72 THEORY AND PRACTICE IN LANGUAGE STUDIES

phrasal (prepositional and adverbial) idiomatic expressions ‫ ﻡﻥﺕﺡﺖ ﺍﻟﻁﺍﻭﻟﺔ‬under the table (Faied, 2003). Some problems
resulted from the translation of idiomatic expressions from English into Arabic in Reuters news agency are investigated
(Khalaf, 2010). He highlighted the most appropriate method for translating idiomatic expressions. The corpus of his
study consists of 50 newspaper articles that are taken from Reuters news agency. The study concluded that the level of
language manipulation, retrieval of standard equivalents from the linguistic system and the level of interpreting style are
appropriately used in translating the idiomatic expressions. He recommended that translators should choose a proper
method when doing translation according to the levels of language manipulation, and should be fully aware of and
comprehend the various aspects of the idiomatic expressions of both languages (as cited in Aubed, 2012).
The difference and variance of idiomatic expressions among languages can cause problems for translating them (Al-
Shawi & Mahadi, 2012). Therefore, if the gap between the source language and the target language is big, then
transferring the message from SL into TL will be more difficult. They mentioned many factors involved in translating
idiomatic expressions like social and religious cultures. The study concluded that the translator’s knowledge of a
specific culture can contribute in many cases to detecting the meaning of the idiom, especially the one that has non-
literal meaning because that idiom demands more than any other feature of language that the translator is not only
accurate but highly sensitive to the rhetorical hints of the language. The study proposed some strategies that can help in
translating idioms like using a parallel idiom in the TL, using similar meaning but a different form, and using notes.
These strategies can offer some solutions and guidelines for translators when translating idioms (as cited in AL-Shawi,
2012).
Idioms in the Saudi press are investigated from a structural perspective by examining their structural patterns and
their internal and external grammatical relations (Alqahtani, 2014). The sample of the study consists of 440 idioms that
appeared in Al-Riyadh newspaper. The study focused on Arabic syntax and grammatical structures. When examining
structural variations, the study used fixedness and compositional/non-compositional techniques. He applied Halliday
and Hasan’s model of cohesiveness during the analysis of this feature in the idiomatic expressions. The study concluded
that half of the overall structures were verbal patterns. Moreover, idioms were found to display the same structural and
grammatical relations as other linguistic units. The study found that verbal structures form 47.6% (nearly half) of the
overall structural patterns, genitive structures 17.2%, preposition phrase patterns 8.4%, and adjectival patterns 7.7%.
The findings demonstrate that the sample's idioms varied in their structure, changing from one pattern to another. There
are also some difficulties which face the students when translating idiomatic expressions (Ali et al., 2016). The study
showed that the differences between the Arabic and English language, i.e., syntactic structure, culture-bound idiomatic
expressions, the social and religious cultures, etc. play a vital role in these difficulties. They come to the result that
literal translation could sometimes convey the meaning of the idiom. This finding is not consistent with Howwar’s
(2013) claim that idioms can never be translated literally. In fact, a literal translation of the idiom ‘you scratch my back
and I’ll scratch yours’, and ‘flog a dead horse’ did not produce meaningless translations. The Arabic reader is still able
to grasp the meaning although the effect is not the same. Nevertheless, in some cases, the literal translation was not
effective as in the case of ‘break the ice’, ‘it is raining cats and dogs’, and ‘kick the bucket’ since it led to ambiguity.
They recommended some strategies in order to handle these difficulties such as paraphrasing, literal translation and
translation by omission.

III. METHODOLOGY
The sample of the current study consists of 158 Arabic idiomatic expressions extracted from three literary texts
Mawsim al-Hijrah ilâ al-Shamâl (Season of Migration to the North) by Tayeb Saleh, Banat AlRiyadh (Girls of
AlRiyadh) by Rajaa Alsanea and Suqut Al-lmam (The Fall of the Imam) by Nawal El Saadawi. This study is
unidirectional study and, therefore, the data will be translated from Arabic into English using three NMT platforms
including Bing (Microsoft translator), Facebook and Google translate. These systems generate 474 translations which
will be compared with human translation, which is provided by three English versions of the selected novels. The
purposeful sampling strategy is used in order to select the most relevant examples, which reflect the problems. Black
box evaluation approach is followed which means that only the input and outputs of the NMT software are subject to
assessment while the internal mechanisms of NMT process are ignored. This leads us to evaluate the machine
translation output as any user or professional translator would experience the situation.

IV. RESULTS AND DISCUSSION

After the translation process was performed using the selected systems for the study, it was found that among 474
translations of the idiomatic expressions, NMT systems have successfully rendered 156 idioms into English. On the
other hand, they failed to render 318 idioms. The following table represents the number and percentage of successful
and failed translations of the idiomatic expressions extracted from the selected novels which form the corpus of the
study.

© 2024 ACADEMY PUBLICATION

THEORY AND PRACTICE IN LANGUAGE STUDIES 73

Validity of translation Frequency Percentage

Successful translations 156 33%
Failed translations 318 67%
Total 474 100%

The success percentage is approximately close between Bing and Google, whereas FB translator achieves low
percentage compared to the other two systems. The following chart shows the success percentage for all systems.

Although NMTs have successfully rendered some idioms into the TL, they faced some problems that can be
summarized as follows:
Problems Season of The fall of imam Girls of AlRiaydh Total Percentage
migration to the
north
Unwanted literal 44 53 42 139 44%
translation
Unjustified omission 12 15 18 45 14%
Wrong lexical 12 8 7 27 8.5%
prediction
Ambiguity resulted 9 6 12 27 8.5%
from H and P
Culture bound 9 3 10 22 7%
expressions
Lack of cohesion 6 7 8 21 6.5%
Wrong choice of 7 4 6 17 5%
words
Colloquial idiomatic 1 0 14 15 4.5%
expressions
Others 2 1 2 5 2%
Total 102 97 119 318 100%

© 2024 ACADEMY PUBLICATION

74 THEORY AND PRACTICE IN LANGUAGE STUDIES

A. Unwanted Literal Translation

Literal translation is referred to in Cambridge Online Dictionary as "a text is done by translating each word
separately, without looking at how the words are used together in a phrase or sentence.'' Although literal translation is
described as a useful procedure by many scholars such as Nida (1964, 1969, 1989) and Newmark (1988) who asserts
that ‘'literal translation is correct and must not be avoided, if it secures referential and pragmatic equivalence to the
original'' (Newmark, 1988, p. 68).
Thus, literal translation in many cases –as in the current study- produces a translation for the meaning out of its
context which will confuse the TL readers at the end of the day. Unwanted literal translation is the translation that is
rendered a word for word and strive to stay formally (lexically and syntactically) close to the source text as much as
possible (Arffman, 2012). There are certain expressions should not be taken literally. For instance, an idiomatic
expression “snake in the grass” means a traitor. As a translator, you can either look for an appropriate idiomatic
expression in the TL or just translate the actual meaning and not the idiom itself. Otherwise, it will sound awkward and
won’t make any sense. In literal translation, the words and phrases of the SL are translated taking no account of the
context but respecting the syntactic structure of TL. Therefore, many examples of literal translation can be meaningless
to TL readers. The following examples will illustrate how unwanted literal translation hinders the SL AFEs:
‫ﺏﺍﺥﺕﻱﺍﺭﻥﺍ ﺍﻭﺭﻍﻡ ﺍﻥﻮﻓﻥﺍ‬‫ﺕﺱﻱ‬
‫ﺍﻟﺪﻥﻱﺍﺭ‬
Ad-dunya tasῑr bixtiyārina aw raġma aunūfina
Bing: The world is going by our choice or despite our noses.
FB: The world is going by our choice or despite our noses.
Google: The world goes by choosing us or despite our noses.
HT: The world goes on whether we choose for it to do so or in defiance of us.
Nearly, all NMT systems translated the idiomatic expression literally. In other words, the phrase "despite our noses"
means nothing in English, whereas the idiomatic expression "raġma aunūfina" which means ''by force" is meaningfully
used in the Arabic text. In one way or another, it could be said that the human translation renders the idiomatic
expression successfully and equivalently into English. Another idiomatic expression ‫ ﺥﻠﺹﺖ ﻡﻥﻩ ﺍﻟﻘﺪﻳﻡ ﻭﺍﻟﺠﺪﻳﺪ‬is rendered
into English by the three systems as:
Bin: I got rid of it old and new.
FB: Finished it old and new.
Google: I got rid of it old and new.
It is clearly appeared that the three systems translate the idiomatic expression word for word while the HT for the
same idiomatic expression is "Paid him in full". Such rendering of the idiomatic expression literally from Arabic into
English without conveying the sense of the original text is misleading. In addition, the generated translations lack
cohesion and coherence.
B. Unjustified Omission
Omission may be used frequently by translators in a justified way to avoid ambiguity, taboo, and repetition, but it
may sometimes lead to incomplete ideas in the original text if it is used in an unjustified way. Newmark argues that
there are cases where the omission may be inevitable due to the different claims in each language (Newmark, 1981, pp.
7-8). Omission helps in submitting the information in a more concise way as well as in focusing on essential
information and avoiding unrequired culture, time and space bumps (Dimitriu, 2004).
Unjustified omission in NMT leads to a loss and shortage in the meaning. As deleting some words or phrases may
ignore important messages which are necessary for the text in order to be communicative. In these cases, the readers in
the TL may be deprived of many details that might help them understand the original text correctly. The following
examples illustrate this idea:
‫ﺍﺵﺕﺏﻙﺍ ﻓﻱﺱﺍﻉﺔﻥﺡﺱ‬ ‫ﺍءﺱﻤ‬
‫ﻥﻥﺍ ﻓﻠﻙﺍﻥ ﻓﻱ ﺍﻟ‬
‫ﻙﺍ‬
ka'anana fulkan fi alsama' ashtabaka Fi sāʿat an-naḥas
Bing: It's like we were in the sky.
FB: It's like we were in the sky.
Google: As if we were in the sky clashed in an hour jinx.
The whole idiomatic expression is omitted by Bing and FB. On the other hand, it is translated literally by Google.
The idiomatic expression which means in evil hour is rendered professionally in the English version of the novel "We
were two celestial bodies that had merged in an ill-omened moment of time". Furthermore, dropping the SL idiomatic
expression "Fi sāʿat an-naḥas" by the NMT systems is not justified. Another example where omission is committed by
NMTs is the following idiomatic expression from "Girls of Riyadh":
ً ‫ﺱﻤﻉ ﻡﻥ ﺍﻟﺠﻱﺭﺍﻥ ﻙﻼﻡﺍ ً ﺍﺱﺕﺵﺍﻁ ﻟﻩ ﻍﺽﺏﺍ‬
Bing: He heard from the neighbors talking about his son.
FB: He heard from the neighbors about his son.
Google: He heard from his neighbors talk about his son, who angered him.
The idiomatic expression "ً ‫ " ﺍﺱﺕﺵﺍﻁ ﻟﻩ ﻍﺽﺏﺍ‬in the Arabic version is totally omitted in the translations provided by Bing
and FB. Such omission of the idiomatic expression is considered as betrayal for the Arabic text and by doing so, NMT

© 2024 ACADEMY PUBLICATION

THEORY AND PRACTICE IN LANGUAGE STUDIES 75

deprived the English readers from an important part of the Arabic text; which is the strong feeling of displeasure,
hostility or antagonism of the father towards his family. Such idiom is well and fully rendered in the HT as: The father
heard things by way of the neighbors, though, and what he heard put him into a fury.
C. Culture Bound Expressions
Linguists have fully discussed the problems related to culture in various linguistic aspects. Also, many dealt with
translating the cultural specific expressions from either Arabic into English or vice versa. Cultural specific expressions
can be challenging to translate; even professional translators occasionally face some difficulties to translate them
equivalently. This is due to the fact that the cultural context is too vague, and it represents the world view of a society,
its beliefs, emotions and values. Cultural specific expressions are “the words, terms or expressions which refer for
concrete objects or abstract aspects that may be related to religious beliefs, social habits, customs and traditions or
social situations” (Nida, 1964, p. 90). Thus, context plays a very important role in translating cultural idiomatic
expressions. The selected corpus is rich by cultural specific idiomatic expressions, especially those which are related to
religious beliefs, social habits, and traditions. Among these expressions are the followings:
bāyaʿnāh ‫ﺏﺍﻳﻉﻥﺍﻩ‬
This idiomatic expression is translated by the three NMT systems as follows:
Bing: By our way.
FB: Seller it.
Google: Meaning it.
The failure of NMT systems appeared clearly in the renditions of the above idiom. They provide translations which
have no meaning or even any relevance to the SL idiom. Moreover, the idiom is rendered equivalently and functionally
as "we have sworn eternal loyalty for him" which is the human translation in the English version of the novel. Another
example which shows the problem which faced NMT systems when translating cultural specific expression is the
idiomatic expression ‫ﺙﻠﺜﻥﺍ ﻭﻍﺪﺍ ﺍﻟﺵﺭ‬ṯilṯna w-ġada aš-šarr which was translated by the three systems as follow:
Bing: One-third of us are evil bastards.
FB: Like and tomorrow evil.
Google: A third of us, tomorrow is evil.
The above idiomatic expression is used when something happened for the third time. The number in that idiomatic
expression is very important since it refers to something that happened peacefully for the third time and become a
source of optimism in the Arabic culture. This idiomatic expression could be replaced by an equivalent idiomatic
expression like a third time lucky in English. Unfortunately, NMT systems do not translate the above idiomatic
expression appropriately which may confuse the readers of the TL and make them perplexed about the content of the
target text. Therefore, relying on NMT systems in translating idioms which had cultural concepts into another language
would be problematic as most people are not familiar with these concepts. Thus, if there is a lexical gap, i.e. if words or
phrases are not known or if lexical equivalents do not exist in the target culture and language, such culture-specific
items cause serious challenge for NMT systems.
D. Lack of Cohesion
Cohesion refers to the text-internal relationship of linguistic elements that are overtly linked via lexico-grammatical
devices across sentences in order to be understood as a text, and it occurs where the interpretation of some elements in
the text is dependent on that of another (Halliday & Hasan, 1976). Thus, if such grammatical and linguistic links
between the sentence elements are missing, the text will no longer be a text at all. In this study there are many examples
which illustrate this problem in the NMT product, such as the following examples:
‫ﺝﺭﻉﺕﻩ ﺍﻟﻤﻩﺍﻥﺍﺕ‬
jarraʿathu al-mahanāt
Bing: His insults.
FB: His dose of dishonored.
Google: His dose of humiliations.
We notice that there are no connections between the words, on one hand, and NMT systems do not differentiate
whether the text refers for a male or female addressee on the other hand. Another issue in the translations is the overlap
related to the word jarraʿathu which used to mean "She had made him swallow such degradations", but NMT systems
translate it as a noun "His dose" in the second and third translations.
‫ﻡﺱﺍﻙﻱﻥ ﻳﻘﻁﻉﻮﻥ ﺍﻟﻘﻠﺏ‬
masākῑn ygaṭʿūn al-galb
Bing: Poor cut heart.
FB: Poor cut heart.
Google: Poor people cut the heart.
HT: They are pitiful and they make you feel sorry for them.
The three translations provided by the three systems need ties in order to make the idea clear and comprehensible to
the reader. Comparing the first and second translations offered by Bing and Facebook translators, there are instances of
repetition and missing articles. Google translation also misses some important details about the meaning. Such

© 2024 ACADEMY PUBLICATION

76 THEORY AND PRACTICE IN LANGUAGE STUDIES

translations are deemed to be failed following the seven standards of textuality set forth by de Beaugrande and Dressler
in their seminal work introduction to Text Linguistics.
E. Ambiguity Resulted From Homonymy and Polysemy
Homonyms are “lexical items that are identical in spelling and pronunciations but have different meanings,” (Yule,
2006, p. 107). The word bank for example might mean the river side or a financial institution or even a seat. While
polysemy is defined as a “situation” in which the same word has two or more different meanings, like the word 'head'
which could be a part of the body, a leading position in an institution or something else (Ullman, 1967, p. 159). They
are ambiguous words whose different senses or meanings are far apart from each other and not obviously related to each
other in any way (Farghal, 1998). Polysemy is closely connected to homonymy and it occurs when a word has more
than one meaning (Kharma & Hajjaj, 1989). Making the distinction between polysemy and homonymy is not an easy
process, according to many academics. Finch is one of them who said “it is difficult for linguists to distinguish between
polysemy and homonymy” (Finch, 2000, p. 165). Homonymy and polysemy are described as “a major source of
ambiguity which hinder the translation process” (Salim, 2013, p. 75). As a matter of fact, homonymy and polysemy
could pose a serious problem for human translator as well as NMT systems. The following examples illustrate this
problem minutely:
‫ﺡﻑﺍﻅﺍ ً ﻉﻠﻯ ﺍﻟﻉﻱﺵ ﻭﺍﻟﻤﻠﺢ‬
ḥfāẓān ʿlā ālʿyš wālmlḥ
Bing: For the sake of living and salt.
FB: To keep living and salt.
Google: In order to save life and salt.
The three systems translate the word "‫ﻉﻱﺵ‬ ‫ "ﺍﻟ‬as living, living and life respectively. The idea of the SL idiomatic
expression is totally managed and misleading in the TL. All systems understand only one meaning of the polysemic
word and neglect the other one which is the intended meaning. The concept of "bread and salt" (‫ﺥﺏﺯ ﻭﻡﻠﺢ‬or ‫ )ﻉﻱﺵ ﻭﻡﻠﺢ‬in
Arabic culture symbolizes the rapprochement and alliance between two persons when eating together. Eating bread and
salt create a moral obligation which should not be betrayed. This attitude is also expressed by Arabic phrases such as
"there are bread and salt between us". It is worth noting that such polysemic words might create a sort of difficulty even
for Human translation. For this expression, there was no translation provided for this expression in the English version
of the novel.
Another example which serves the same idea is the idiomatic expression "milḥ wa qublah ‫ "ﻡﺍﺵﺍء ﷲ ﻡﻠﺢﻭﻕﺏﻠﺔ‬which
translated by the three systems as:
Bing: God willing, salt and accept it.
FB: What Allah wished salt and before him.
Google: God wills salt and before him.
The above example shows how polysemy plays a major role in the ambiguity of the SL idiomatic expression. The
word which is being in polysemy is the word "‫" ﻕﺏﻠﻩ‬. The word "‫"ﻕﺏﻠﻩ‬denotes three ideas which are: before him
(sequence), kiss and pretty (the degree of beauty). Unfortunately, the NMT systems were able to catch two of the
meanings stated in the SL expression, and eliminate the third meaning which is the intended one in this regard.
However, HT renders that idiomatic expression successfully and eliminates the ambiguity when translating it as: Ma
shaa Allah, she is so pretty into the TL.
F. Wrong Choice of Words
Words are powerful tools when used correctly. Thus, it is important in translation to choose the right word which
convinces your TL audience. However, words can also be dangerous weapons which results in significant differences
between the SL and TL. It’s important to learn how to spot problematic words and replace them with clear, powerful
ones so that your translation can be as effective as possible. In many cases NMT uses words which mean something
different because they sound similar to the systems’ stored data. The following is one of the most common mistakes that
NMT committed while rendering AFEs into English:
‫ﺙﻠﺝ ﻕﺪﺕﻙﺱﺭ‬‫ﺕﺵﻱﺭ ﺍﻟﻯ ﺍﻥﺙﻤﺔ‬
‫ﺡﺍﺩﻳﺙﺍﻟﺕﻱ‬.‫ﻭﻍﻱﺭﻩﺍ ﻡﻥ ﺍ‬
Bing: And other conversations that suggest that there's snow could break.
FB: And other conversations that indicate that snow may be broken.
Google: And other hadiths indicating that there is snow has broken.
The right HT rendition for this expression is “the conversation indicates that problems between them are overcome
and things went in ease”.
G. Wrong Lexical Prediction
The process of lexical prediction is “the pre-activation of upcoming words -their meaning, and to some extent, their
form- during online sentence comprehension” (Kochari & Flecken, 2018, p. 1). Several examples related to this
problem have been investigated in the current study, among them are the following examples:
The idiomatic expression Zēr an-nisāʔ ‫ ﺯﻳﺭﺍﻟﻥﺱﺍء‬extracted from the season of migration to the north is rendered by
Bing, FB and Google respectively as woman's minster, woman minster and womanizer. The only one which can be

© 2024 ACADEMY PUBLICATION

THEORY AND PRACTICE IN LANGUAGE STUDIES 77

considered as correct translation is the third one (Google translation). Whereas, translations provided by Bing and FB
are far from the purpose of the SL expression. Such failure is resulted due to the system wrong lexical prediction that
might happen in the aforementioned example due to the interference between two Arabic words ‫ ﺯﻳﺭ‬Zēr and ‫ ﻭﺯﻳﺭ‬which
have a similar spelling or pronunciation except for omitting the first letter in ‫ﻭﺯﻳﺭ‬. Another example from the same
novel, is the idiomatic expression ʿišna w-raʔēna ‫ﻟﻥﺱﺍءﺕﺥﻁﺏ ﺍﻟﺭﺝﺍﻝ‬ ‫ﻉﺵﻥﺍ ﻭﺭﺃﻳﻥﺍ ﺍ‬which is translated by FB as 'we lived
behind women engaged men'. The system prossessed the word '‫ ﻭﺭﺃﻳﻥﺍ‬w-raʔēna ' as if it is a preposition (‫ (ﻭﺭﺍء‬. In fact,
the word ' ‫ 'ﻭﺭﺃﻳﻥﺍ‬consists of a coordination conjunction, a verb and an attached plural pronoun ( ‫ ﻭﺭﺃﻳﻥﺍ‬and we saw).
Accordingly, the translation is not accurate and is misleading as a result of the system wrong lexical prediction.
Additionally, Bing and FB committed the same error when rendering the Arabic idiom "Gargarat lhātha baḍ-ḍiḥik
‫ﻕﺭﻕﺭﺕ ﻟﻩﺍﺕﻩﺍﺏﺍﻟﺽﺡﻙ‬ " ,"she gave out throaty, gurgling laughs". The two systems translated the idiomatic expression
respectively, as "I read to her with laugher, and I read her with laugher". The two translations do not convey the
meaning of the Arabic idiom. The action of laughing is rendered as if it is reading action. This is might be due to the
interference between the two words " ‫ﻕﺭﻕﺭ‬ " and " ‫ ;"ﻕﺭﺃ‬which causes such wrong lexical prediction.
H. Colloquial Expressions
Colloquial expressions refer to any informal word or expression used soundly in conversations among ordinary or
educated people (Nofalli, 2012). Conveying the SL meaning is really a very important matter in the translation process.
Unfortunately, misunderstanding of the colloquial expressions by NMT leads to catastrophic outputs that lead to bad
and sometimes awful translations that could not meet the expectations of the TL readers. There are many colloquial
expressions in the selected novels especially in the Girls of Riyadh '‫ 'ﺏﻥﺍﺕ ﺍﻟﺭﻳﺍﺽ‬which is rich with these colloquial
expressions, for instance: "‫ﺍﻟ ﺭﻩﺵ‬, ‫ﺙﻘﺍﻟﺔ ﺍﻟﻁﻱﻥﺔ‬, ‫ﺍﻟﻤﺹﺍﻟﺔ‬, ‫ﺍﻟﺱﻤﺍﺝﺔ‬, ‫ﻑﻱﻥﺍﺏﻱ ﺍﻉﺭﺱ‬
‫"ﻳﻤﻩﺕﻙ‬. NMT systems were unable to render such
colloquial idiomatic expressions from Arabic into English. Let us consider the following example and how all systems
translate it:
‫ﺕﻙﻑﻱﻥﺃﺏﻱ ﺃﻉﺭﺱ‬ ‫ﻩ ﻳﻤ‬
takfῑn abi aʿras
Bing: Yamit is enough for my father to get weddings.
FB: mother shrouds my father's wedding.
Google: Blind shroud Abi wedding.
This idiomatic expression is all about "a guy appeal to his mother to get married". This idea is totally missed in the
translation provided not only by NMT systems, but also by human translation which omit the whole idioms.

V. CONCLUSION AND RECOMMENDATIONS

The translation of idiomatic expressions using NMT systems is not an easy task for these platforms, since rendering
these idiomatic expressions from SL to the TL should sound natural. In other words, the translation of idiomatic
expressions in literary texts is not just disposition or replacing of words from one language into another; and it should
not be a word for word translation. It is worth mentioning that Google and Bing translators significantly outperformed
the Facebook translator in translating idioms. Generally, although NMTs have successfully rendered some idioms into
the TL, but they still face some problems such as: unwanted literal translation, unjustified omission, culture bound
expressions, wrong lexical prediction, Lack of cohesion, Ambiguity resulted from homonymy and polysemy, wrong
choice of words, and colloquial idiomatic expressions.

REFERENCES
[1] Abu-Saad, A. (1987). Moʿjam altarakib walʿ'ibarat aliṣṭilaḥiyyah ala'rbiyyah alqdīm minha walmowllad. Lebanon: Dar al ‘lm
lilmalyīn.
[2] Al-Anbar, J. (2001). Al taʿbir al iṣṭilaḥi fi musnad Al-Imam Ahmad. Riyadh: King Saud University.
[3] Al-Hamzawi, A. (2000). Al mathal wa al taʿbir al iṣ ṭilaḥi fi altorath al ʿarabi. Available online at:
[Link]
[4] Ali, H. and Al- Rushaidi, S. (2016). Translating Idiomatic Expressions from English into Arabic: Difficulties and Strategies.
Arab World English Journal (AWEJ), pp. 187-201.
[5] Alqahtani, H. (2014). The Structure and Context of Idiomatic Expressions in the Saudi Press. PhD Thesis: University of Leeds.
[6] Al-Qassimi, A. (1979). Al taʿabir al iṣṭilaḥiyyah wa alsiyaqiyyah wamuʿjam ʿarabi laha. Journal: majallat al lisan al ʿarabi, 17,
17-34.
[7] Alsanea, R. & Booth, M. (2007). Girls of Riyadh. New York: The Penguin Press.
[8] AL-Shawi, M. and Mahadi, T. (2012). Strategies for Translating Idioms from Arabic into English and Vice Versa.
AMARABAC, Journal of American Arabic Academy for Sciences and Technology, [Link]. Volume 3, Number 6,
(2012), pp. 139-147.
[9] Arffman, I. (2012). Translating International Achievement Tests Translators' View. Jour.
[Link]
W/citation/download
[10] Aubed, M. (2012). Polite Requests in English and Arabic: A Comparative Study. Theory and Practice in Language Studies, 2(5)
pp. 916-922.

© 2024 ACADEMY PUBLICATION

78 THEORY AND PRACTICE IN LANGUAGE STUDIES

[11] Balfaqeeh, N. (2009). Strategies for Translating Idioms and Culturally-Bound Expressions Within the Human Development
Genre. MA Thesis. University of Birmingham.
[12] Cowie, P. and Mackin, R. (1975). Dictionary of Current Idiomatic English, Volume 1, Phrasal Verbs. Oxford: Oxford
University Press.
[13] Dawood, M. M. (2003). Muʿjam altaʿbir aliṣṭilaḥi, Cairo: Dar Gharib Books for Printing and Publishing.
[14] Dimitriu, Rodica. (2004). Omission in Translation, Perspectives: Studies in Translatology, 12(3), 163-175.
[15] Domyati, M. (2009). Attaʿbīrāt aliṣṭilāḥiyyah: mafhūmuhā wa dawāfiʿuhāwa maṣādiruhā wa ʾnmāṭihā attarkībiyyah. Available
at: [Link]
[16] El-Saadawi, N. (1988). The Fall of the Imam. UK: Methuen Publishing.
[17] Faied, W. K. (2003). Baᶜḍ ṣuwar al taᶜbirat al iṣṭilaḥiyyah fi alᶜarabiyyah al muᶜṣirah. Majalat majmaᶜ al loghah al ᶜarabiyyah:
Damascus, 896-916.
[18] Farghal, M and Shunnaq, A. (1998). Translation with reference to English & Arabic. Irbid: Dar Al-Helal.
[19] Halliday, M. A. K. (1985). An Introduction to Functional Grammar. London: Edward Arnold.
[20] Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman Publishing.
[21] Howwar, M. (2013). Seeking the nature of idioms: A socio-cultural study in idiomatic English and Arabic meanings.
International Journal of Scientific and Research Publications, (3)2, p. 13.
[22] HussamEldin, K. Z. (1985). Al taʿabir al iṣṭilaḥi: dirasah fi ta'ṣil almuṣṭalaḥ wa mafhūmihi, Cairo: Anglo-Egyptian Library.
[23] Khalaf, M. (2010). Investigating the problems of translating idiomatic expressions from English into Arabic in news articles
from Reuters News Agency. (Unpublished master’s thesis). Universiti Sains Malaysia, Penang, Malaysia.
[24] Kharma, N. N., & Hajjaj, A. H. (1989). Use of the Mother Tongue in the ESL Classroom. International Review of Applied
Linguistics (IRAL), 27.
[25] Kochari, A. & Flecken, M. (2018). Lexical Prediction in Language Comprehension: a Replication Study of Grammatical
Gender Effects in Dutch. Journal of Language, Cognition and Neuroscience. Routledge Taylor and Francis Group.
[26] Newmark, P. (1988). A Textbook of Translation. New York: Prentice Hall.
[27] Newmark, P. (1981). Approaches to Translation. Oxford and New York: Pergamon Press.
[28] Nida, E. and Taber, C. (1982). The Theory and Practice of Translation, Leiden: E.J. Brill.
[29] Nida, E. (1969). Science of Translation. In Anwar S. Dil (ed.), Language Structure and Thought. Essays by Eugene A. Nida.
Stanford: Stanford University Press, 1975.
[30] Nida, E. (1964). Toward a Science of Translating with Special Reference to Principles and Procedures involved in Bible
Translating. Leiden: E. J. Brill.
[31] Nofalli, T. S. (2012). Translation Method Used in Translating Colloquial Expressions in Subtitle Text of Transformers Movie.
M.A. Paper. Faculty of Humanities Dian Nuswantoro University Semarang. Retrieved July, 16, 2019 from
[Link]
[32] Quwaydar, H. (2000). Al-ʿIbārah al-iṣṭilāḥīyyah fī al-lughah al-ʿarabīyyah:māhīyatuhā, khaṣāʼiṣuhā, maṣādiruhā, aṣnāfuhā.
Damascus: Dār Kinān.
[33] Salih, T. (1969). Season of Migration to the North. London: Heinemann Education.
[34] Salim, A. (2013). Homonymy in Jordanian Colloquial Arabic: A Semantic Investigation. Canadian Center of Science and
Education.
[35] Ullman, S. (1962). Semantics: An Introduction to the Science of Meaning. Oxford: Basil Blackwell.
[36] Yule, G. (2006). The Study of Language: An Introduction. CUP, Cambridge.

Abdullah S. Aldelaa is currently a Lecturer in the Language Center–The Hashemite University. He received
his BA degree in English language and literature from Yarmouk University (YU) in Irbid, Jordan 2006. He
also received his MA in English language/ Translation from Yarmouk University 2010. He holds a PhD in
Translation Studies, granted by University of Science Malaysia, Penang, Malaysia in 2022. His research
interests are Machine Translation (MT), Computational linguistics, and Computer Aided Machine Translation
(CAMT). Email: abdullahs@[Link]

Mai A. Malkawi is currently a lecturer of English language at The Hashemite University (HU) in Zarqa, Jordan. She was born in
Irbid, Jordan. She received her BA degree of English language and literature/ French language minor from Yarmouk University (YU)
in Irbid, Jordan, followed by her MA degree in English language / translation from the same university. She has published 5 papers
so far focusing on literary translation (Arabic, English, French) and machine translation which are among her research interests. Now
she is a certified legal translator under the Arab Translators Association. Moreover, she is a member of The Jordanian Translators
Association as (Arabic, English, French) translator. Email: mmalkawi@[Link]

© 2024 ACADEMY PUBLICATION