Abstract:
This study examines the use of the cohesive devices in an Arabic/English parallel corpus of the
United Nations texts and compares and contrasts them to identify cohesion shifts in the English
translations, to justify their occurrences and to find out how they are interpreted as equivalents.
Two research hypotheses are tested. The first one is that since each language employs its own
cohesive devices, English and Arabic would reveal differences in both their types and the
frequency of their use, which would considerably affect translation. The second one is that
because Arabic and English belong to two different language families, many differences would
appear in translation. Therefore, shifts of Arabic cohesive devices would occur. These shifts
would probably be due to the translators’ intention to meet the accuracy, transparency and
formality of the UN texts. With the help of corpus linguistics, a quantitative method and a
qualitative descriptive one are employed to demonstrate the extent to which source language
norms influence the use of cohesive devices in translation and to describe the accuracy of the
translation of these devices and how translators cope with the differences. Findings reveal that
Arabic and English have more similarities than differences in terms of the cohesive devices
used but there are significant differences in the frequency of their occurrence. The similarities
are significantly preserved for the purpose of accurateness, transparency and formality that
characterise the language of legal texts. The results also show that English translated texts have
a major tendency towards both explicitation and implicitation. This is demonstrated in the
occurrence of three types of shifts, namely addition, omission and substitution of the cohesive
devices used. Some extracts from the Parallel Corpus of the United Nations Texts are used for
the purpose of designing Data-driven Learning activities in translation classes
