Discursive figures are categories of language which underscore the imaginative and poetic use of words. Their acceptability and accuracy in literature are proofs of their mastery by writers. This requires that every translator should have a wide knowledge and the ability to understand the intention of a text better. First, it has to do with set expressions; this is besides the stylistic aspects of which the object is to add aesthetic value to a work. Presented with a literary text in which discursive figures abound, the translator is not only called upon to give the message but to also express it in a similar figurative style using the exact equivalent in the target language. This implies that one is never permitted to translate figurative elements in any other form than their equivalences. We reconsider some discursive figures used in our corpus based on the theory of meaning and the translation principle that figurative elements have stylistic value and a universal acceptance; their meaning being a priori predetermined. Yet we note some imbalance in the translation of figurative expressions in some writers’ work, particularly those who translate their work as self-translators. Consequently, the translated text hardly renders the intended meaning of the original work which, curiously, is the “brainchild” of the same author. Les coïncidences et les polarités on which this study is based has many figures of speech wrongly translated into the target text. This makes us conclude that it is not enough to render the message; the beauty that comes with these expressions is also of equal importance.