Many Western scholars such as Dryden show little interest in imitations, and express their preference for translations, i.e. paraphrases that are faithful to the sense of the source text. However, they consider imitations as a viable category of translation. It is the degree of freedom, or departure from the original, that differentiates a translation from an imitation. This paper is concerned with issues that are central to the understanding of English-Arabic translation errors when rendering expository text. Not surprisingly, when translating exposition, errors recur, especially those relating to the linguistic competence of the students. But not all errors were the same neither was their distribution. Each text-type shows different idiosyncrasies and error distributions which indicate that performance in translation depends largely on the type of text and the rhetorical purposes as well as patterns which follow from the source text. To this end, an error corpus of linguistic structure was collected from the translation project of students majoring in translation. Syntactic, semantic, pragmatic and discoursal criteria were used to judge imitation and authenticity strategies adopted by the students during the translation process. Implications for increasing students' awareness of the pragmatic and syntactic constraints in translating structures will also be provided.