The current study aims to verify the claims, in the literature, about the repetitive nature of the Arabic language, its oral origin, as well as its potential transfer into Arabs' EFL discourse. To this end, using a mixed-methods research design, the study investigated the use of lexical repetition (LR) in 40 EFL essays, 40 EFL speeches, 30 Arabic as first language (AL1) essays and 30 AL1 speeches produced by Tunisian university students majoring in English and in Arabic. Through a quantitative analysis of the data, LR was measured by calculating the frequencies, as well as densities of simple and complex lexical repetition (SLR and CLR) and of overall LR. Furthermore, the frequencies and average length of lexical repetition chains (LRCs) were quantified. The statistical tests revealed an absence of a significant difference in the use of LR between AL1 written and oral productions, which implies a possible interplay between speech and writing in Arabic. Similarly, there was no statistically significant difference between AL1 and EFL written essays in LR measures. In contrast, the findings demonstrated a significant difference between the EFL written and oral corpora, with writing unexpectedly displaying higher rates of LR. Overall, the corpora exhibited a prevalence of SLR over CLR. The findings also uncovered longer LRCs and a higher density of CLR in writing in comparison to speech. The qualitative analysis unveiled two major patterns of distribution of LR, namely even dispersions and (densely) clustered distributions. It also accounted for the comparable CLR densities in EFL writing and AL1 productions, despite the highly derivational nature of Arabic. Finally, the study offered some examples of Arabic-specific LR.