Aspect-oriented programming (AOP) can effectively solve code-tangling and code-scattering that are caused by the attributes of crosscutting concerns or aspects. To date, research in the field of aspect orientation has concentrated on the implementation stage of the software development life cycle. Also, quite a good number of studies on aspect orientation have been conducted on the later stage of software development. However, less attention has been paid to aspect orientation at the requirement analysis stage, i.e., the early stage of software development. Therefore, the aim of this research is to propose an approach for analyzing and extracting crosscutting aspects from software requirements specifications by using a natural language processing technique, namely, the educated text stemmer algorithm. Moreover, this research focuses on the auto analysis and extraction of aspects from specifications written in the Arabic language because a lot of companies and organizations are still using English language-based requirements approaches for Arabic systems. The results of the proposed approach show promise because the proposed Arabic miner tool was able to extract the greatest number of candidate aspects from Arabic requirements specifications. The approach was evaluated by comparing the results that it automatically extracted with those extracted manually by software engineering experts. The accuracy of the experts' efforts was 88%, whereas the accuracy of the proposed approach was 36%, which can be ascribed to the poorness of the Arabic data set and Arabic WordNet tool. Nevertheless, the proposed approach has potential, and the experimental results offer insights on how to further develop the proposed tool.