The self-similar Qur'an project: Automatic detection of internal similarity in classical Arabic texts

A historically important tradition in exegesis, rooted in a number of scriptural passages, considers the Qur'an to be a self-similar text. This claim, while being sharply debated in literature, has never been independently tested. This paper proposes a strategy to measure self-similarity in classical Arabic texts, based on Leven-shtein distance, within the Self-Similar Qur'an (SSQ) project. The significance of self-similarity is checked against two tests: the similarity found in another religious text written in classical Arabic (the Arabic translation of the Diatessaron) and blind similarity judgments of uniformed readers on the translations of pairs of verses.