He first studied in the Mahâvihâra and eventually requested permission to translate the Sinhalese commentaries. To prove his competence for the task he composed the celebrated Visuddhi-magga, and, this being considered satisfactory, he took up his residence in the Ganthâkara Vihâra and proceeded to the work of translation. When it was finished he returned to India or according to the Talaing tradition to Thaton. The Buddhaghosuppatti adds two stories of which the truth and meaning are equally doubtful. They are that Buddhaghosa burnt the works written by Mahinda and that his knowledge of Sanskrit was called in question but triumphantly proved. Can there be here any allusion to a Sanskrit canon supported by the opponents of the Mahâvihâra?
Even in its main outline the story is not very coherent for one would imagine that, if a Buddhist from Magadha went to Ceylon to translate the Sinhalese commentaries, his object must have been to introduce them among Indian Buddhists. But there is no evidence that Buddhaghosa did this and he is for us simply a great figure in the literary and religious history of Ceylon. Burmese tradition maintains that he was a native of Thaton and returned thither, when his labours in Ceylon were completed, to spread the scriptures in his native language. This version of his activity is intelligible, though the evidence for it is weak.
He composed a great corpus of exegetical literature which has been preserved, but, since much of it is still unedited, the precise extent of his labours is uncertain. There is however little doubt of the authenticity of his commentaries on the four great Nikâyas, on the Abhidhamma and on the Vinaya (called Samanta-pâsâdikâ) and in them[78] he refers to the Visuddhi-magga as his own work. He says expressly that his explanations are founded on Sinhalese materials, which he frequently cites as the opinion of the ancients (porânâ). By this word he probably means traditions recorded in Sinhalese and attributed to Mahinda, but it is in any case clear that the works which he consulted were considered old in the fifth century A.D. Some of their names are preserved in the Samanta-pâsâdikâ where he mentions the great commentary (Mahâ-Aṭṭhakathâ), the Raft commentary (Paccari, so called because written on a raft), the Kurundi commentary composed at Kurunda-Velu and others[79]. All this literature has disappeared and we can only judge of it by Buddhaghosa's reproduction which is probably not a translation but a selection and rearrangement. Indeed his occasional direct quotations from the ancients or from an Aṭṭhakathâ imply that the rest of the work is merely based on the Sinhalese commentaries.