In his article, “The Doctorinal Canonization of the Kathāvatthu” (IBK, Vol.63, No.3, 2015, pp.1243–1249), Toshifumi Shimizu investigated a discussion about the Kathāvatthu in the Aṭṭhasālinī, and concluded that, in order to accept what was spoken by the Buddha’s disciples (sāvakabhāsita) as “the word of the Buddha,” Theravādins required three conditions: (1) being based on māṭikā demonstrated by the Buddha, (2) corresponding to sabbaññutañāṇa, and (3) being retroactively approved with anumodanā given by the Buddha. However, his argument is not reasonable. The Aṭṭhasālinī explicitly mentions the reason why the Kathāvatthu composed by Moggaliputtatissa is buddhabhāsita as that, after the prediction of the Buddha, Moggaliputtatissa, following the summary (māṭikā) established by the Buddha, expounded it with the method given by the Buddha. The commentator compares it with the Madhupiṇḍikasutta of the Majjhimanikāya. In this sutta, the Buddha gives his disciples a brief discourse, and later Mahākaccāna expounds it to them in detail. That is finally confirmed by the Buddha saying, “I would have explained it in the same way that Mahākaccāna did. Such is exactly the meaning of this. Receive it as it is.”
It is noteworthy that the Pāli commentator did not intend to establish a general principle to regard sāvakabhāsita as buddhabhāsita, and also the Pāli commentator, in spite of his exertion to regard the Kathāvatthu as buddhabhāsita, agrees that what the Pāli tipiṭaka called “Buddhavacana” (the word of the Buddha) consists not only of buddhabhāsita, but also of sāvakabhāsita, and so on. Thus acknowledging a text to be buddhabhāsita is not equivalent to its canonization. Relating to the canonization of the Kathāvatthu, we may rather note that Moggaliputtatissa was pre-approved with the Buddha’s prediction and was entrusted as a successor by the arahats at the second council to purge the saṅgha by holding the third council with this work.