He said:

"Fast when you sight it and break the fast when you sight it; if it is hidden from you by cloud, then complete thirty days. If two witnesses testify, then fast and break the fast (according to their testimony)."

Among the supporting reports is the ḥadīth of Ibn ʿAbbās, who said:

"A Bedouin came to the Prophet—may Allah bless him and grant him peace—and said, ‘I saw the crescent tonight.’ The Prophet asked, ‘Do you bear witness that there is no god but Allah and that Muḥammad is His servant and Messenger?’ He replied, ‘Yes.’ The Prophet then said, ‘Bilāl, announce to the people that they are to fast tomorrow.’”

Al-Tirmidhī recorded it and remarked, “Its chain is disputed, for a number of scholars transmitted it as mursal.”

Another proof-text is the ḥadīth of Ribʿī ibn Ḥirāsh, which Abū Dāwūd transmitted from Ribʿī ibn Ḥirāsh, from a man among the Companions of the Messenger of Allah—peace and blessings be upon him—who said:

"On the last day of Ramadan, two Bedouins came forward and testified before the Prophet—peace and blessings be upon him—that they had seen the crescent the previous evening. The Messenger of Allah—peace and blessings be upon him—thereupon ordered the people to break the fast and to return to the prayer ground.”

Faced with these narrations, the scholars followed one of two approaches: preferring one report over another (tarjīḥ) or reconciling them (jamʿ).

Al-Shāfiʿī reconciled the report of Ibn ʿAbbās with that of Ribʿī ibn Ḥirāsh in their outward meanings, obligating the fast on the testimony of a single witness and ending it on the testimony of two.

Mālik gave precedence to the ḥadīth of ʿAbd al-Raḥmān ibn Zayd by way of analogy—that is, by likening the matter to legal testimony in property cases.

It appears that Abū Thawr likewise saw no contradiction between the reports of Ibn ʿAbbās and Ribʿī ibn Ḥirāsh. The latter records a judgment based on two witnesses, while the former records a judgment based on one, which indicates that both options are permissible. It does not amount to a contradiction, nor does it imply that the first ruling applies solely to the beginning of the fast and the second to its end; such a notion rests only on an imagined conflict. Similarly, there seems to be no clash between the report of ʿAbd al-Raḥmān ibn Zayd and that of Ibn ʿAbbās except by an argument from implication (dalīl al-khiṭāb), which is weak when confronted by an explicit text. Hence, in spite of its rarity, Abū Thawr’s view appears clearer, for comparing the one who sights the moon to a narrator is sounder than comparing him to a legal witness.

With respect to testimony, one must either say that stipulating a specific number of witnesses is an unreasoned act of worship, in which case analogy is impermissible, or say that the numeric requirement is due to the contentious nature of property disputes and the suspicion that arises from one litigant’s claim. The number was therefore set so that the presumption of truth would be stronger and the tendency toward one party’s evidence would be firmer, yet it was not raised above two lest establishing testimony become too difficult and rights be lost. In the sighting of the moon there is no opposing claimant whose doubt necessitates bolstering the evidence with multiple witnesses.

It likewise seems that al-Shāfiʿī distinguished between the crescent of breaking the fast and the crescent of