A third group held: when a person is unsure, he is not required to go back to what is certain, nor to make any investigation; he need only perform the prostration of forgetfulness whenever doubt arises.

The root of their disagreement lies in the seeming conflict among the textual proofs transmitted on this question. In this chapter there are three key reports:

1. The report that commands “building on certainty.” It is the ḥadīth of Abū Saʿīd al-Khudrī, who related that the Messenger of Allah—may Allah bless him and grant him peace—said:
   “If any one of you is uncertain in his prayer and does not know whether he has prayed three or four (rakʿahs), let him dismiss the doubt and build upon what he is certain of. Then let him perform two prostrations before giving salām. If he has in fact prayed five, the two prostrations will make his prayer even; and if he has completed four, they will be a humiliation for Satan.”
   Muslim narrated it.

2. The ḥadīth of Ibn Masʿūd, that the Prophet—may Allah bless him and grant him peace—said:
   “If one of you errs in his prayer, let him investigate, and then let him perform two prostrations.”
   In another wording from him:
   “Let him consider which is more likely to be correct, then give salām, then perform the two prostrations of forgetfulness, recite the tashahhud, and give salām again.”

3. The ḥadīth of Abū Hurayrah, recorded by Mālik and al-Bukhārī, that the Messenger of Allah—may Allah bless him and grant him peace—said:
   “When one of you stands to pray, Satan comes to him and confounds him until he no longer knows how much he has prayed. If that happens to any of you, let him perform two prostrations while he is seated.”
   Conveying the same meaning is the ḥadīth of ʿAbd Allāh ibn Jaʿfar, reported by Abū Dāwūd, in which the Messenger of Allah—may Allah bless him and grant him peace—said:
   “Whoever is uncertain about his prayer, let him perform two prostrations after it and then give salām.”

Faced with these aḥādīth, scholars followed two broad approaches: harmonization (jamʿ) and preference (tarjīḥ). Among those who preferred (tarjīḥ) are some who ignored the opposing texts, others who sought to interpret the opposing texts so as to align them with the preferred one, and still others who combined some reports and set aside the ruling of others.

As for those who both reconciled some reports and preferred others—while interpreting the non-preferred so that it supports the preferred—such as Mālik ibn Anas: he took the ḥadīth of Abū Saʿīd al-Khudrī to apply to the person whom doubt does not habitually assail, and the ḥadīth of Abū Hurayrah to the one whom doubt frequently overcomes. This is an instance of harmonization. He further interpreted the ḥadīth of Ibn Masʿūd by taking the “investigation” (taḥarrī) mentioned there to mean “returning to certainty,” thereby validating all the aḥādīth within his own legal framework.

As for the scholar who reconciled some reports but set aside others—thus exercising preference without interpreting the text that is set aside—that is Abū Ḥanīfah, for he held that the ḥadīth of Abū Saʿīd