The ḥadīth of al-Faḍl: the strongest report on the other side
A woman spoke with the Prophet while al-Faḍl b. ʿAbbās looked at her, and she was not told to cover. It is in al-Bukhārī. Those who hold covering obligatory answer it four ways, rated unevenly here.
Most pages here set out the case for the practice. This one sets out the report that tells against it, because a reference that only presented one side would not be worth consulting.
The report
In Ṣaḥīḥ al-Bukhārī: a woman came to ask the Prophet a question. Al-Faḍl b. ʿAbbās looked at her and she at him, and the Prophet reached back, took hold of al-Faḍl's chin, and turned his face away.
A woman of Khathʿam came seeking a ruling, and al-Faḍl b. ʿAbbās began looking at her and she at him. The Prophet reached back, took hold of al-Faḍl’s chin, and turned his face to the other side.
The inference drawn from it is direct. Her face was visible: otherwise there would have been nothing to look at, and nothing to turn away from. And she was not instructed to cover it. The correction fell on the man's gaze, not on the woman's face.
This is the strongest evidence on that side, and Ibnu Ḥazm builds his position on it. Its authenticity is not in dispute.
Four answers, and what they are worth
Those who hold covering obligatory answer it four ways, and the four do not hold up equally well.
One — she uncovered deliberately, hoping he would marry her. Supported by a narration in Abū Yaʿlā's Musnad in which a bedouin presents his daughter to the Prophet. The grading: the base report is jayyid, but Yūnus b. Abī Isḥāq is ṣadūq yahimu qalīlan, and the addition "hoping he would marry her" is rejected as munkar, appearing only in a chain not strong enough to carry it. The answer fails on the very detail it needs.
Two: she was an enslaved woman, to whom the covering rule did not apply. Al-Ṭarīfī's argument, resting on transmissions describing her as jāriyah shābbah. One such report, through Mālik and al-Zuhrī, is graded sound on the conditions of al-Bukhārī and Muslim. The linguistic argument is that jāriyah alone can mean either a girl or a slave, but paired with shābbah (young) it must mean slave, since otherwise shābbah adds nothing. The verdict: worth considering.
Three — she was inside a hawdaj, a covered litter. Described as untidy, because it would imply a woman may show her face to any man so long as she is in one — though it does weaken the inference that the ḥadīth licenses uncovering in public generally.
Four — she was a child. Judged not strong, since its basis is a narration in Aḥmad's Musnad reaching at best ḥasan.
What this page is doing
Two of the four answers fail on this site's own assessment. One is worth considering. One is partial.
That is the honest state of the argument, and it is set out here arguing for the practice — not by an opponent. A page that reported four rebuttals without saying which of them are rated poorly here would be misrepresenting the case in its own favour.
Check your understanding
Those who hold niqab obligatory offer four answers to the ḥadīth of al-Faḍl. How does the note rate these four answers?