Sūrat al-Nisāʾ, Verses 17 to 18

1. The Questions This Verse Raises

The two verses state whose repentance Allah accepts and whose He does not. The questions that follow examine the words on which this depends, in the order of the verses; the table summarizes the outcome of each discussion.

Q1. What بِجَهَالَةٍ means, and whether the promise extends to one who sins knowingly

The first verse states that the acceptance of repentance is 'upon Allah' for those who do evil بِجَهَالَةٍ (bi-jahālatin), "in jahāla", and it opens with the restrictive particle إِنَّمَا (innamā), "only". If jahāla here meant ignorance that the act is a sin, one who knew that his act was a sin would appear to be excluded from the promise, a difficulty the commentators raise themselves (al-Rāzī, Ibn ʿĀdil).⁴⁰ Two questions therefore need answering:

1. The word: what jahāla means in بِجَهَالَةٍ.
2. The scope of the promise: whether the promise extends to one who sinned knowing that the act was a sin.

Since the answer to the second follows from the first, the positions below address both together.

The earliest explanations understood jahāla in ways that do not exclude one who sins knowingly. One identified jahāla with the deliberate act itself: ignorance of the lawful and the forbidden is not the sinner's jahāla; his entering into the sin is (Mujāhid, al-Ḍaḥḥāk).⁴¹ A second took 'in jahāla' to mean in this world, the whole of which is jahāla, so that the phrase describes when the evil is done rather than what the sinner knew (ʿIkrima).⁴² A third held that the sinner knows the act to be a sin but does not know the full extent of its punishment (al-Farrāʾ, al-Kalbī).⁴³ This third explanation cannot be the complete meaning of the word, for it would deny repentance to anyone who fully knew the punishment, contrary to the Prophet's ﷺ established teaching that the door of repentance is open until the sun rises from the west (al-Ṭabarī).⁴⁴ It describes one of the states the word includes, as the first position below holds, but not the only one. None of the three, then, restricts the promise to one who did not know that his act was a sin; whether jahāla here has that restricted meaning is what the two positions below address.

ʿAbduh includes ignorance of the prohibition as well as the folly of a person who knows it. This gives him a point of contact with both explanations; it does not make him a holder of the whole second position.⁴⁷

The understanding of the Companions. The earliest evidence for the meaning of the word is the understanding of the Companions, who heard the verse revealed in their own Arabic, and for them the word included every sin, deliberate or not. Abū al-ʿĀliya reports that the Companions رضي الله عنهم of the Prophet ﷺ used to say that every sin a servant commits is committed in jahāla,⁴⁸ and Qatāda reports that they gathered and concluded that anything by which Allah is disobeyed is jahāla, whether deliberate or not.⁴⁹ Qatāda's report has a sound isnād to him but is mursal, since he did not hear it from the Companions whose agreement he reports; Abū al-ʿĀliya's report has a ḥasan isnād (Ḥikmat Bashīr Yāsīn).⁵⁰ Several commentators transmit Qatāda's words as a consensus of the Companions; Successors who hold the view are named (Ibn Kathīr, al-Qurṭubī), and it is given as the view of most of the Companions, the Successors and the exegetes (Makkī).⁵¹ According to those who first heard the verse, therefore, a deliberate sin is included in the word.

Why a knowing sinner is said to act in jahāla. The reason for the first position is a rule of Arabic usage. Arabic does not describe one who does something deliberately as ignorant of it, unless it means that he does not know the true measure of its benefit and harm; when it says that one who sinned knowingly acted in jahāla, it means that he behaved as the ignorant behave, not that he lacked knowledge. Those who do evil are therefore in jahāla in the very act, whether they sin deliberately or without knowing what Allah has prepared for those who commit it (al-Ṭabarī).⁵² The word describes the nature of the act, not a deficiency in the knowledge of the one who commits it. Later commentators identify that quality as folly, safāha: evil is committed only when appetite or anger overcomes reason and displaces forbearance, so that the act is prompted by folly and never by wisdom (al-Zamakhsharī, Ibn ʿAṭiyya, Abū Ḥayyān, al-Marāghī).⁵³

The lexicons confirm this meaning. Besides its root meaning, lack of knowledge, jahl denotes doing something other than as it ought to be done, whether one's belief about it is right or wrong, as when a person deliberately abandons the prayer (al-Rāghib).⁵⁴ Arabic poetry uses the verb in the same sense, for rash and wrongful conduct against another:

The line is by ʿAmr b. Kulthūm, and it is cited as evidence that jahāla means rash action and wrongdoing, the opposite of forbearance, which is its meaning in this verse (Ibn ʿĀshūr).⁵⁵ The Qurʾān uses the word in the same way: when Yūsuf says "and I be of the ignorant" (12:33), the ignorant are those who disobey (al-Wāḥidī).⁵⁶ Usage outside the commentaries therefore confirms that jahāla describes a way of acting, not a lack of knowledge.

Why ignorance of the ruling cannot be meant. Legal reasoning leads to the same conclusion, for the repentance the verse speaks of is repentance from the evil deed itself. One who does wrong without knowing that it is wrong is excused for the deed and incurs no sin by it (Ibn ʿUthaymīn, Ibn Kamāl Pāshā).⁵⁷ If he was able to learn, what he failed in was the duty to learn, since every Muslim is required to know what is forbidden; that obligation is why jahāla here cannot mean ignorance of the prohibition (al-Zuḥaylī).⁵⁸ The deed done in ignorance is therefore not what he repents of, since one who does not know needs no repentance (al-Ālūsī, al-Qāsimī).⁵⁹ A promise of acceptance restricted to such people would be a promise to those who incurred no sin by the deed the verse describes. The rules of liability show the same from the other side: the jahāla of the verse is not an ignorance that excuses. The verse cannot mean that those who do evil lack discernment, since one who lacks reason or discernment is not liable to a ḥadd; the ignorance present in a sin consists rather in preferring a passing pleasure to a lasting one, and that ignorance does not remove liability to punishment, otherwise no one would ever be punished (al-Zajjāj).⁶⁰

The objection from the literal sense. The second position, held by al-Rāzī, takes jahāla in its literal sense: committing a sin without knowing that it is a sin, while being able to know. This position prefers the definition because it keeps the word in its literal sense; the deliberate sinner then falls outside the verse, and his case is known only by analogy.⁶¹ On this definition the restrictive particle إِنَّمَا has a clear purpose: one who sins without knowing is in a lighter state than one who sins knowingly, so the promised acceptance is reserved for the lighter case.⁶² The two verses then describe two classes of sinners: those whose repentance Allah has promised to accept, and those whose repentance He refuses. Between them lies a third class, those who do evil deliberately and then repent, to whom Allah has neither promised acceptance nor refused it, leaving their case to His will.⁶³

The objection is limited in scope: it concerns the guarantee of acceptance, not acceptance itself. It does not claim that Allah refuses the repentance of one who sinned knowingly; indeed, al-Rāzī himself states that the apparent sense of إِنَّمَا would imply such a refusal, and that this is false by consensus.⁶⁴ It claims only that the acceptance of such a person's repentance is left to Allah's will rather than guaranteed by the promise of this verse.

The response to the literal sense. The response accepts that lack of knowledge is one meaning of jahl, but denies that it is the meaning intended here, for four reasons.

1. Consensus on the meaning of the word. The scholars agree that jahāla here does not mean ignorance that the acts are sins, because one who commits a sin without knowing that it is a sin deserves no punishment, since the burden of error has been lifted from this community (al-Wāḥidī). The literal definition has accordingly been judged weak, as contrary to what the exegetes agreed upon (al-Rummānī, as Darwīsh reports him).⁶⁵
2. Folly is an established meaning of the word. Since the lexicons define jahl as doing something other than as it ought to be done, knowingly or not, folly is not a figurative meaning imposed on the word. Even if it were a departure from the literal sense, the evidence already presented would justify it: the Arabic use of jahl for rash conduct, the Companions' understanding of this verse, and the Qurʾān's use of the word for Yūsuf's disobedience. Furthermore, the literal meaning the objection relies on belongs to jahl, not to jahāla, the form the verse uses: a lack of knowledge about what one has done is called jahl, and it is not called jahāla (Ibn ʿĀshūr).⁶⁶
3. There is no third class between the two verses. The restriction of إِنَّمَا would exclude the knowing sinner only if 'in jahāla' were a condition of acceptance, and it is not. Both qualifiers in the first verse, 'in jahāla' and 'soon', are admonitions on how a Muslim should conduct himself, not conditions of acceptance, and the second verse is the counterpart of the first, with no intermediate class between them (Ibn ʿĀshūr).⁶⁷ One who sinned knowingly and then repents is therefore not left in a middle class between the promise and the refusal.
4. The concern behind the narrower definition is addressed elsewhere in the verses. A modern definition restricts jahāla to a sin committed without deliberate intent, persistence or indifference. Its concern is one who sins knowingly and persistently and does not repent after the sin, and who may therefore die without repenting (al-Jazāʾirī).⁶⁸ The two verses address that concern through their other terms: 'soon', which Question 2 examines, and the second verse's refusal of one who delays repentance until death is present, which Question 4 examines. That concern therefore gives no reason to restrict the meaning of jahāla.

Which interpretation is stronger? The first position is stronger, because it agrees with Arabic usage, with the Companions' understanding of the verse and with the consensus on the meaning of the word. The second position, by contrast, leaves outside the verse the deliberate sinner, whose repentance the scholars agree is not refused, and reaches his case only by analogy; Ibn ʿĀdil, who draws the second position's consequences, reports that agreement himself (Ibn ʿAṭiyya, Abū Ḥayyān, Ibn ʿĀdil).⁶⁹ This interpretation is adopted here. It is al-Ṭabarī's preference, based on Arabic usage and shared by al-Shāfiʿī and al-Wāḥidī, and al-Ālūsī and Ibn ʿĀshūr defend it against the narrower definition, as shown above.⁷⁰

Conclusion: Every sin is committed in jahāla, the folly of one who disobeys Allah, whether he sins deliberately or without knowing what Allah has prepared for those who commit it; the promise therefore extends to one who sinned knowingly and then repents.

And Allah knows best.

Q2. How soon مِنْ قَرِيبٍ is, and the report of the death rattle

The first verse promises acceptance to those who repent within a short time: ثُمَّ يَتُوبُونَ مِنْ قَرِيبٍ (thumma yatūbūna min qarībin), "then they repent soon", literally "from near". The verse does not state the point from which this nearness is measured, whether from the sin or from death. Three questions therefore need answering:

1. The last moment: until when repentance is accepted.
2. The measure: whether 'soon' is measured from the sin or from death.
3. The late penitent: whether one who repents long after the sin, but before death, is included in this verse's promise.

The two verses and the Sunna give one answer to the first question, so the positions below address the second and the third.

Explaining ‘soon’ as near the sin does not by itself mean that a later sound repentance falls outside the promise. The following discussion distinguishes the scholars’ conclusions from the lexical explanation they share.

When the time for repentance ends. The second verse answers the first question. It denies repentance to those who continue in evil deeds until the moment described as حَتَّى إِذَا حَضَرَ أَحَدَهُمُ الْمَوْتُ, "until, when death is present to one of them", when the sinner says, "I repent now". The presence of death is therefore the moment after which repentance is no longer accepted. From this the first position infers that all the time before that moment counts as near (al-Zamakhsharī, al-Bayḍāwī, Abū al-Suʿūd).⁷³ It takes the preposition مِنْ in its partitive sense, meaning "in some part of" that time, so that whoever repents at any point between the sin and the presence of death repents 'from near' (al-Zamakhsharī, al-Maẓharī, al-Zuḥaylī).⁷⁴

The period ends at the presence of death because of what repentance is. Repentance is regret for past sins and a resolve not to return to them, made by a person who understands what he regrets and chooses to abandon the sin (al-Ṭabarī).⁷⁵ Regret and resolve are possible as long as hope remains; once death takes hold of the sinner, neither is possible, and repentance therefore becomes impossible (Ibn ʿAṭiyya, with al-Qurṭubī reporting his scholars).⁷⁶ Consequently, a dying person who can still reason as a healthy person does, and who turns to his Lord in repentance, is included in the promise, if Allah wills (al-Ṭabarī).⁷⁷

Why a whole lifetime counts as near. A whole lifetime counts as near, for three reasons: the appointed term is approaching, and whatever approaches is near; a life, however long, is short when set between the eternity before it and the eternity after it; and death may come at any moment. Measured against what comes after it, the whole of life is near (al-Rāzī, although he measures 'soon' from the sin; al-Maẓharī, Ibn Kamāl Pāshā).⁷⁸ Death, by contrast, is what is truly far, and Arabic poetry uses the word in this sense:

As his mourners bury the poet, they tell him "do not go far", and he answers that his own place, the grave, is the only place that is truly far. The line is by Mālik b. al-Rayb, as Ibn ʿAṭiyya identifies him when he quotes its second half, and it is cited as evidence that death is what is far (Ibn Rajab).⁷⁹ For the first position, then, a repentance made at any point in life is made 'soon'.

The hadith of the death rattle. The Sunna confirms, in the Prophet's ﷺ own words, that the time for repentance ends with the throes of death. Ibn ʿUmar رضي الله عنهما relates that the Prophet ﷺ said:

Al-Tirmidhī grades the hadith ḥasan gharīb. Its isnād is graded ḥasan rather than ṣaḥīḥ only because of one narrator, Ibn Thawbān, while the rest of its narrators are trustworthy (Shuʿayb al-Arnaʾūṭ and his colleagues, in their notes to the Musnad), and later critics grade the hadith ḥasan or ṣaḥīḥ. It is therefore sound evidence for the last moment of repentance.⁸¹ The verb yughargir refers to the soul reaching the throat, where it moves as a drink moves when it is gargled at the back of the throat without being swallowed (al-Harawī, as al-Qurṭubī reports him; the image of the drink is in the notes of Shuʿayb al-Arnaʾūṭ and his colleagues).⁸²

Scholars who measure 'soon' differently still agree on this last moment. Al-Rāzī, who measures 'soon' from the sin, reports a consensus on the end of the period, not on its measure: the nearness in the verse ends with the arrival of death and the sight of its terrors.⁸³ The scholars also agree that repentance is accepted as long as the dying person has not reached the death rattle, as the hadith states (al-Nawawī).⁸⁴

The early explanations, and where they agree. None of the early explanations places the last moment of repentance after the death rattle. They place the end of the period at different moments: while the person is healthy, before illness; before he sees the angel of death; at any point in this life; or before the death rattle.⁸⁵ One of these is not a last moment at all: "before illness" refers to the best time for repentance, whereas the explanation most scholars give, until the death rattle or the sight of the angels, refers to the last time (Ibn ʿAṭiyya, Abū Ḥayyān).⁸⁶ Illness itself does not end the time for repentance. Ibn Taymiyya states that repentance remains accepted before the death rattle, even during an illness feared to be fatal, as the Mawsūʿa records. Al-Shawkānī likewise rejects ‘before illness’ as the final limit because the hadith sets it at the death rattle.⁸⁷

The other explanations end at the same point, the moment of dying. Once a person despairs of life, sees the angel, and his soul rattles in his throat, no repentance is accepted from him (Ibn Kathīr).⁸⁸ One point remains, the sight of the angel: some end the time at that sight (Ibn ʿAbbās رضي الله عنهما, al-Ḍaḥḥāk, Abū Mijlaz),⁸⁹ while one modern explanation accepts a repentance made after it, so long as the rattle has not begun (al-Zuḥaylī).⁹⁰ The sight is taken here as the limit, with Ibn Kathīr, who joins it to the rattle: the angel is seen when death is present, the moment the hadith of the death rattle specifies, and every early explanation ends at or before it.⁹¹

The objection from the literal sense of 'near'. The second position takes 'near' in its literal sense. In Arabic the primary sense of the word is nearness to the time of the sin; the verse therefore requires repentance immediately after the sin, without delay, because one who postpones repentance becomes persistent in the sin, and on this view the interpretation of many exegetes departs from the wording of the verse (ʿAbduh as Rashīd Riḍā reports him, al-Qāsimī, and al-Rāzī for the reason of persistence).⁹² On this interpretation مِنْ marks the beginning of the period: the repentance must start from a time near the sin (al-Samīn).⁹³

The objection takes two forms, according to what each says of one who repents long after the sin but before death. In the first form, this verse does not promise him acceptance, and his case is one of hope. The first verse states when repentance is always accepted and the second when it is never accepted; between them lies a middle state of hope and fear, and the late penitent belongs to it, among those of whom Allah says, "it may be that Allah will turn to them" (9:102), a hope rather than a promise (al-Rāzī, Rashīd Riḍā; al-Samīn, who sets it out on the Kūfan view of the particle).⁹⁴ In the second form, his repentance is accepted, but that acceptance is known from the second verse and the hadiths, not from this verse (al-Qāsimī).⁹⁵ Both forms deny that this verse's promise includes him; only the first also denies that his acceptance is assured.

The response to the literal sense. The response accepts that 'near' can mean near to the sin, but denies that one who repents later is excluded from this verse's promise, for three reasons. The first reason answers both forms of the objection, the second answers its first form, and the third answers its analysis of مِنْ.

1. 'Soon' is an admonition, not a condition. Even if 'soon' means a time near the evil deed, it is an admonition on how a Muslim should conduct himself, like the verse's first qualifier, 'in jahāla' (bi-jahālatin), which Question 1 examines. It is not a condition of acceptance, because the second verse is the counterpart of the first, and no third group lies between the two verses (Ibn ʿĀshūr, who himself takes 'soon' to mean a time near the deed).⁹⁶ Since 'soon' is not a condition, this verse's promise includes one who repents later.
2. The hadith states acceptance, not hope. The hadith of the death rattle states that Allah accepts the servant's repentance until the rattle. It therefore leaves no period before the rattle in which a sound repentance is only hoped for, and the middle state of hope and fear on which the first form rests has no place between the two verses.
3. The grammar does not require the other sense. Using مِنْ to mark the beginning of a period of time is permitted by the Kūfan grammarians and not by the Baṣran grammarians, whereas the partitive sense of مِنْ needs no such exception (al-Samīn).⁹⁷ The objection's analysis of مِنْ therefore rests on a rule that one school of grammar rejects, while the partitive sense does not.

The call for immediate repentance is kept, but as an exhortation rather than a condition, because the hadith of the death rattle was given to prevent despair, not to permit delay. It means that no one should despair of Allah's mercy and of the acceptance of repentance while he is alive; it does not mean that one who persists in sin has nothing to fear if he repents an hour before death (al-Marāghī).⁹⁸ Repenting quickly is also obligatory, because Allah has tied acceptance to a term that no one knows (Ibn ʿUthaymīn).⁹⁹ This obligation is not a condition of acceptance: one who delays falls short of it, and his repentance before the death rattle is still accepted.

Which interpretation is stronger? The first position is stronger, both on the measure of 'soon' and on the late penitent, because the words of the two verses, the hadith of the death rattle and the consensus all extend acceptance until the rattle, and none of them leaves a period in which a sound repentance is only hoped for. The concern of the second position is not discarded: repentance immediately after the sin is obligatory, but delay does not itself invalidate a sound repentance before the limit. This interpretation is adopted here. It follows al-Ṭabarī's tarjīḥ, shared by Ibn Kathīr and al-Maẓharī, that repentance is accepted before death, while the sinner still understands Allah's command and prohibition and before he is overcome and the throes of the death rattle begin. Like Ibn ʿAṭiyya, Abū Ḥayyān and Ibn ʿĀshūr, it distinguishes the best time for repentance from the last.¹⁰⁰

Conclusion: ‘Soon’ includes any time before the throes of death, while the sinner can still regret the sin and resolve to leave it. The hadith marks the death rattle as the point at which acceptance ends. Repenting at once is obligatory, but delay does not itself invalidate a sound repentance made before that limit.

And Allah knows best.

Q3. What عَلَى اللَّهِ means, and how certain the acceptance of repentance is

The first verse opens with إِنَّمَا التَّوْبَةُ عَلَى اللَّهِ (innamā al-tawbatu ʿalā Allāhi), "repentance is only upon Allah". Here 'repentance' refers to Allah's acceptance of repentance, not to the servant's act of repenting (Ibn Kamāl Pāshā, Abū al-Suʿūd, Maḥmūd Ṣāfī).¹⁰¹ In Arabic the preposition عَلَى (ʿalā), "upon", is used for what is due from a person, as in the expression "I owe you such-and-such"; the verse therefore appears to state that the acceptance of repentance is due from Allah (al-Biqāʿī, al-Ālūsī, Ibn ʿĀshūr).¹⁰² Two questions therefore need answering:

1. The kind of necessity: whether 'upon Allah' means an obligation that reason imposes on Allah, or a promise that He has made binding on Himself.
2. The certainty: whether the acceptance of a sound repentance is certain, or only very likely.

The positions below address the first question; the second is examined after them.

The argument for an obligation. The first position takes عَلَى in its sense of obligation. On this view the first clause announces that acceptance is obligatory on Allah, as some acts of obedience are obligatory on the servant, and the second clause, 'those are the ones to whom Allah turns', is His promise that He will fulfil what is obligatory on Him (al-Zamakhsharī).¹⁰⁵ The structure of the verse is offered as a second argument. If the first clause meant only that Allah accepts repentance, it would say the same as the second, which also reports that acceptance takes place; only if the first states the obligation and the second its occurrence are the two clauses distinct, so that the second is not a restatement of the first (al-Qāḍī, as al-Rāzī reports him).¹⁰⁶ The obligation is held to be established by reason: a sincere repentance is accepted with certainty by rational proof. This is the view of the Muʿtazila, the school of theologians who held that reason can establish what is binding even on Allah, as Ibn ʿĀshūr reports them.¹⁰⁷ On this view, acceptance would be a debt that the penitent could claim from Allah.

Why reason imposes no obligation on Allah. The response denies that reason can impose any obligation on Allah, for three reasons.

1. An obligation requires one who imposes it. An obligation is imposed by one of higher rank than the one obliged, and Allah is the Creator and Owner of His creatures and the One who assigns them their duties; He cannot, therefore, be described as bound by any obligation (al-Qurṭubī, al-Zuḥaylī).¹⁰⁸ An obligation is also something whose omission deserves blame. Since blame cannot attach to Allah, an acceptance He was obliged to give would be one He could never withhold, and He would then act under necessity rather than by His own choice; to say instead that He could deserve blame is absurd (al-Rāzī, Ibn ʿĀdil).¹⁰⁹
2. The servant's act cannot bind Allah. Whatever Allah grants, He grants out of grace, not in return for a prior desert (Ibn al-Munayyir).¹¹⁰ On the first position's own premise, repentance is the servant's free act; if that act made acceptance obligatory on Allah, an act of the servant would bind the Creator, which no rational person accepts (al-Rāzī).¹¹¹
3. The penitent seeks pardon, which reason does not require. What the penitent seeks is the pardon of a punishment he had already earned before he repented. Reason does not require that past sins be pardoned because a person resolves to abandon them in the future; the acceptance of repentance is therefore known from revelation, not from reason (Ibn ʿĀshūr).¹¹²

The first position therefore fails where it rests on reason. The word 'upon' remains in the verse, however, and it still has to be explained.

What 'upon Allah' means. The second position explains the word by Allah's promise. Allah has promised to accept repentance, and when He promises something, He fulfils it and cannot break it; a promise that cannot be broken resembles an obligation, and that resemblance is why عَلَى is used (al-Rāzī, al-Nasafī).¹¹³ Every expression of obligation used here refers to the necessity that Allah's promise come true. Necessity in this sense means only that the thing cannot fail to be, as when Allah's existence is called necessary, meaning that He cannot fail to exist; it does not mean a duty owed to anyone, for no one can earn anything from Allah (Ibn al-Munayyir).¹¹⁴ The promise itself is stated in the Qurʾān: "Your Lord has written mercy upon Himself" (6:54). 'Upon Allah' therefore means that He has made the acceptance binding on Himself by His grace (ahl al-maʿānī, as al-Wāḥidī reports them).¹¹⁵ On this explanation عَلَى keeps its sense of commitment: the promise of acceptance is so firmly confirmed that it is made like a right upon Allah, while nothing is binding on Allah except His own promise, given out of grace (Ibn ʿĀshūr).¹¹⁶ Other explanations reach the same result through an implied word. The phrase may mean 'upon Allah's grace and His mercy to His servants' (Ibn ʿAṭiyya),¹¹⁷ and the grammarians analyse it as 'the acceptance of repentance rests on the bounty of Allah', not by way of obligation (Abū Ḥayyān; al-Samīn, for the implied word).¹¹⁸ In each explanation the necessity is that of a promise given freely, not of a debt.

The two clauses of the verse. The same explanation answers the argument from the structure of the verse. The first clause, 'repentance is only upon Allah', states what Allah has made binding on Himself, and the second, 'those are the ones to whom Allah turns', states that He fulfils it. The two clauses therefore stand to each other as a promise stands to its fulfilment, and no repetition arises (al-Rāzī).¹¹⁹ Since it is impossible for Allah to break His promise, the second clause stands to the first as a result stands to its cause, and it confirms what Allah has bound Himself to (al-Maẓharī, Ibn ʿUthaymīn).¹²⁰ The difference between the two clauses therefore requires no obligation imposed by reason.

An obligation that Allah imposes on Himself. Some explanations keep the word 'obligation' but assign it to Allah's own will. Allah may make binding on Himself whatever He wills, while His servants can make nothing binding on Him, and the acceptance of repentance is binding in this sense (Ibn ʿUthaymīn, citing 6:12).¹²¹ On this basis the difference between the Muʿtazila and Ahl al-Sunna, the Sunni theologians, on this question has been called merely verbal: the acceptance is upon Allah not by the command of any authority over Him, but as part of the perfection that He has made binding on Himself by His will and choice, and expressions that appear to make something obligatory on Allah follow the Arabs' way of speaking (Rashīd Riḍā).¹²² The difference is verbal only when 'obligation' means what Allah binds on Himself. It is not true of the claim that reason imposes the obligation, which the arguments against any obligation of reason refute. In substance, both explanations belong to the second position.

How certain the acceptance is. The second question concerns a believer's repentance from his sins. The acceptance of a disbeliever's faith is certain, since revelation states it (Ibn ʿAṭiyya, Abū Ḥayyān; Ibn al-Faras, as Ibn ʿĀshūr reports him).¹²³ For a believer who repents sincerely and fulfils every condition of repentance, one group holds that acceptance is certain, as Allah has stated of Himself, and this is reported as the view of the majority of scholars (Ibn ʿAṭiyya, following his father, and Ibn Juzayy, who reports the majority).¹²⁴ Another group holds that these texts give only a strong likelihood of acceptance, not certainty (Abū al-Maʿālī al-Juwaynī, as al-Qurṭubī, Ibn Juzayy and Abū Ḥayyān report him).¹²⁵ Both groups agree that nothing is binding on Allah by reason; the question between them is only whether the many texts that promise acceptance, taken together, give certainty (Ibn ʿĀshūr, as his framing of the two groups implies).¹²⁶

The question becomes clearer once the senses of 'acceptance' are distinguished. When acceptance means the forgiveness of past sins that Allah has promised, it is known from revelation and not from reason, and this is the sense to which the question of certainty applies. When it means that a commanded act, performed with its legal conditions, is valid and discharges the command, the acceptance is certain, although the reward for the act remains a matter of likelihood (Ibn ʿĀshūr).¹²⁷ Repentance is itself a commanded act, so in this second sense its acceptance is certain. The acceptance of repentance, moreover, is not a question of the foundations of the religion, so it does not require a proof that yields certainty (Ibn ʿĀshūr, who prefers in theory that the texts give only a strong likelihood).¹²⁸

Which interpretation is stronger? The second position is stronger. It explains the word 'upon' by a promise that the Qurʾān itself states; it explains the difference between the two clauses as the difference between a promise and its fulfilment; and it avoids what the first position cannot avoid, an obligation imposed on the Creator by reason or by the act of His servant. This interpretation is adopted here. It follows al-Rāzī's refutation of the obligation, Ibn al-Munayyir's answer to al-Zamakhsharī, and Ibn ʿĀshūr's observation that what the penitent seeks is a pardon that reason cannot require.¹²⁹ On the second question, the acceptance of a repentance that meets its conditions is certain, because the texts of the Qurʾān and the Sunna that promise it agree and together yield certainty. This is Ibn ʿAṭiyya's tarjīḥ, following his father, and Abū Ḥayyān's, against Ibn ʿĀshūr's preference in theory for a strong likelihood.¹³⁰

Conclusion: 'Upon Allah' means a promise that Allah has made binding on Himself out of His grace; it is not an obligation that reason imposes on Him.

And Allah knows best.

Q4. Whom the second verse excludes, whether it was abrogated, and why repentance ends when death is present

Verse 4:18 denies the acceptance of repentance to two groups. It names the first group لِلَّذِينَ يَعْمَلُونَ السَّيِّئَاتِ, "those who do evil deeds": they continue in these deeds until death is present, and then they say, "I repent now". It names the second group الَّذِينَ يَمُوتُونَ وَهُمْ كُفَّارٌ, "those who die while they are disbelievers". The verse ends with a threat: "for them We have prepared a painful punishment". Four questions therefore need answering:

1. The first group: who 'those who do evil deeds' are.
2. Abrogation: whether a later verse, 4:48, abrogated this one.
3. The last moment: what ends the acceptance of repentance when death is present, and why.
4. The second group and the threat: what 'those who die while they are disbelievers' refers to, whether a disbeliever's faith is accepted once death is present, and for whom the punishment is prepared.

The positions below address the first question; the other three are examined after them.

Why the first group are believers. The wording of 4:18 decides the question.

1. The hypocrites are already disbelievers. If the first group were the hypocrites, the separate mention of وَلَا الَّذِينَ يَمُوتُونَ وَهُمْ كُفَّارٌ, "nor those who die while they are disbelievers", would have no meaning of its own.
2. The verse names the two groups differently. Allah called the second group disbelievers. He described the first group as people of evil deeds and did not call them disbelievers. This shows that the two groups are different (al-Ṭabarī).¹³⁵

The 'evil deeds' of the first group are therefore sins short of disbelief (al-Saʿdī).¹³⁶ From this al-Rāzī draws a further consequence: a Muslim who sins is not a disbeliever. This refutes the Khawārij, the early sect that held that a Muslim who commits a grave sin becomes a disbeliever.¹³⁷

What the early reports say. The reports from the early authorities point in different directions, so they cannot settle the question by themselves.

1. The reports from Sufyān al-Thawrī support both sides. Asked about the verse, he answered "shirk". In another report he said, "They are the Muslims", and he cited as evidence the words 'nor those who die while they are disbelievers'.¹³⁸
2. The reports from Ibn ʿAbbās رضي الله عنهما are divided on this question. By two routes he is reported to have said that the evil deeds here are shirk.¹³⁹ His report about the later verse, 4:48, however, says that 4:48 left the people of tawḥīd to Allah's will. That report assumes that 4:18 concerned the believers, and it is related among the reports of those who held this view (al-Ṭabarī).¹⁴⁰
3. Abū al-ʿĀliya and al-Rabīʿ divide the two verses among three groups. They are reported to have said that 4:17 and 4:18 speak of the believers, the hypocrites and the disbelievers in turn. Abū al-ʿĀliya's report that 4:18 concerns the hypocrites has a good isnād (Ḥikmat Bashīr Yāsīn).¹⁴¹

The choice among the reports therefore rests on the argument from the wording. On that argument al-Ṭabarī prefers Sufyān's report that the verse concerns the people of Islam.

The answers to the other three positions. There is a clear answer to each of the other three positions.

1. The hypocrites (the second position). The first group cannot be the hypocrites. First, if the hypocrites meant are those who persist in hypocrisy until death, they have no repentance for the verse to deny. If they repent of hypocrisy before death, they are like any other sinner (al-Ālūsī).¹⁴² Second, the hypocrites are already included in the verse's second group, since they die while they are disbelievers (Abū Ḥayyān).¹⁴³ A defender of the position replies that a hypocrite shows faith, so whoever does not know his inner state gives him the ruling of a believer (Ibn ʿĀdil).¹⁴⁴ That reply concerns how other people see the hypocrite. The verse, however, concerns what Allah accepts, and with Allah a hypocrite who dies in his hypocrisy dies a disbeliever, a conclusion both answers confirm.
2. The disbelievers (the third position). This position rests on the word of the Companions and Successors who witnessed the revelation (al-Wāḥidī).¹⁴⁵ Their word, however, is not unanimous. The same authorities are reported on both sides, and none of the reports answers the argument from the wording.
3. The sinners and the disbelievers together (the fourth position). The exact wording of the verse answers this position as well: 4:18 names the second group by their disbelief, and the first group by their deeds alone. According to this position a disbeliever's faith is refused once death is present. That result follows on the first position too, as the discussion of a disbeliever's faith once death is present shows.

Whether the verse was abrogated. The claim of abrogation rests on a report from Ibn ʿAbbās رضي الله عنهما. By a ḥasan isnad of ʿAlī b. Abī Ṭalḥa, he is reported to have said that after this verse Allah revealed 4:48: "Indeed, Allah does not forgive association with Him, but He forgives what is less than that for whom He wills". By that verse, he said, Allah forbade forgiveness to whoever dies a disbeliever and left the people of tawḥīd to His will.¹⁴⁶ Ibn Abī Ḥātim relates the same report without the words "and left the people of tawḥīd to His will" (ʿAbd Allāh al-Turkī, in his notes to al-Ṭabarī).¹⁴⁷ On the basis of this report, the threat at the end of 4:18 is held to remain in force for the disbelievers. For the disobedient Muslims the threat is held to be abrogated, and their punishment is subject to Allah's will (Ibn Juzayy).¹⁴⁸

Whether the first group are the disbelievers or the believers, the verse needs no abrogation.

1. If the first group are the disbelievers (the third position), 4:48 showed that the general wording of 4:18 did not include the people of tawḥīd. That is specification, not repeal, and the report itself does not use the word 'abrogation' (al-Wāḥidī).¹⁴⁹
2. If the first group are believers (the first position), the two verses fit together without abrogation. Verse 4:18 does not deny forgiveness to a sinner who did not repent. It denies only that one who repents when death is present counts as a penitent, and his forgiveness is then left to Allah's will under 4:48 (Ibn ʿAṭiyya, Abū Ḥayyān).¹⁵⁰

A further objection concerns أَعْتَدْنَا, “We have prepared”: a statement of fact cannot be abrogated (an objection al-Ālūsī records). That objection is enough. A sentence phrased as a statement can establish a legal ruling, and such a ruling can in principle be abrogated (Ibn ʿAṭiyya, Abū Ḥayyān). Here, however, the two verses can be read together in the ways already explained that removes the need to claim abrogation.¹⁵¹

What ends the acceptance of repentance. Repentance is no longer accepted once death is present. At that moment the dying person's breath rattles, and he sees the angels of his Lord who have come to take his soul. Whatever he says then is said outside the state in which repentance is possible (al-Ṭabarī).¹⁵² This is reported as the view of Ibn ʿAbbās رضي الله عنهما, Ibn Zayd and the majority of the exegetes (al-Qurṭubī).¹⁵³ At that moment no repentance is accepted from a sinner, and no faith from a disbeliever. In the same way, Pharaoh's declaration of faith when he was drowning did not benefit him (al-Baghawī).¹⁵⁴

Proximity to death does not preclude acceptance. The presence of death in 4:18 is not the same as its nearness. One who is only near death may still repent. Once death is present, however, the dying person sees the states from which there is no return to this world (al-Khāzin, al-Ālūsī, al-Durra).¹⁵⁵ The pains of dying do not preclude repentance by themselves either. Allah caused some people to die and then brought them back to life, among the Children of Israel and the children of Ayyūb, and they were charged with obligations again. The pains of one near death, moreover, are like the pains of colic or of childbirth, and those pains leave obligation in place (al-Rāzī, reporting the rigorous scholars). The argument also points to Allah’s answering the distressed person in 27:62. Pain by itself is therefore not a reason for refusing repentance; on this account, distress gives a further reason to hope for acceptance.¹⁵⁶

The refusal at that moment rests on Allah's word. By His grace He promised in 4:17 to accept a repentance made before that moment. By His justice He stated in 4:18 that He would not accept a repentance made in it. The refusal does not rely on the theological premise of the Muʿtazila. In their account, the sight of that moment makes the dying person's knowledge of Allah compelled, and this by itself ends his accountability (al-Rāzī; Abū Ḥayyān, who records that the Muʿtazila were answered).¹⁵⁷ Al-Rāzī gives three answers to the claim that the dying person’s knowledge necessarily ends his responsibility. First, even those raised from the dead see that they live again and see the Fire; the conclusion that these things require a Maker still involves inference, according to most of the Muʿtazila’s own teachers. Second, even immediate knowledge of Allah would not itself remove choice: a person might still sin while relying on His generosity and knowing that obedience does not benefit Him nor sin harm Him. Third, genuinely certain knowledge is not made more certain merely by being immediate rather than inferred. These arguments lead him back to revelation: the time of acceptance ends because Allah has declared that limit, not because the proposed explanation of knowledge proves that repentance must be impossible.¹⁵⁸

What that moment is. The scholars describe that moment in several ways. Each description names something that repentance requires and that the dying person no longer has. That is why a repentance made at that moment is not the repentance that 4:17 promises to accept, while the limit itself was set by Allah's word.

1. Choice has ended. A repentance made at that moment is compelled by what the dying person sees, like the act of one who is coerced. Acceptance was promised as a reward only to one who repents by choice (al-Nasafī, Ibn ʿUthaymīn).¹⁵⁹
2. The next world has begun. The presence of death is the first state of the next world. The dying person sees its realities, and his obligations have ended, as they have for one who has already died (al-Zamakhsharī, al-Wāḥidī).¹⁶⁰
3. The sin can no longer be abandoned in practice. Repentance becomes real when the sinner abandons the sin in his conduct. At that moment this is no longer possible (al-Zajjāj, Makkī).¹⁶¹
4. The purpose of resolve is gone. The resolve not to return exists for its effects in this life, in the reform of one's conduct. Once life is despaired of, that purpose is gone (Ibn ʿĀshūr).¹⁶²

One explanation goes further than these descriptions. It argues that 4:18 denies that a sound repentance occurs from such a person at all. This is why 4:18 does not say عَلَى اللَّهِ, "upon Allah", as 4:17 does. What the dying person offers is not repentance in the law's sense (ʿAbduh, as Rashīd Riḍā reports him; Ibn ʿUthaymīn).¹⁶³ Another explanation takes 4:18 to deny only that such a repentance is certainly accepted. Denying that something is certain does not establish that it does not happen, and so this explanation leaves the case to Allah (al-Marāghī).¹⁶⁴ The immediate sense of the verse, however, is the first explanation's: such a repentance counts for nothing and is not accepted (Rashīd Riḍā, following ʿAbduh, and al-Ālūsī).¹⁶⁵

All these descriptions lead to one result. Verse 4:18 makes one who repents only when death is present equal to one who dies a disbeliever (al-Zamakhsharī, al-Bayḍāwī, al-Shanqīṭī).¹⁶⁶

A view in the Ḥanafī school. This view accepts the repentance of a believer who has despaired of life, while it refuses a disbeliever's faith in that state. It holds that hope still remains with the believer, so his regret and his resolve to abandon the sin are still sound. It also holds that his repentance renews a bond with his Lord that already exists. A disbeliever's faith, by contrast, would have to begin a bond that never existed (a view al-Ālūsī reports, with al-Bazzāziyya's judgment that acceptance is the sound view).¹⁶⁷ Al-Ālūsī himself holds that the apparent sense of the verse supports the other view: the repentance of one who has despaired is refused, like his faith.¹⁶⁸ Within the school, ʿAlī al-Qārī answers the distinction from the Prophet's ﷺ words, "Allah accepts the repentance of the servant so long as he does not rattle". These words cover the repentance of the believer and the faith of the disbeliever alike (as Ibn ʿĀbidīn reports him).¹⁶⁹ Its supporters also reason that, if intercession is accepted at the Resurrection when worldly life is beyond hope, repentance before death has an even stronger claim. They appeal to the general promise that Allah accepts repentance in 42:25. Ibn ʿĀbidīn records the believer’s case as a matter of conjectural judgment; the agreement he reports concerns the refusal of a disbeliever’s faith at despair.¹⁷⁰

The disagreement concerns the believer who witnesses the terrors of death, not merely someone dangerously ill. This tafsīr follows ʿAlī al-Qārī’s answer: the hadith sets a specific limit for the servant. On this interpretation, the general promise of acceptance remains confined within that limit. Acceptance of intercession is a separate question from acceptance of a repentance made after the appointed limit.

ʿAbduh offers a further interpretation of ‘disbelief’ here, below rejection of the prophetic message. Its reasoning is developed under How the Commentators Read This Verse. The interpretation followed here retains the ordinary sense of dying in disbelief.¹⁷¹

The second group. The phrase وَلَا الَّذِينَ يَمُوتُونَ وَهُمْ كُفَّارٌ, "nor those who die while they are disbelievers", is explained in two ways.

1. Those who actually die in disbelief. This is the apparent sense of the phrase. The repentance denied to them is the remorse they show on the Day of Resurrection, when the time for deeds has ended (Ibn ʿUthaymīn and Ibn Kathīr, with al-Rāzī in one of his explanations).¹⁷²
2. Those who come near to death in disbelief. In this explanation 'die' means 'come near to death'. A disbeliever's faith is then refused at the point of death, as a sinner's repentance is (al-Rāzī in his other explanation).¹⁷³

The second explanation begins with the sentence’s structure: repentance is denied to those described as dying in disbelief, yet no repentance can be performed after death. Its proponents therefore understand ‘die’ as approaching death, comparing the anticipatory wording in 4:9 and 2:240. Pharaoh’s case makes a further argument, but it is not their only reason.¹⁷⁴ Ibn ʿĀshūr answers the proof from Pharaoh by distinguishing faith at the descent of punishment from faith at an ordinary death. He cites Allah’s refusal of faith after punishment has descended, while expressly excepting the people of Yūnus in 10:98. His preferred apparent sense also allows the denied repentance to mean remorse shown in the Hereafter.¹⁷⁵

The disbelievers are named in a verse about repentance because a disbeliever's faith is his repentance from disbelief, and faith is the noblest kind of repentance (Ibn ʿĀshūr).¹⁷⁶ Arabic usage agrees. Islam was called "the time of repentance", and a poet uses the phrase in this sense:

Al-Zamakhsharī cites the line, by al-Jaʿdī, for "the time of repentance" as a name for Islam, because in Islam people repent of shirk.¹⁷⁷

A disbeliever's faith once death is present. The apparent sense of the phrase leaves one case to settle: a disbeliever who says "I believe" when death is already present, but before he has died. Before that moment the door is open to both the sinner and the disbeliever. After death the door is closed to both. The difference concerns only the disbeliever, and only at that moment.

The majority hold that the disbeliever's faith is refused at that moment, as a sinner's repentance is, and their view is adopted here. It is al-Shanqīṭī's tarjīḥ: he takes the repentance of the disbelievers that 3:90 refuses to be a repentance made when death is present.¹⁷⁸ The majority's view rests on five proofs:

1. The verse pairs the two groups. Verse 4:18 sets those who put off their return until death is present beside those who die in disbelief. It does so to show that a return at that moment counts for nothing (al-Bayḍāwī, al-Shanqīṭī).¹⁷⁹
2. The time of choice has passed. The reason holds for faith as it does for repentance. The presence of death is the first of the states of the next world, when the time of obligation and choice has passed (al-Zamakhsharī, al-Qurṭubī).¹⁸⁰
3. Faith at the sight of punishment did not benefit. The Qurʾān says of those who declared their faith once they saw Allah's punishment that their faith did not benefit them (40:84–85). Pharaoh's declaration of faith when drowning overtook him did not benefit him either (al-Baghawī, Ibn Kathīr).¹⁸¹
4. The Prophet ﷺ set one limit. He set the death rattle as the limit for the servant, and he did not distinguish a believer's repentance from a disbeliever's faith.¹⁸²
5. The Ḥanafī view agrees on the disbeliever. Even the Ḥanafī view that accepts a believer's repentance at that moment refuses a disbeliever's faith, and Ibn ʿĀbidīn reports agreement on it.¹⁸³

Ibn ʿĀshūr's view, and the answer to it. Ibn ʿĀshūr draws the opposite conclusion from the apparent sense. In his view nothing closes the door on a disbeliever's faith but death itself, so the disbeliever's faith is accepted even once death is present.¹⁸⁴ Among the commentators examined here, Ibn ʿĀshūr alone explicitly adopts this general distinction. Ibn ʿĀshūr gives three reasons for it,¹⁸⁵ and there is a clear answer to each of his points.

1. The two clauses are worded differently. Verse 4:18 joins وَلَا الَّذِينَ يَمُوتُونَ وَهُمْ كُفَّارٌ, "nor those who die while they are disbelievers", as a case different from the sinner's حَتَّى إِذَا حَضَرَ أَحَدَهُمُ الْمَوْتُ, "until, when death comes to one of them". The different wording, however, only names the second group by the state in which they die. It does not say that a disbeliever who turns at the presence of death is received. What 4:18 refuses in the sinner is a return made once death is present, and a disbeliever's faith at that moment is such a return.
2. The invitation to Abū Ṭālib. Ibn ʿĀshūr cites the Prophet's ﷺ invitation to Abū Ṭālib to believe when death came to him, as evidence that faith is still accepted at that moment. Historical Context explains the invitation as occurring before the terrors of death. It also records Ibn ʿUthaymīn’s possibility of an exceptional case. Neither answer establishes a general extension of the time for faith.
3. Faith is not a sinner’s repentance. Faith is an act of the heart and a word of the tongue. While the person still lives, those acts have been completed: he joins the Muslims, strengthens their community and spreads the reputation of Islam among disbelievers. Ibn ʿĀshūr contrasts these immediate effects with a dying sinner’s resolve to reform a life that is ending. These effects explain his distinction, but do not by themselves establish an exception to the revealed time limit. This tafsīr follows the majority in taking the hadith’s limit for the servant to include both the sinner’s repentance and the disbeliever’s faith.¹⁸⁶

Ibn ʿĀshūr’s qualification concerning Pharaoh has been explained above. It does not establish that faith is accepted once death is present: the majority apply the refusal of faith at the sight of punishment to the dying person who sees the angels coming for his soul (al-Baghawī, Ibn Kathīr).¹⁸⁷

A similar view is reported from some Sufis, together with a hope that Ibn ʿUmar رضي الله عنهما expressed for a polytheist who declares Islam at the rattle. Al-Ālūsī, who reports both, holds the apparent sense of the verse against them, and the report of Ibn ʿUmar's hope has an unnamed narrator in its chain.¹⁸⁸ Some support the alternative with a report from Abū Dharr, who reports the Prophet ﷺ saying the words below. Its explanation of the veil is part of the transmitted report itself:

On that wording they place the limit at the soul’s departure. The report’s grading is disputed: al-Ḥākim judged its chain authentic, while the editors of Musnad judged it weak because two narrators are unknown. The adopted explanation therefore does not rest on treating this report as uncontested evidence.¹⁹⁰

For whom the punishment is prepared. Verse 4:18 closes with the threat أُولَئِكَ أَعْتَدْنَا لَهُمْ عَذَابًا أَلِيمًا, "those: for them We have prepared a painful punishment". The word أُولَئِكَ, "those", can refer back in two ways.

1. The nearer group only. 'Those' refers to those who die while they are disbelievers, because a demonstrative, like a pronoun, refers back to the nearest thing named. The punishment is then the everlasting punishment of the disbelievers (al-Rāzī, Ibn ʿUthaymīn, and al-Samīn, who allows both references).¹⁹¹
2. Both groups. The punishment is then everlasting for the disbelievers. For the sinners, it is a punishment that does not last forever (Ibn ʿAṭiyya, al-Qurṭubī, al-Zuḥaylī).¹⁹²

The waʿīdiyya were the theologians named for the threat (waʿīd), because they held that Allah always fulfils His threats against sinners. They inferred from 4:18 that the sinners among the people of prayer who neglect repentance until their death share the punishment of the disbelievers (as al-Rāzī, al-Wāḥidī and al-Khāzin report).¹⁹³ Verse 4:48 answers that interpretation, because it is the decisive verse to which such threats are referred. Every threat made to sinners who died as Muslims is qualified by its words, "He forgives what is less than that for whom He wills" (Abū Ḥayyān).¹⁹⁴ On the second reference, too, the sinner's punishment is not everlasting. A second objection asks what the refusal of late repentance would accomplish if Allah might still forgive the sinner. Al-Rāzī answers that the possibility of punishment remains a real cause for fear even when pardon is possible. Refusing repentance does not have to make punishment inevitable in order to matter. He also objects that this reasoning depends on inferring an unstated consequence from the wording, a method which, he says, the Muʿtazila reject.¹⁹⁵ The first reference, however, is the stronger, because 'those' refers back to the nearest group named (al-Rāzī, Ibn ʿUthaymīn).¹⁹⁶ The painful punishment is therefore prepared for the disbeliever. The believer who persisted until death is a sinner whose end is left to Allah's will. Fear predominates in his case, and revelation shows that Allah forgives some such people by His grace (Ibn ʿAṭiyya, Ibn ʿUthaymīn).¹⁹⁷

Which interpretation is stronger? Regarding the first question, the initial position is stronger. It uniquely preserves an independent meaning for the clause 'nor those who die while they are disbelievers,' and it respects the explicit distinction 4:18 draws between the two groups. Because the hypocrites already fall under the second group, and the early reports remain conflicted, this interpretation is adopted here, reflecting al-Ṭabarī's formal tarjīḥ and Ibn ʿĀshūr's later agreement. Furthermore, the verse is not abrogated: 4:48 simply determines whose end is left to Allah's will, as Ibn ʿAṭiyya and al-Wāḥidī explain, each according to his own interpretation of the first group. Finally, the second group designates those who actually die in disbelief, which is the apparent sense preferred by Ibn ʿĀshūr.¹⁹⁸ A disbeliever's faith once death is present is refused, as the majority hold. Ibn ʿĀshūr's further conclusion, that it is accepted until death itself, is not acceptable.

Conclusion: The second verse denies the acceptance of repentance to the believer who persists in sin until death is present and to whoever dies a disbeliever; the painful punishment is prepared for the disbeliever, while the believer's end is left to Allah's will.

And Allah knows best.

Q5. What makes a repentance sound, and whether it is sound from one sin while another is kept

The first verse promises acceptance to those who do evil and then يَتُوبُونَ (yatūbūna), "repent", soon after. The second refuses it to the one who, when death comes to him, قَالَ إِنِّي تُبْتُ الْآنَ, "says, 'I repent now'". The Qurʾān uses 'says' for the dying man rather than 'repents': it attributes to him the words of repentance and not repentance itself, and so declines to call what he offers a repentance (Ibn Kamāl Pāshā, Abū al-Suʿūd).¹⁹⁹ Repentance is therefore an act a person performs, not a phrase he speaks. Two questions need answering:

1. The conditions: what a repentance must include in order to be sound.
2. One sin among others: whether a repentance from one sin is sound while the penitent persists in another, and whether a later relapse undoes it.

The definitions answer the first question in different lists that describe one act; the second question raises different views about repentance from one sin while another continues, which is settled after the conditions are explained.

Repentance is an obligation, and an immediate one. Every believer is obliged to repent. The community agrees on this, and its agreement establishes that the command "And turn to Allah in repentance, all of you, O believers" (24:31) is an obligation and not only a recommendation (as Ibn ʿAṭiyya, al-Qurṭubī, Abū Ḥayyān, al-Shawkānī, al-Zuḥaylī and the Majmaʿ's al-Tafsīr al-Wasīṭ report it).²⁰⁰ The obligation is also immediate: that repentance is due at once is beyond doubt (al-Qāsimī).²⁰¹ An immediate duty, however, is not a condition of the promise. Question 2 found that the promise extends to any repentance made before the throes of death; the immediacy of the duty means only that delaying repentance is itself a fault. The conditions explained below are therefore what this duty requires of the sinner as soon as he has sinned.

What the word 'repentance' means. The meaning of the word confirms that repentance is something done. التَّوْبَةُ (al-tawba), "repentance", means return: the penitent turns back from disobedience to obedience, in apology to Allah (Ibn Fāris, Ibn Manẓūr, al-Biqāʿī, Mawdūdī).²⁰² Its most precise sense is desisting: the person ceases from the sins he was committing, so that repentance is the abandonment of persistence in disobedience (Jabal, who prefers the definition of al-Miṣbāḥ).²⁰³ The second verse shows the same meaning from the other side. The dying man can no longer abandon the sin in his conduct, which is one of the reasons Question 4 gives for the refusal, and his word الْآنَ (al-āna), "now", suggests that he has no resolve to keep away from the sin were he to live on, a resolve that a sound repentance requires (a view al-Ālūsī reports).²⁰⁴ The conditions of repentance are therefore acts of the heart and of conduct, and words of repentance without them are not what the first verse promises to accept.

The three conditions at the centre. The definitions centre on three conditions: the sinner regrets the sin, abandons it, and resolves not to return to it (al-Nawawī, al-Jazāʾirī).²⁰⁵ Each condition has a definite object. The regret must be for the act as disobedience to Allah: regret because the act harmed one's body or property is not repentance. Where the act is one the sinner could commit again, the resolve to abandon it is a condition, since without that resolve he is persisting in the sin, and persistence cannot coexist with repentance (Ibn ʿAṭiyya).²⁰⁶ The sin must also be abandoned at once, and all of this must arise from shame before Allah and not before anyone else; if any one of these conditions is missing, the repentance is not sound (al-Qurṭubī).²⁰⁷

What these conditions exclude is a repentance of the tongue alone. Allah accepts a sincere repentance, in which the sinner recognizes the ugliness of what he has done, regrets it and loses his love for it, so that he would not return to it if he lived on; He does not accept asking forgiveness with the tongue, or a few acts of expiation such as charity or words of remembrance, while the sinner persists in his sins (al-Marāghī, and ʿAbduh as Rashīd Riḍā reports him).²⁰⁸ A sincere repentance is therefore a change in the person and not a formula of words: past mistakes are sincerely regretted, and the regret changes his whole attitude (Sayyid Quṭb).²⁰⁹ Regret, abandonment and resolve all concern the same act and together turn the sinner away from it, so they form one act of return, and a repentance that lacks any of them is not the return the word denotes.

Making good what can be made good. Where a sin has left something owed, the return is not complete until it is restored. Repentance in the law also includes making up, as far as possible, the duties that were neglected, by performing them; this is a fourth element beside regret, abandonment and resolve, and with it the conditions of repentance are complete (al-Rāghib).²¹⁰ Where the sin wronged another person, repentance includes returning his right to him as far as possible (al-Durra, Muḥammad al-Hilāl, Muftī Shafīʿ).²¹¹ The definitions do not all list the same elements, but they describe one act: shame before Allah, which one definition lists as a condition of its own, is the motive of the whole repentance rather than an additional act, making good applies wherever a right or a neglected duty remains, and the time, examined next, is the frame within which every condition must be met.

The time within which the conditions count. A repentance is sound only if it is made at a time when repentance is accepted, and not after the appointed term has come or the sun has risen from the west (Ibn ʿUthaymīn).²¹² Put positively, the accepted repentance is one made while there is still hope of life, together with the wish to reform one's conduct afterwards (the Majmaʿ's al-Tafsīr al-Wasīṭ).²¹³ Al-Ṭabarī connects the limit to the ability to understand regret and choose to leave the sin. He expects the agony of dying to overwhelm that ability, but includes the person who still reasons like a healthy person and sincerely turns back within the promise, if Allah wills.²¹⁴ The time limits established in Questions 2 and 4 are therefore themselves a condition of a sound repentance.

Repentance from one sin while another is kept. The second question arises because a person may repent of one sin while he continues in another. One view holds that such a person is not a penitent at all (the Muʿtazila, as Ibn ʿAṭiyya and al-Qurṭubī report them).²¹⁵ On this view repentance is not sound until the sinner has abandoned every other sin, so that his reform is real, and the view appears to rest on what the Muʿtazila take acceptance to mean: Allah's pleasure with the penitent and his inclusion among the righteous, which a repentance that leaves other sins in place would not achieve (Ibn ʿĀshūr, who infers it from their condition).²¹⁶

The definition of repentance answers this view. The regret is for the act that was committed, and the resolve is to abandon that act; neither condition concerns any other sin. A repentance that meets these conditions for one sin is therefore sound for that sin, even while the penitent persists in another sin of a different kind. This is the position of Ahl al-Sunna, against the Muʿtazila and some who ascribe themselves to the Sunna but follow them on it, and it is reported as the view of the majority (al-Qurṭubī, al-Nawawī, Ibn ʿAṭiyya, Abū Ḥayyān, al-Zuḥaylī, with al-Shawkānī, who reports the majority).²¹⁷ A further view distinguishes between sins of the same kind and sins of different kinds. On that account, abandoning an unlawful gaze while continuing unlawful touching is not a full withdrawal from that kind of sin. Ibn ʿUthaymīn reports this distinction before giving his own conclusion below.²¹⁸ Ibn ʿUthaymīn states the tarjīḥ and its limit: the sound view is that repentance from one sin is accepted while the sinner persists in another, but he does not then earn the unqualified description of 'the penitents'; he is a penitent from that particular sin, because to describe him as a penitent from that sin only is justice, and Allah commands justice.²¹⁹ The limit falls on the name and not on the repentance: the one who persists in other sins is not called a penitent without qualification, yet the repentance he has made from that sin is sound and accepted.

A relapse does not undo a repentance. A related question follows from the resolve. If the resolve not to return is a condition, a later return to the sin might seem to show that the resolve was never real and that the repentance was void, and the Muʿtazila held this view as well (al-Nawawī).²²⁰ A repentance made with its conditions, however, is an act of obedience that was completed and was sound. Returning to the sin afterwards is a new sin, which requires a new repentance; it does not undo the first (Ibn ʿAṭiyya, al-Nawawī, Abū Ḥayyān).²²¹ The resolve is therefore required when the sinner repents, and a later fall does not close the way back to Allah.

Conclusion: A repentance is sound when the sinner regrets the sin as disobedience to Allah, leaves it, resolves not to return and makes good what he can while repentance is still accepted; it is sound from one sin while he keeps another, though he is not called a penitent without qualification, and a later relapse does not undo it.

In the rulings of this world, a repentance that is shown is accepted by its outward form, since no human being knows its reality except the one who repents, and knowledge of the unseen belongs to Allah alone (al-Shāfiʿī).²²² Repentance does not, however, remove a ḥadd once testimony has established it, so that a thief or a slanderer who repents still receives the penalty (al-Qurṭubī, who knows of no difference on it, though his muḥaqqiqs point to the difference he records on the question elsewhere).²²³

And Allah knows best.

2. How the Commentators Read This Verse

The following accounts take up whether the second verse's two clauses can describe a single group, and how its word 'disbelievers' can be applied to believers without declaring them disbelievers.

3. Iʿrāb: Grammatical Structure

The word and clause tables use Dār al-Ṣaḥābah’s Iʿrāb al-Qurʾān al-Karīm as their baseline.²³³ In the tables, rafʿ, naṣb and jarr are the three grammatical states of a noun, shown mostly by its last vowel: rafʿ is the state of a subject or a predicate, naṣb that of an object and of words such as a ḥāl, and jarr that of a noun after a preposition or after another noun it completes. The explanations after them show the alternatives and what they do, or do not, change in English.

Word / Grammar / What it does

4:17

4:18

Clause / Position

A clause can stand as a fresh statement, complete another clause, or occupy a noun’s grammatical position. The table follows the baseline.²³⁵

Where the grammar changes the meaning

What completes “repentance”?

In the first verse, al-tawba, “repentance”, is the subject, and two phrases stand after it: عَلَى اللَّهِ (ʿalā Allāhi), “upon Allah”, and لِلَّذِينَ (li-lladhīna), “for those who”. One of them completes the statement as its predicate. In the baseline, “upon Allah” is attached to the understood predicate and “for those who” to an understood ḥāl: repentance rests upon Allah, for those who do evil in ignorance.²³⁶ Ibn ʿUthaymīn allows either phrase to serve as the predicate. Al-Samīn prefers ‘upon Allah’ as the predicate, but attaches ‘for those who’ to the same understood governor. He regards the alternative treatment of ‘for those who’ as a separate ḥāl as unnecessarily strained.²³⁷

The second account reverses the roles. “For those who” is the predicate, and “upon Allah” describes the repentance as a ḥāl, through an understood “when it is”: repentance, when it is upon Allah, is for those who do evil (al-ʿUkbarī, Darwīsh).²³⁸ A ḥāl cannot normally come before a governor that works through meaning alone, that is, a word that is not a verb and governs through its meaning. al-ʿUkbarī meets the difficulty with the Arabs’ saying “this, as unripe dates, is better than it as ripe ones”, where “as unripe dates” is likewise a ḥāl that comes first, resting on an understood “when it was”.²³⁹

On the first account a question remains: repentance is the servant’s act, so how does it rest upon Allah? An answer understands two dropped words, “the acceptance of repentance rests upon the bounty of Allah”, which lets ʿalā keep its proper sense of what lies upon something (Abū Ḥayyān, whose account al-Samīn marks as open to question without saying on what ground; al-Jadwal).²⁴⁰

Effect on English. Neither account changes the rendering, because English word order does not show which of the two phrases is the predicate. “Repentance is only upon Allah for those who …” is right on both, and the dropped words make explicit what the English already implies: what Allah takes upon Himself is the acceptance of repentance. What ʿalā makes of that acceptance is argued in Question 3. In the second verse there is no such choice, since “for those who” is the predicate of laysa.

In ignorance, or because of ignorance?

The baseline attaches بِجَهَالَةٍ (bi-jahālatin) to an understood ḥāl.²⁴¹ On that account the bāʾ is one of accompaniment: they do evil clothed in jahāla, with it at their side (al-Samīn, Ibn ʿĀshūr, Darwīsh).²⁴² The ḥāl could also describe the object, “evil clothed in jahāla”, but ignorance belongs to the one who acts and not to the deed, and al-Samīn calls that account far-fetched and loose.²⁴³

The other account makes it the باء السببية (bāʾ al-sababiyya), the bāʾ of cause: jahāla is what drove them to the evil, because had they kept in mind what follows a sin while doing it, they would not have ventured on it. Reason at that moment is overcome or taken away (Abū Ḥayyān).²⁴⁴

Effect on English. The ḥāl gives “they do evil in ignorance”; the cause gives “they do evil out of ignorance”, and the table follows the baseline’s ḥāl. Neither construction decides what jahāla is, want of knowledge or folly; that is argued in Question 1.

“Soon”: the adjective without its noun

قَرِيبٍ (qarībin), “near”, describes a noun that is not expressed, “a near time”. An adjective standing in for its noun in this way is taken from the Arabs’ usage, not formed by rule (al-Samīn, Ibn ʿĀshūr).²⁴⁵ The two accounts of min, the start of a span or a part of it, and the rule of the grammarians of Basra that weighs against the first, are set out in Question 2. The English “soon” renders the adjective together with its unexpressed “time”.

Why “turns to them” takes ʿalā

When the servant turns in repentance, he turns ilā Allah, “to” Him. When Allah turns, the verse says يَتُوبُ اللَّهُ عَلَيْهِمْ. The verb is given the sense of yaʿṭifu, “He inclines in kindness”, and that verb takes ʿalā (al-Samīn; Lisān al-ʿArab).²⁴⁶ English has no preposition that does this, so “turns to them” keeps the turning and loses the kindness that the Arabic ʿalā adds to it.

What “until” limits

حَتَّى (ḥattā) is here a particle of starting, not a preposition. The whole conditional sentence after it, “when death comes to one of them, he says: I have repented now”, is the limit of what comes before: repentance is not for people whose doing of evil runs on until that saying. إِذَا (idhā), “when”, is attached to “he says”, so the saying falls at the moment death comes (al-Samīn, who calls this a good account; Dār al-Ṣaḥābah).²⁴⁷ English “until” marks the same limit.

“Nor those who die”: joined to the denial, or a new sentence?

In the adopted account, وَلَا (wa-lā) is the wāw of joining with the negating lā, which repeats the negation, and الَّذِينَ (alladhīna) is in the place of jarr, joined to the first “for those who” (Dār al-Ṣaḥābah, al-Ṭabarī, Ibn ʿUthaymīn, al-Samīn, al-Jadwal). Ibn Hishām prefers it in al-Mughnī, as al-Durra reports.²⁴⁸ The sense is: repentance is not for these, nor for these. One negation covers both groups, the one who delayed until death came and the one who died a disbeliever, and so the verse makes the two equal in that their repentance is not accepted (al-Samīn).²⁴⁹

Another account takes the lām as the لام الابتداء (lām al-ibtidāʾ), the lām of emphasis that opens a sentence: “and surely those who die while they are disbelievers, those, We have prepared for them a painful punishment”. Alladhīna is then a subject, and أُولَئِكَ (ulāʾika) with what follows is its predicate. Ibn al-Jazarī records this from al-Akhfash and found it in Abū al-Baqāʾ al-ʿUkbarī’s Iʿrāb as well.²⁵⁰ al-Akhfash also made “I have repented now” a complete stop, a point at which the sentence is finished, so that the reciter may pause there and begin afresh with what follows; only his account allows it: if “nor those who die” is joined to the first “those who”, the sentence runs on and there is no stop there (Makkī).²⁵¹

The account fails on the written text. A lām of the start could stand only if a lām had been written on alladhīna, making it wa-la-lladhīna; the maṣāḥif write wa-lā, a lām and an alif, the negating lā (al-Samīn, al-Suyūṭī, and Ibn al-Jazarī, who grants that the account is sound as grammar but for the written form).²⁵² al-Nasafī reports that some maṣāḥif write it with two lāms, which would give the subject account.²⁵³ His report names no muṣḥaf, while the scholars of the written text describe the maṣāḥif as writing wa-lā, so the adopted account follows them.

Effect on English. Here the difference is real. On the joined account the English is “nor for those who die while they are disbelievers”: both groups are denied repentance. On the account of the lām of the start, the verse would no longer say that repentance is not for the disbelievers; it would only threaten them: “and those who die as disbelievers, for them We have prepared a painful punishment”. Whom “those” points to on the adopted account, the nearer group or both, is covered in Question 4.

4. Ṣarf: Roots and Word Formation

The table follows the root-bearing words of the two verses in order, repeated words included. Particles, demonstratives, pronouns, relative nouns and the divine name are left to the grammatical tables. رَفْع (Rafʿ), نَصْب (naṣb) and جَرّ (jarr) are a noun’s grammatical states; the imperfect verbs of these verses are all in رَفْع (rafʿ). A “hollow” root has a weak letter in its middle.

Prepared: from الْعَتَاد (al-ʿatād), or from أَعْدَدْنَا (aʿdadnā)?

أَعْتَدْنَا (aʿtadnā) means “We have prepared”, and lexicographers trace its construction to two distinct origins. On the first, it is أَفْعَلْنَا (afʿalnā) from الْعَتَاد (al-ʿatād), the thing stored up before it is needed. The root is then one of its own, whose single sense is presence and nearness: what is prepared is at hand (Ibn Fāris, al-Rāghib, Jabal).²⁶⁸ al-Azharī held this the more correct: عتد (ʿatada) is a structure of its own and عدّ (ʿadda) a doubled one.²⁶⁹ Lisān al-ʿArab cites the verb itself from an unnamed poet, after “We have prepared” of 18:29, for a thing kept ready at hand:

The Arabs’ usage stands behind the first way: the noun عَتَاد (ʿatād) and the adjective عَتَد (ʿatad) both name what is made ready and kept at hand, and al-Thaʿlabī cites a line for each at this verse. ʿAdī b. al-Riqāʿ uses الْعَتَاد (al-ʿatād) for the equipment gathered for war:

al-Asʿar al-Juʿfī uses عَتَد (ʿatad) for the horse kept ready for running:

On the explanation recorded in Lisān al-ʿArab, they leave their father’s blood unavenged, while the poet pursues vengeance. The witness concerns عَتَد (ʿatad), the horse kept ready for war.²⁷³

Under the second derivation, the verb is أَعْدَدْنَا (aʿdadnā), “We have made ready”, its first د (dāl) changed to ت (tāʾ) because the two letters share one place of articulation and often replace each other (al-Jadwal, Darwīsh, and Yaʿqūb as Lisān al-ʿArab reports him, while al-Rāghib and al-Samīn give both ways).²⁷⁴ Lisān al-ʿArab also reports the reverse claim, that أَعَدَّ (aʿadda) is itself أَعْتَدَ (aʿtada) with ت (tāʾ) assimilated, and that others reject it: أَعَدَّ (aʿadda) comes from ع (ʿayn) and two instances of د (dāl), because the Arabs say أَعْدَدْنَاهُ (aʿdadnāhu) and pronounce both.²⁷⁵ al-Wāḥidī gives that proof with a line he ascribes to Imruʾ al-Qays, which shows both instances of د (dāl); the poet, he notes, did not say أَعْتَدْتُ (aʿtadtu):

What the form establishes. The two instances of د (dāl) of أَعْدَدْنَاهُ (aʿdadnāhu) settle that أَعَدَّ (aʿadda) is not a changed أَعْتَدَ (aʿtada); they leave open whether أَعْتَدْنَا (aʿtadnā) is a changed أَعْدَدْنَا (aʿdadnā) or a root of its own. The lexicographers leave the choice open, with one preference, al-Azharī’s for the separate root, and the English is the same on either way: “We have prepared”. The first adds that what is prepared is present and at hand; a further sense is reported from الْعَتَاد (al-ʿatād) as what is fixed and binding, “We have fixed it and made it a settled matter” (al-Samīn).²⁷⁷ On either way the verb is past: the punishment prepared for them is the Fire, and because it is already prepared, the Fire is already created (al-Rāzī, Ibn ʿAṭiyya).²⁷⁸

Painful: the punishment that causes pain

أَلِيمًا (alīman) is on the pattern فَعِيل (faʿīl) in the sense of مُفْعِل (mufʿil): the painful is what causes pain, as سَمِيع (samīʿ) is used for مُسْمِع (musmiʿ), the one who makes others hear (Lisān al-ʿArab, al-Rāghib, al-Zajjāj; Abū ʿUbayda, as al-Samīn reports him).²⁷⁹ The verb أَلِمَ (alima) describes the one who suffers, and the one who passes the pain on is أَلِيم (alīm) (Ibn Fāris, from al-Khalīl).²⁸⁰ Lisān al-ʿArab cites Dhū al-Rumma for a blaze that pains:

The blaze does not suffer pain; it inflicts it, and so does the punishment of the verse. Ibn ʿArafa took أَلِيم (alīm) as “possessed of pain”, the one who has it, and said he did not know what Abū ʿUbayda meant. The blaze answers him, since it has no pain and gives it, and al-Samīn replies that Abū ʿUbayda’s statement is clear.²⁸² The lexicographers add a degree: the painful punishment is the one whose paining reaches the utmost (Lisān al-ʿArab, al-Zajjāj).²⁸³

In ignorance: the two senses of جَهْل (jahl)

The جَهَالَة (jahāla) of the verse needs a sense in which a person can know that an act is forbidden and still do it in جَهَالَة (jahāla), and the root has one. Ibn Fāris gives the letters ج (jīm), ه (hāʾ) and ل (lām) two roots: the opposite of knowledge, and lightness, the opposite of calm. The stick that stirs the embers is a مِجْهَل (mijhal), and the wind اسْتَجْهَلَتْ (istajhalat) the branch when it moved it and it shook.²⁸⁴ His witness for the second root is a line of al-Nābigha’s:

اسْتَجْهَلَتْكَ (Istajhalatka) means “made you light and stirred you”: the abodes did not leave the poet without knowledge of anything; they stirred him and made him light. al-Zamakhsharī, in Asās al-Balāgha, cites the same line as a figurative use, “made you light”, so for him the lightness grows out of one root rather than standing as a second.²⁸⁶ Jabal also finds one root and gives the link: an inside empty of what is useful, with a dryness, from which lightness follows, so that جَهْل (jahl) as the opposite of حِلْم (ḥilm) shows in conduct as lightness, rashness and folly.²⁸⁷ Whether lightness is a second root or a branch of the first, the usage is the same: جَهْل (jahl) can name the loss of self-command in someone who knows. That is the folly Question 1 finds in every sin, deliberate or not.

Now: the form of الْآنَ (al-āna)

الْآنَ (al-āna) names the time the speaker is in. It is an adverb that does not take the full range of endings (al-Jawharī in Lisān al-ʿArab), and its فَتْحَة (fatḥa) is fixed (Dār al-Ṣaḥābah).²⁸⁸ Its origin is أَوَان (awān), with one ا (alif) dropped and و (wāw) changed to ا (alif) (al-Farrāʾ, as Lisān al-ʿArab reports him).²⁸⁹ Jarīr, whom Ibn Barrī names, uses the same word for the present moment:

With الْآنَ (al-āna), the past تُبْتُ (tubtu) is a past that touches the moment of speaking, “I have repented now” (al-Sāmarrāʾī).²⁹¹ The one who says it places his repentance in the moment of his speaking, and the verse has just named that moment: when death comes to him.

5. Balāghah: Rhetorical Analysis

The two verses are a promise and its threat. Each gives its outcome under one demonstrative, "those", and the first closes on its own limit. Between these two frames the wording marks, or sets up, at five places the line between those inside the promise and those outside it: the far "those", the "then" before "soon", the imperfect "do", the moment of death, and the naming of those who die as disbelievers. The contrast between "evil" in the singular and "evil deeds" in the plural is developed among the Gems, and "he says" in place of "he repents" in Deeper Study, Question 5.

Contrasting the Promise and the Threat

Where it is. فَأُولَئِكَ يَتُوبُ اللَّهُ عَلَيْهِمْ (fa-ulāʾika yatūbu Allāhu ʿalayhim), "it is those to whom Allah will turn in forgiveness" (4:17), and أُولَئِكَ أَعْتَدْنَا لَهُمْ عَذَابًا أَلِيمًا (ulāʾika aʿtadnā lahum ʿadhāban alīman), "those: for them We have prepared a painful punishment" (4:18).

The plain way. "Allah accepts the repentance of the first group and refuses the others" would state both outcomes, but in two unrelated sentences.

What it does. Each verse gives its outcome under the same word, أُولَئِكَ (ulāʾika), "those", followed by what is decided for them, so the reader meets the two outcomes as a pair. al-Zamakhsharī takes the match to show that both are bound to come.²⁹² For the promise, that certainty rests on Allah's promise and not on any obligation, as Deeper Study, Question 3 concludes. For the threat, it holds for the disbeliever, for whom the punishment is prepared; the end of the believer who persisted is left to Allah's will, as Question 4 concludes. The heavier the threat, the more the repentance of the first verse is seen to gain, and the threat's wording adds weight of its own. "For them" is placed before the punishment, which shows that this punishment is prepared for them in particular; the indefinite عَذَابًا (ʿadhāban), "a punishment", magnifies it in itself, and "painful" magnifies it in its description (Ibn Kamāl Pāshā, Abū al-Suʿūd).²⁹³

The effect. The promise and the threat function as a single, inseparable pair. Each sharpens the other: the reader fully grasps what is saved through the repentance of the first verse only by seeing what awaits in the second.

"Those", pointing far (إشارة البعيد, ishārat al-baʿīd)

Where it is. فَأُولَئِكَ (fa-ulāʾika), "it is those", in 4:17, and أُولَئِكَ (ulāʾika), "those", in 4:18: in both verses the demonstrative for what is far, though the people it points to have just been described.

The plain way. A demonstrative for what is near would suit people named a moment before.

What it does. Because the same far demonstrative gives both outcomes, the difference lies in what its distance marks. In 4:17 it may mark the height of the penitents' rank (Ibn ʿUthaymīn, al-Ālūsī).²⁹⁴ It may also mark only that their description is complete, so that once the verse has finished describing them they are treated as distant (Abū al-Suʿūd).²⁹⁵ The two explanations do not exclude each other, since a description can be complete and still be one of high rank, and the effect below rests on the first. In 4:18 the effect depends on whom "those" points to. If it points to both groups, its distance tells how far their state reaches in awfulness and how low their place is in evil (Abū al-Suʿūd).²⁹⁶ If it points only to the nearer group, the disbelievers, as the first reference in Deeper Study, Question 4 has it, the remoteness belongs to them alone.

The effect. Under the interpretation of rank, a single demonstrative achieves two opposing rhetorical effects: it elevates the penitents in the first verse and banishes the others in the second.

"Then" beside "soon"

Where it is. ثُمَّ يَتُوبُونَ مِنْ قَرِيبٍ (thumma yatūbūna min qarībin), "and then repent soon".

The plain way. "And they repent soon", with the simple "and", would join the repentance to the sin without specifying an interval between them.

What it does. ثُمَّ (thumma) is the particle that joins with an interval. One explanation takes the interval as time. Because Allah used the particle of delay here, the verse tells how wide His pardon is (al-Samīn al-Ḥalabī).²⁹⁷ So a repentance that came late is still accepted, because it came before the death throes and the sight of death (Darwīsh).²⁹⁸ The explanation from rank denies that the particle marks time at all, because "soon" follows it, and takes it as the distance in rank between the evil deed and the repentance (al-Hilāl).²⁹⁹ That objection assumes that "soon" leaves no room for delay. Deeper Study, Question 2 measures "soon" to the throes of death, so an interval after the sin does not contradict it, and the explanation from time stands (al-Samīn al-Ḥalabī, al-Ālūsī, Darwīsh).³⁰⁰ The answer removes only the denial of time; the distance in rank between the sin and the return is not touched by it.

The effect. The particle leaves room for time between the sin and the return, and the whole of that time lies within the promise.

The imperfect "do" (صيغة المضارع, ṣīghat al-muḍāriʿ)

Where it is. يَعْمَلُونَ السُّوءَ (yaʿmalūna al-sūʾa), "who do wrong", in 4:17, and يَعْمَلُونَ السَّيِّئَاتِ (yaʿmalūna al-sayyiʾāti), "who do evil deeds", in 4:18: the verb is in the imperfect in both verses.

The plain way. The past tense, ʿamilū al-sūʾa, "who did wrong", would present the sin as one finished event, and the verse does not use it.

What it does. In 4:17 the imperfect presents the doing as a scene before the reader, or as a doing that is renewed and repeated (al-Hilāl).³⁰¹ On the second sense the renewal is of sins that come one at a time, each broken off by a return, which is the first verse's case as Gem 2 sets it against the sins kept up in the second. In 4:18 it has been taken to signal that they persist in doing evil deeds until death comes to them (Abū Ḥayyān).³⁰² But the same form stands in both verses, so the tense alone cannot separate the two groups. The persistence of the second group comes from "until", whose limit is set out in Iʿrāb; the tense adds to it only a doing that goes on. What separates the groups is the clause after the verb: in the first verse "then repent soon", in the second "until, when death comes to one of them".

The effect. The imperfect keeps both groups in the act rather than behind it. The question the two verses put is not whether a person has sinned, but whether the sinning is broken off by a return before death.

"When death comes to one of them" (تقديم المفعول, taqdīm al-mafʿūl)

Where it is. حَتَّى إِذَا حَضَرَ أَحَدَهُمُ الْمَوْتُ (ḥattā idhā ḥaḍara aḥadahumu al-mawtu), "until, when death comes to one of them". The object, أَحَدَهُمُ (aḥadahum), "one of them", stands before the subject, الْمَوْتُ (al-mawt), "death".

The plain way. Arabic would ordinarily name the doer before the object, "death came to one of them", and the same death could have come to "them".

What it does. The verb is ḥaḍara, "became present". al-Hilāl contrasts it with jāʾa, "came": jāʾa has the sense of suddenness, severity and the taking of the soul, while ḥaḍara has the sense of a slower approach, the stage of dying before the soul is taken.³⁰³ The verse bears this out, since the dying person still speaks and says "I have repented now", when it is already too late. The object stands first, so the attention falls on the person death reaches before it falls on death, and every hearer can take it as falling on him; he is left watching for it and fearing what follows (al-Hilāl, al-Biqāʿī).³⁰⁴ "One of them" in place of "them" does the same work: the plural would bring to mind the rare death of a whole group at once, which does not suit a warning to those who fell short (Ibn Kamāl Pāshā).³⁰⁵

The effect. The fronted object and "one of them" put each hearer singly in the dying person's place, and ḥaḍara gives him time enough to speak and none left in which a repentance is accepted.

Naming those who die as disbelievers (المبالغة, al-mubālagha; تكرير حرف النفي, takrīr ḥarf al-nafy)

Where it is. وَلَا الَّذِينَ يَمُوتُونَ وَهُمْ كُفَّارٌ (wa-lā alladhīna yamūtūna wa-hum kuffārun), "or of those who die while they are disbelievers", joined to those who repent only when death comes.

The plain way. "Repentance is not accepted from those who delay it until death" would state the ruling on the delayer alone.

What it does. Those who die as disbelievers died without the repentance that faith is, and Deeper Study, Question 4 gives the reason the verse names them in a verse on repentance. A second reason stands beside it: the naming is for stress, so that a repentance made when death is present is counted as no repentance (al-Bayḍāwī, Ibn Kamāl Pāshā, Abū al-Suʿūd).³⁰⁶ The equation of the two groups is weighed in Question 4, and the one negation over both is shown in Iʿrāb.

The negating particle is also repeated, وَلَا (wa-lā), "and not": the one negation of laysa covers both groups, and lā repeats it for the second. Abū al-Suʿūd sees in the repetition a hidden hint that the delayer is, if anything, the clearer case of a repentance that brings no benefit, and he gives no ground for it beyond the repetition itself; al-Ālūsī reports the hint as something that has been said.³⁰⁷ The hint concerns what the late repentance achieves, which is nothing. It does not concern the delayer's end, which for the believer who persisted until death is left to Allah's will.

The effect. A repentance made at death is set beside those who died without faith, and a reader who means to repent later learns what a repentance delayed to that moment is worth.

The close, "Knowing and Wise"

Where it is. وَكَانَ اللَّهُ عَلِيمًا حَكِيمًا (wa-kāna Allāhu ʿalīman ḥakīman), "and Allah is ever Knowing and Wise", at the end of 4:17.

The plain way. A verse that only encouraged repentance might be expected to close on غفورًا رحيمًا (ghafūran raḥīman), "Forgiving, Merciful".

What it does. Ibn Kamāl Pāshā sees in these two names the answer to the word with which the verse opens, innamā, "only", whose grammar is covered in Iʿrāb. The repentance that Allah wrote upon Himself to accept is only the one the verse describes, and so no other is accepted; the close then gives the reason for that limit. Allah is Knowing of the repentance that meets the conditions of acceptance and of the one that does not, and Wise, so He accepts only the first. For, he reasons, what suits acceptance is encouragement and promise, and had the verse not limited acceptance and given the two names as its reason, it would have closed on "Forgiving, Merciful". This close, as he hears it, is a threat and a warning to whoever falls outside the limit.³⁰⁸

The effect. On his understanding, the promise remains a promise, and whoever repents as the verse describes is within it; yet the verse does not end on reassurance alone. Its last words already face the second verse, and the reader crosses into it knowing that there is a repentance which Allah knows and, in His wisdom, does not accept.

6. Qirāʾāt: Variant Readings

Ibn al-Jazarī's al-Nashr records the ten readings. His entries for the words of Sūrat al-Nisāʾ pass from "the two among you" in 4:16 to "by compulsion" in 4:19, and at these two verses he names one word, الْآنَ (al-āna), "now", which he sends to his chapter on moving the hamza's vowel to the letter before it.³⁰⁹ The difference there lies in how the word is pronounced, not in which word is recited. It is set out here because he names the word at this sūra, and because at this word Ibn Wardān, who does not move the vowel as a rule, has two ways. Other general rules of pronunciation, which recur wherever their conditions recur, are not set out word by word in this section.

الْآنَ: the hamza's vowel moved to the lām

Some Arabs lighten a single hamza by moving its vowel to the letter before it, and of the ten readings only Warsh's riwāya makes this a general rule. The preceding consonant must be unvowelled, must end a word, must not be a letter of prolongation, and must stand before a hamza that opens the next word; the lām of the article is one such letter. It takes the hamza's vowel, and the hamza drops from pronunciation.³¹² In "I have repented now" the lām of al- stands before the hamza of āna, so Warsh, joining the two words, recites tubtu-lāna.

Ibn Wardān, who transmits from Abū Jaʿfar, has the transfer at this word as well, and his transmitters give it in two ways. Ibn al-Jazarī records them:

His "rest of the Qurʾān" is every place of the word other than its two places in Sūrat Yūnus, which he has just treated, and his entry for Sūrat al-Nisāʾ sends this verse's al-āna to that chapter.³¹⁴ Ibn Wardān is therefore listed in both rows of the table.

What changes in meaning? Moving the vowel changes how the word sounds and no letter of the written word. In either way the dying man says "I have repented now", the word that, as Ṣarf showed, places his repentance in the moment he speaks, and the verse refuses it in both. The written وَلَا (wa-lā) of "nor those who die", with the report that some copies wrote it with two lāms, is a question of the written text, taken up in Iʿrāb under "Nor those who die"; Ibn al-Jazarī's entries record no difference among the ten at it.³¹⁵

And Allah knows best.