66. Sūrat al-Taḥrīm

Twelve verses. It is one of the Medinan sūras. It takes its name from the verb تُحَرِّم (tuḥarrimu — "you are making forbidden") in the first verse; it is transmitted that in some sources it is also known as Lima tuḥarrimu.

The sūra is one of the hardest texts in the Qurʾān to place — and the difficulty comes not from its language but from whom it addresses. It opens with a correction directed at a prophet, begins to recount the incident out of which that correction came, then leaves the incident half told, then turns to all the believers, and at the very end closes by recounting four women out of history.

It is easy to take this strange structure for disorder. It is not. What the sūra does, as I shall try to show below, is to conceal an incident and keep the principle that comes out of it. And it describes this choice word for word in the third verse.

The sūra proceeds in four steps.

First step (1-5): a domestic matter and the answer given to it. The Prophet has forbidden to himself something God had made lawful to him. The verse corrects this. Then the way an oath is loosed is stated, then a secret and the spreading of that secret are recounted, then repentance is offered to two persons.

What must be noticed is this: through five verses no person is named, no statement is made of what was made forbidden, and no explanation is given of the content of the secret. The text leaves out every defining detail of the incident.

Second step (6-7): the address widens. Now the addressee is "O you who believe", and the command passes from a single house to all houses: guard yourselves and your families.

Third step (8): the command to repent. What was offered to two persons in the fourth verse is here commanded to everyone. The same verb, the same root, an enlarged addressee.

Fourth step (9-12): four women. Two women in the houses of two prophets perish; in the most tyrannical house a woman is saved; and the sūra closes with a woman who never married — with Mary (Maryam).

That is, the sūra begins inside a single house and opens out onto four houses in history. And the two commands standing between the two ends — guard and repent — lie exactly in the middle, in the sixth and eighth verses. This is not disorder; it is a passage from a narrowing subject to a widening principle.

It is widely transmitted in the sources that the first verses of the sūra came down upon a particular incident, and that this incident concerns the Prophet's house. The reports agree at this point. What comes after that they do not.

The transmissions come down through two main branches that differ considerably from one another: in one branch it is said that the thing made forbidden was a food, in the other that it was a different matter. There are also differences among the variants about which of the wives the incident took place between; in some reports names are given, in others they are not, and the names given are not the same in every report.

I do not go into the detail here and I give no name. My grounds are three:

One. As method requires. The rule to which this text is bound is clear: a transmitted report of which one is not sure is not relayed. To pick one from among mutually contradictory reports and present it as "this is the incident" would be to present a preference as though it were a transmitted report — and that is the error this text counts most grave.

Two. The detail is of no use to the verse. The ruling of the verse does not depend on which detail the incident turned upon. The sentence "do not make forbidden to yourself what God has made lawful" is understood exactly without knowing whether the thing made forbidden was honey or something else. The particularity of the occasion does not narrow the generality of the ruling.

Three, and the real ground: the text itself does not do this. For five verses the Qurʾān chooses not to state all the details it has in hand. It does not give the names, it does not give the subject, it does not give the secret. We are not at a place where the exegete must open up what the text conceals — we are at a place where the text does the concealing deliberately.

And this is no idle inference. The third verse of the sūra describes this very behaviour in a single sentence:

عَرَّفَ بَعْضَهُ وَأَعْرَضَ عَن بَعْضٍ "He made a part of it known and turned away from a part."

In the verse this is recounted as a behaviour of the Prophet: he spoke a part of what he knew and did not enter upon a part. And the sūra does exactly the same thing to its reader — it makes a part known and turns away from a part.

I record this as my own reading, not as a classical transmission. But the coincidence between the text's behaviour and the behaviour the text describes is a fact worth pausing over. And its practical consequence is this: to plunge into detail in this sūra runs against the sūra's own method.

For that reason I write below without the least approach to the language of gossip. What is at issue is speaking about people who have died, by means of a text that is not in their hands; and this requires care.

The address-structure of the sūra is by itself material for a commentary. In twelve verses the addressee changes six times:

That Arabic possesses a dual form is of use here: the fourth verse states by grammar that it is said to exactly two persons — not to one, not to three. English has no such distinction and it is lost in translation.

And what the map says is this: the sūra narrows and widens. The narrowest point is the fourth verse (two persons), the widest the sixth (all believers and all families). In passing from the most particular matter to the most general command the text builds no bridging sentence; it merely changes the form of address.

Yā ayyuha'n-nabiyyu lima tuḥarrimu mā aḥalla'llāhu laka tabtaghī marḍāta azwājik, wa'llāhu ghafūrun raḥīm

"O Prophet! Why do you make forbidden to yourself what God has made lawful to you, seeking the good pleasure of your wives? God is much forgiving, most merciful."

The question of the places where the Qurʾān criticises its own prophet was taken up in ʿAbasa; the findings there — the fact itself, the list of examples (al-Tawba 9/43, al-Anfāl 8/67-68, al-Aḥzāb 33/37 and this verse), and the note that this must be presented as an observation and not as a proof — are not repeated here. That section had already taken this verse into its own list.

What must be done here is not repetition but distinction. For the criticism in ʿAbasa and the criticism here are not of the same kind.

The difference gathers here: what is criticised in ʿAbasa is a prioritisation supposed to be correct. What is criticised here is a behaviour that looks more pious. To deprive oneself of something looks, from outside, not like a deficiency but like a surplus. The verse objects to this.

And the direction of the objection must be read carefully: the verse does not say "you acted harshly", nor does it say "you wronged yourself". It says this: you narrowed a place God had widened.

This is no ordinary warning. Drawing the boundary between lawful and forbidden belongs, by a principle repeated many times over in the Qurʾān, to God alone. The Qurʾān criticises openly, in other places, those who set this boundary on their own: "Do not say, on the strength of the lie your tongues invent, 'this is lawful and this is forbidden'" (al-Naḥl 16/116). That the same Qurʾān corrects the lightest and most private form of the same act in its own messenger too shows that the principle is not bent for anyone.

The finding recorded in al-Baqara at 2/124 — that the sentence "My covenant does not reach the wrongdoers", having been said to a prophet who was asking, shows that the rule is bent for no one — meets a counterpart here: the rule that the authority to set boundaries belongs to God is applied first of all to the messenger himself.

The body of the verse is built between two verbs and the two come from opposite roots. Both must be opened out, because the English words "forbidden" and "lawful" have lost the picture the roots carry.

Root ح ر م (ḥ-r-m). At the centre of the root there is no prohibition. At its centre lie setting apart and making inviolable: taking a thing out of general use, drawing a boundary round it and closing it to everyone's access.

The derivatives of the root show this plainly:

• حَرَم (ḥaram) — the region around Mecca in which no hunting is done, no blood is shed, no tree is cut. Not a forbidden zone; an inviolable zone.
• إحرام (iḥrām) — the state entered upon in the pilgrimage and the ʿumra. A person takes himself out of ordinary life for a time.
• مَحْرَم (maḥram) — a close relative whom it is forever forbidden to marry. But the English borrowing harem (a part of the house set apart, closed to outsiders) stands closer to the root's original sense: set apart, not open to everyone.
• حُرْمَة (ḥurma) — reverence. English keeps the two apart with two unconnected words — "reverence" and "prohibition" — but the same root gives both.
• حَرِيم (ḥarīm) — the protected zone around a house, a well, a property; in law it is still used in this sense.
• مَحْرُوم (maḥrūm) — the one to whom something is not given, who is left without his share.

That the same root gives both "sacred" and "forbidden" is no coincidence. Both are the result of the same act: drawing a line round something. What is inside the line may be inviolable because it is sacred, or because it is dangerous. The act is the same, the reason different.

This is needed in order to understand the verse: what the Prophet did was to declare inviolable around himself something lawful to him. That is, not an ordinary "I shall not do it" but a change of status.

The pattern: تَفْعِيل (tafʿīl). Not ḥaruma but ḥarrama. The tafʿīl form states causation: not "it became forbidden" but "he made it forbidden". That is, the verb has an agent, and that agent is the one the verse addresses. The choice of word takes the act out of being a state and makes it a decision.

Root ح ل ل (ḥ-l-l). The opposite pole. The concrete sense of the root: the untying of a knot, the opening of a bond, the setting down of a load in order to make camp.

• حَلَّ العُقْدَة — he untied the knot.
• حَلَّ بِالمَكَان — he made camp at a place; the traveller comes to have "alighted" there because he unties the animal's load. From the same root come maḥall (place) and ḥulūl (settling in a place).
• حَلَال (ḥalāl) — untied, without a bond, carrying no restriction upon it.

That is, in Arabic ḥalāl means, before it means "permitted", "whose bond has been untied". And ḥarām means, before it means "prohibited", "round which a line has been drawn". The two words have been thought out by way of a knot and a fence.

This picture clarifies what the verse says: God untied a bond; the one addressed tied it again.

And it is worth recalling here an observation recorded in al-Baqara: the Qurʾān's choices of word most often build abstract notions by way of concrete physical operations. The pair lawful-forbidden is one of the clearest examples of this — the most abstract distinction in law is recounted with a rope and a boundary stone.

لِمَ = li (for) + mā (what), with its final alif dropped. A particle asking for a cause.

Questions of this kind directed at the Prophet are few in the Qurʾān and each is weighty. The form of the question is worth pausing over, because other forms could have been chosen:

• Prohibition: "Do not make forbidden to yourself what God has made lawful." — A direct prohibition.
• Statement: "You have made forbidden to yourself what God has made lawful." — A finding.
• Question: "Why do you make it forbidden?" — What the verse chose.

All three carry the same information; the three do not do the same work.

A prohibition leaves the one addressed passive: a command comes, it is obeyed. A statement is an accusation: this is what you did. A question, however, asks the one addressed for a ground. And to ask for a ground is to put the one addressed inside a reasoning — he is expected to notice for himself a ground he cannot give.

The finding recorded in al-Māʿūn about the rhetorical function of a question holds good here too: a question does not make one concede; it makes one find. And in al-Muzzammil another face of the same technique was recorded: a question that has no answer silences the one addressed.

Here there is a third function as well, peculiar to this verse: the question softens the correction but does not lighten it. There is no reduction whatever in the dose of criticism in the sentence; the one addressed is named openly ("O Prophet"), the act is openly attributed to him (tuḥarrimu), even its ground is stated (tabtaghī marḍāta azwājik). The only thing softened is the form of the sentence.

The finding about "the architecture of courtesy" recorded in ʿAbasa meets us here in another form: "courtesy is not the bending of what is true but the correcting of the way in which what is true is carried."

This detail is often skipped, but it is important.

When the Qurʾān addresses the Prophet it uses two formulas: يا أيها النبي (O Prophet) and يا أيها الرسول (O Messenger). And there is a difference of function between the two formulas — not a fixed rule, but an observable tendency:

• Yā ayyuha'r-rasūl generally comes in contexts having to do with the task of conveying: transmitting, announcing, making known.
• Yā ayyuha'n-nabiyy comes more in contexts directed at the person and his position: family, house, war, legal regulation.

In this sūra the address comes twice and both are in the nabiyy form (verses 1 and 9). The next sūra, al-Ṭalāq, opens with the same word. This is a datum that makes one think the distinction is not accidental; but I do not impose it as a rule, because there are also crossing usages of the two formulas in the Qurʾān.

And another point: even while criticising the one it addresses, the verse calls him by the name of his office. It does not say "O Muḥammad"; it says "O Prophet". That is, the criticism is not made by suspending the position. The position stays where it is, and the correction is made inside it.

This is the complement of the architecture recorded in ʿAbasa: there the criticism was said and left, here the criticism is said from inside the title.

Root ن ب أ (n-b-ʾ): nabaʾ — news that is carried, that has weight. The full analysis of the root and the distinction between nabaʾ and khabar were made in al-Nabaʾ; I do not repeat them. I record only this: the view that the word nabī derives from this root was given there as one of two derivations. And the third verse of this sūra will use the same root four times at once — there is a link between the word of address and the weave of the third verse.

The verse does not content itself with establishing the act; it states its cause as well. And the cause stated is not a cause to be blamed.

Root ب غ ي (b-gh-y). The concrete sense of the root is to want, to seek, to go after. In al-Bayyina the negative branch of the same root — baghy, that is, exceeding the limit, wanting more than one's own share — was taken up in connection with al-Baqara 2/213. The root has two directions and the difference between them lies in the legitimacy of what is sought: to seek is legitimate, to exceed is not.

The pattern: the افتعال (iftiʿāl) form, ibtighāʾ. This form most often states reflexivity and effort: a person's doing something for himself, with exertion. Kataba (he wrote) → iktataba (he had something written for himself). That is, ibtighāʾ is not a passive "wanting"; it is going after something, trying to obtain it.

And the verb is imperfect: tabtaghī. Continuity. Not a single reckoning.

مَرْضَاة (marḍāt) — root ر ض و (r-ḍ-w): to be well pleased, to be content. Marḍāt is the mīmī verbal noun of riḍā: good pleasure, the state of being well pleased.

And here there is an observation which I record as my own reading — I do not present it as transmitted:

This word is used in the Qurʾān predominantly for the good pleasure of God. Ibtighāʾa marḍāti'llāh — "seeking God's good pleasure" — is a formula repeated in the Qurʾān and it describes a behaviour that is praised. The raḍiya'llāhu ʿanhum wa raḍū ʿanh treated at the last verse of al-Bayyina is from the same root; there the word stands as the highest recompense set above the description of the garden.

Now look at this verse. The same lofty word comes, together with the same verb of seeking, but with its object changed: marḍāta azwājik.

That is, in order to recount what it censures the verse chooses a word that is not censured, and only changes its direction. The fine criticism the sentence carries lies here: what was done is not a bad work — it is a good work done to the wrong address.

And this takes the verse out of being a criticism of family. The verse does not say "do not please your wives"; it says do not change a boundary God has set in order to please your wives. The two are entirely different things, and if they are confused the verse comes to say something it never said.

Indeed the Qurʾān itself mentions contentment between spouses as a value; the whole regulation the next sūra — al-Ṭalāq — sets up rests on the parties not harming one another. That is, the matter is not the seeking of contentment but by what means contentment is sought.

Three points:

1. A circumstantial clause. Tabtaghī marḍāta azwājik stands grammatically as a circumstantial (ḥāl): it states in what state the act is done. That is, "you make forbidden" and "you seek" are not two separate works; the second is the state within which the first stands.

This has a consequence: the verse builds the two verbs not as cause and effect but one inside the other. The act of making forbidden takes place within the seeking.

2. The object left indefinite. Mā aḥalla'llāhu laka — "what God has made lawful to you". A relative pronoun is used, that is, the object is not named.

This is the first instance of the attitude the sūra keeps up from beginning to end. The verse could have said what was made forbidden; it does not. And its not saying has a consequence: the ruling becomes independent of the object. The sentence is now not about a particular food or a particular behaviour; it is about everything that is lawful.

The measure of the tradition of legal method works here of itself: al-ʿibratu bi-ʿumūmi'l-lafẓ lā bi-khuṣūṣi's-sabab — the ruling goes by the generality of the wording. This rule was already recorded in al-Māʿūn. The difference is this: in al-Māʿūn we were applying the rule; here the verse comes already applied. The wording has been left general from the start.

3. "Laka" — to you. The verse does not say "what God has made lawful"; it says "what God has made lawful to you". That is, the being-lawful is mentioned as attributed to the one addressed.

This detail has an effect: the lawful is positioned as the addressee's own share. He is closing to himself something that was given to him. And as with the root ʿafw treated at al-Baqara 2/109, the picture of maḥrūm here (from the same ح ر م root) appears of itself: a person is depriving himself of his own share.

The most striking place in the verse is perhaps its end.

It does not end as you would expect a sentence of rebuke to end. It does not end with a threat, it does not end with "do not do it again", it does not end with a sanction. It ends with forgiveness and mercy — and moreover not in a separate verse but within the same verse.

This is another example of the finding recorded in ʿAbasa: "the criticism ends and no apology is asked, no repentance commanded, no expiation set. It is said, and left." Here a further step is taken: it is said and closed in the same breath.

Note the structure of the sentence: both adjectives are in the faʿūl and faʿīl patterns, that is, they state intensity — much forgiving, greatly merciful. And both come as God's permanent attributes, not as a reaction peculiar to this incident. That is, the sentence does not say "this time He forgave"; it says "He is so in any case".

An observation, as my own reading: this close makes one read the question at the beginning of the verse again in retrospect. The question lima tuḥarrimu, together with these two names appended at its end, becomes not a calling to account but an offer of opening. The total sense of the sentence is this: you narrowed; but you had no need to narrow, because the one before you is not one who narrows but one who widens.

One. To lay down extra prohibitions is not piety.

This is the most direct consequence of the verse and it is generally skipped, because it clashes with an intuition running the other way. The common intuition is this: religious sensitivity is measured by the number of things not done. Whoever closes more things to himself is counted more careful.

The verse corrects this intuition. And in correcting it, it uses the strongest example: the person doing this is the Prophet, and the reason for what he does is not selfishness but concern to please someone else. That is, the example has been chosen in the best-intentioned form it could take — and it is corrected all the same.

The measure that comes out of this is: the authority to draw boundaries is not in the person himself. A person may, by his own preference, not do something; that is the subject of no verse at all. But when he carries what he does not do into the status of "forbidden to me", he has used an authority that does not belong to him.

The difference is fine but large in practice: the difference between "I do not choose this" and "this is forbidden to me". The first is a choice, the second a ruling.

Two. Bending the principle in order to protect the relationship.

The mechanism the verse records is thoroughly familiar: giving up what one knows to be right, or one's own due, in order not to upset someone close.

This behaviour is most often counted a virtue and most often really is one — self-sacrifice, flexibility, accommodation. The Qurʾān praises these openly elsewhere.

The limit the verse points to is here: self-sacrifice must be made out of one's own property. A person may give up his own share, his own time, his own comfort. But when the thing he gives up is a ruling, he is no longer giving out of his own property.

And the practical counterpart of this is one of the traps relationships fall into most often: that the concessions given in order to please the other side end, past a certain point, by abolishing the person's own measure. A person left without a measure comes after a while to be unable to give anything even to the person he was trying to please — because he has no fixed place left from which to give.

Three. The way criticism is carried.

The verse does the hardest thing that can be done in a single sentence: it says the criticism without softening it at all, without shaking the position of the one addressed at all, and it closes.

The four steps of the technique stand in the grammar:

1. The one addressed is called by his title (the position is preserved). 2. A question is asked, no verdict is given (the reasoning is left to the one addressed). 3. The ground is stated and accepted (the intention is not disparaged). 4. The sentence closes with forgiveness (the matter is closed).

Today criticism generally loses at least two of these four. Either the position is shaken ("you are always like this anyway"), or the intention is disparaged ("really because it suited him"), or the matter is not closed — what was said is stored up to be used again later.

What the verse shows is that all four at once are possible.

Qad faraḍa'llāhu lakum taḥillata aymānikum, wa'llāhu mawlākum wa huwa'l-ʿalīmu'l-ḥakīm

"God has determined for you the loosing of your oaths. God is your mawlā; He is the knowing, the wise."

The most important feature of the verse stands, before its words, in its pronouns.

In the first verse everything was singular: tuḥarrimu (you make forbidden), laka (to you), azwājika (your wives). The addressee was one person.

In the second verse everything is plural: لَكُمْ (to you), أَيْمَانِكُمْ (your oaths), مَوْلَاكُمْ (your mawlā).

There is no transitional sentence in between. One verse ends, the other begins, and the addressee has changed.

This is the first instance showing how the sūra works, and it gives a rule that holds for the whole text: a particular incident is used as the place where a general ruling is born; but the ruling does not stay inside the incident.

And when this transition is joined to the indefiniteness of the object we saw in the previous verse, the picture is complete: the verse does not say what was made forbidden, it does not count the persons, and in the second sentence it already generalises the addressee. The same work is done in three separate ways — the incident is erased, the ruling remains.

Root ف ر ض (f-r-ḍ). The concrete sense of the root is rather far from the word's present religious usage: to cut a notch into something, to open a nick, to score a mark.

The notch in which the bowstring sits is called furḍa. A mark cut into a board or a bone is named with the same word.

Two branches come out of this:

The first branch: to determine, to fix the exact amount. A notch is the mark of a measure; it shows how far a thing goes. In Arabic faraḍa lahū — "he assigned him a definite share". The Qurʾān uses this sense plainly: in the context of the dower the expression "half of the amount you have made obligatory (determined) for them" (al-Baqara 2/237) is in this sense — what is at issue there is not a command but a determined figure.

The second branch: to make binding, to turn into an obligation. What has been determined binds because it has been determined. Farīḍa, farḍ, iftirāḍ are from this branch.

And note: which branch is at work in this verse?

Reading the sentence as a command seems possible at first sight: "God has made it obligatory for you to loose your oaths." But this reading makes the sentence meaningless — no one is under obligation to loose every oath he takes.

The right reading is the first branch: God has determined for you the way in which your oaths may be loosed. That is, what is there is not a command but the fixing of a possibility. A door has been opened and its measure set.

This difference matters, because it changes the tone of the verse entirely. Read as a command the verse imposes an obligation; read as a determination the verse shows a way out. The second suits both the root's original sense and the context.

And here a link may be made with the next sūra: at al-Ṭalāq 65/2 it will be said that "whoever guards himself against God, He makes for him a way out (مَخْرَجًا)". The two sūras stand side by side and both recount a door opened in a tight place. I record this not as a claim about the arrangement of the muṣḥaf but as a possibility of reading.

One of the most technical and most beautiful words of the sūra.

Root ح ل ل — treated at the first verse: the untying of a knot, the opening of a bond. I do not repeat that analysis.

The pattern: تَفْعِلَة (tafʿila). This is one verbal-noun form of the tafʿīl pattern (like taḥliya / taḥilla). That is, the name of the act of untying.

And the picture the word carries is this: an oath is a knot; taḥilla is the untying of that knot.

This picture has spread over all the words to do with oaths in Arabic, and that is no coincidence:

• عَقْد (ʿaqd) — a knot. The Qurʾān uses this word for oaths: ʿaqqadtumu'l-aymān (al-Māʾida 5/89). English keeps the same picture with a different word-family: one "ties the knot", one is "bound" by an agreement.
• حَلّ (ḥall) — untying. The opening of the same knot. English does the same by another road: to solve a problem is, at bottom, to loosen it — solvere means to untie.
• رَبْط / وَثَاق — to bind, a tight bond.

That is, when Arabic counts a word as given, it takes it to have been physically tied. A word is not a sound leaving the mouth; it is a rope knotted at both ends.

The consequence of this is heavy and it stands in the language of the verse: if an oath is a knot, then going back on an oath is not "failing to keep one's word" but untying a knot. And a knot can be untied without being cut. This is what the verse offers: the way to open the knot without cutting the rope.

Root ي م ن (y-m-n). The concrete sense of the root is the right hand. From the same root come yamīn (the right side), maymana (the right wing), yumn (good fortune, blessing) and maymūn (fortunate).

Why is an oath named with the word for "right hand"? Because among the Arabs — and in many cultures — right hands were clasped together when an agreement was made. The word was placed into the hand. In al-Ḥāqqa, where the notions of right and left were taken up, this side of the root was already mentioned.

That the word comes from this root shows that an oath was thought of as a bodily act: a word is not spoken, a hand is given.

And when this is joined with taḥilla the picture is complete: the hand is clasped, the knot is tied; then the knot is untied. The whole sentence is recounted by the movement of two hands.

A note: that the word yamīn occurs in the Qurʾān both in the sense "right hand / right side" and in the sense "oath" shows that this doubleness of the root is alive in the text. In the eighth verse of this sūra bi-aymānihim will occur as well — there in the sense "their right sides". The same root, four verses apart, with two different senses. I record this not as a claim of wordplay but as an indication of the root's liveliness.

Here I must draw a boundary.

The verse states that oaths can be loosed and that there is a way of doing so. This verse does not say what that way is. The content of the expiation for an oath — who is to be fed, how many, the alternatives, and what one who lacks the means is to do — is regulated elsewhere in the Qurʾān, at al-Māʾida 5/89. This verse points there; I do not repeat that ruling here.

And I issue no legal ruling. As the method of this text requires:

• In what cases an oath is binding,
• Whether making something forbidden to oneself counts legally as an oath,
• The details and order of the expiation,

all of these are the subject of the literature of law and there are differences among the schools. To decide these matters is not the business of this text. The reader should consult the sources of law for these questions.

The only thing I say here is what the verse itself says: a way out has been determined.

Root و ل ي (w-l-y). The concrete sense of the root is nearness, adjacency, standing side by side with no gap between. Walī, walāya, awlā, tawallī, mawlā, walī'l-ʿahd are all from this root.

The most concrete use of the word is the ranging of two things with nothing else between them: what comes immediately after a thing is its walī.

The word مَوْلَى (mawlā) covers a wide field in Arabic and is interestingly a reciprocal word:

• Master (one who frees)
• A freed slave
• Helper, protector
• Friend, one who is close
• Owner, one who takes charge of another's affair

That is, the same word is used both for the upper and for the lower party. The reason for this is the sense of the root: the word states not a hierarchy but a bond. Both ends of the bond are named by the name of the bond.

Its function in the verse. Note where the sentence stands. It comes immediately after the statement that an oath can be loosed. That is, the context is this: a person has given a word, the possibility of going back on that word has been stated, and immediately afterwards it is said "God is your mawlā".

There are two readings of this and both are defensible:

The two do not exclude one another. The common conclusion is this: the authority to untie comes from the other end of the bond.

Two names, and their choice suits the context.

الْعَلِيم — root ع ل م: to know. The faʿīl pattern, a knowing that is constant and complete.

الْحَكِيم — root ح ك م. The concrete sense of the root comes from a place often skipped: the bit put in an animal's mouth (ḥakama). A bit does not stop the animal; it steers it. It keeps it from going on its own and puts it into a direction.

The derivatives that come out of this: ḥukm (the decision that puts a matter in its place), ḥikma (the knowledge of putting a thing in its right place), iḥkām (making firm), ḥakam (arbiter).

That is, ḥakīm is not "one who knows much" — it is one who puts knowledge in its right place. The difference between knowledge and wisdom lies exactly here, and by setting the two side by side the verse makes the difference visible.

Its fitness to the context. The verse stated a way of resolution. That an oath can be loosed may look at first sight like a slackness: if the word given does not bind, what is the value of the word?

The two names give the answer to this objection:

• ʿAlīm — the one who knows which oath was given under which conditions. Who knows the states behind the oaths, the tight places, the things said in haste.
• Ḥakīm — the one who opens the door not arbitrarily but in its proper place. To return to the image of the bit: the possibility of untying is not the rein's being let go entirely; it is the correcting of direction.

That is, the verse says this: this door is not a slackness but a measured flexibility.

One. That a word given can be taken back.

The verse says something heavier than it appears: every bond by which a person binds himself does not have to be permanent.

This has two sides and both must be seen at once.

The first side is open to abuse: every arrangement that makes going back on one's word easier can lower the value of the word. The Qurʾān is aware of this risk and criticises openly elsewhere the taking of oaths lightly; moreover, that an expiation is provided shows that untying is not without cost. The door is open but it is not free.

The second side, and the principal one: a person may have bound himself to a wrong thing. There are words given in a moment's anger, out of an excessive scrupulousness, out of a feeling of guilt. To carry such a word for a lifetime is not keeping one's word; it is keeping up a mistake.

What the verse does is to open a road between the two: to preserve the weight of the word but not to lock a person inside his own mistake.

Two. A door opened for excess.

The context of the verse — the extra prohibition in the previous verse — makes this verse face in a particular direction. The oath whose loosing is made easier is not an oath to do evil; it is an oath of deprivation.

That is, the Qurʾān is here giving a way out to a person who is loading extra weight upon himself. This is not a sentence standing alone; at al-Muzzammil 73/20 there was a verse doing exactly the same work, and the finding recorded there was this: the lightening does not come upon a fault, it comes after something done in excess.

The two verses show the same mechanism:

The common conclusion: in correcting a person who loads too much upon himself, the Qurʾān does not use punishment; it recalls the measure.

Three. That a word is bodily.

That the word yamīn means "right hand" points to something we have lost today: giving one's word had a body. A hand is clasped, eyes meet, a witness is found.

Today most undertakings are bodiless: a box is ticked, a button is pressed. And the psychological weight of a bodiless undertaking is different from that of a bodily one — this is a fact observed in the behavioural sciences too, and it explains why rituals such as the signature, the handshake and the witness persist.

I write this not as a "scientific miracle" but as an ancient observation recorded by the word: a word is heavier when it is given by hand.

Wa idh asarra'n-nabiyyu ilā baʿḍi azwājihī ḥadīthā, fa-lammā nabbaʾat bihī wa aẓharahu'llāhu ʿalayhi ʿarrafa baʿḍahū wa aʿraḍa ʿan baʿḍ, fa-lammā nabbaʾahā bihī qālat man anbaʾaka hādhā, qāla nabbaʾaniya'l-ʿalīmu'l-khabīr

"And when the Prophet confided a word in secret to one of his wives. When she gave news of it and God disclosed it to him, he made a part of it known and turned away from a part. When he gave her news of it she said, 'Who told you this?' He said, 'The Knowing, the All-Aware told me.'"

First let us see how little the verse says. For this is its most conspicuous feature.

• To whom the word was said is not stated: baʿḍi azwājih — "to one of his wives". Indefinite.
• What the word was is not stated: ḥadīthan — "a word". Indefinite, without the article.
• Which part was made known and which part was left is not stated: baʿḍahū / ʿan baʿḍ — "a part of it" / "from a part". Both indefinite.
• In the exchange, who is speaking is known only from the conjugation of the verb; no name occurs.

This much indefiniteness in a single verse cannot be accidental. There is indefiniteness in four separate places, and each of them gives the minimum information needed to build the sentence and cuts off the rest.

And exactly in the middle, the verse's own sentence names this attitude: عَرَّفَ بَعْضَهُ وَأَعْرَضَ عَن بَعْضٍ.

I repeat here the observation I recorded in the introductory section, because this is its place: the verse applies the behaviour it describes to the reader as well. It makes a part known and turns away from a part. I present this as my own reading; it is not a classical transmission. But this coincidence between the text's behaviour and its content is the most useful key to hold in hand while reading the verse.

Its practical consequence is this: to look at this verse and try to reconstruct the incident is to spoil the work the verse does. The verse is not recounting the incident; it is recounting an attitude towards the incident.

Root س ر ر (s-r-r). This root divides in Arabic into two branches, and at first sight the two do not resemble one another at all.

The first branch: secrecy.

• سِرّ (sirr) — a secret, what is hidden.
• أَسَرَّ (asarra) — he hid; and he said in secret. Here is the interesting side of the word: the same verb means both "he concealed" and "he whispered". Some of the lexicographers count this under the heading of aḍdād (words carrying opposite meanings together). But in truth there is no contradiction: to whisper a thing to one person is to hide it from everyone else. The verb answers to both faces of the same act at once.
• سَرِيرَة (sarīra) — what a person keeps within.

The second branch: joy.

• سُرُور (surūr) — joy, ease within.
• مَسْرُور (masrūr) — joyful. In al-Inshiqāq this word occurred in recounting the return to his people of the one who takes his record from the right.

How do the two branches come from one root? The link the lexicographers give is this: joy is something within a person; what is hidden is also what is within. Both have to do with the "inside". Moreover one branch of the word sirr carries the sense "the best, the core of a thing" — sirru'l-wādī, the most fertile place in the valley. That is, what is hidden is hidden because it is valuable.

I relay this link not as a certainty but as an explanation the lexicographers record.

The sense in the verse. Asarra ilā — "he said in secret to …". That the preposition is إِلَى is important: the act is done towards a direction. That is, the word is not hidden; it is entrusted to an address.

And this establishes the whole matter of the verse: a secret word is said not in order to be hidden but in order to be shared while staying within a limited circle. This is the definition of a secret — information shared not with zero persons but with a counted number.

حَدِيثًا — root ح د ث (ḥ-d-th): to be new, to come about afterwards, to appear. Ḥāditha (an event), ḥadīth (a word; and what is new), iḥdāth (bringing about), muḥdath (what came about later) are from the same root.

That the word means "speech" comes from this root: a word is something that comes into being at that moment. Speaking is an event brought into being out of nothing.

And its coming indefinite (ḥadīthan, "a word") is here not indefiniteness but a leaving indefinite. The verse does not define the word.

A note: this root will meet us again in the first verse of the next sūra, al-Ṭalāq: laʿalla'llāha yuḥdithu baʿda dhālika amrā — "perhaps God will bring about some matter after that". The same root, in two neighbouring sūras, with two different subjects: here a human being brings about a word, there God brings about a state of affairs.

Root ظ ه ر (ẓ-h-r). The concrete sense of the root is the back: ẓahr, the rear face of the body. And all the derivatives that come out of it carry this picture:

• ظَاهِر (ẓāhir) — the outer face, the visible side. The opposite of bāṭin (the inner).
• أَظْهَرَ (aẓhara) — he brought out, he made manifest. That is, he brought what was inside to the outer face.
• ظَهِير (ẓahīr) — one who backs, who supports. One who stands behind another, who gives him his back.
• تَظَاهَرَ (taẓāhara) — to back one another. (It will come in the fourth verse.)
• ظُهْر (ẓuhr) — noon; the "back" of the day, its highest point.
• مُظَاهَرَة — a public showing; English covers the same ground with demonstration, from another root but with the same idea of bringing something out into the open.

The usage in the verse: aẓhara … ʿalayhi — "He brought it out upon him". That is, what was hidden has been brought to the person's outer face; it is now among the things he knows.

And the point to notice: the agent of the verb is God. The Prophet does not learn this by investigating, by asking, by following a trail. The knowledge was given to him.

This is needed in order to understand the behaviour that follows in the verse: he has in hand a piece of knowledge he did not obtain by investigation. And what he will do is how he will use that knowledge.

And note the root's journey through the sūra. The root ظ ه ر occurs three times in this sūra, in three different forms:

The same root, in one verse "the coming out of knowledge", in the next "giving one's back to one another". What joins the two is the concrete sense of the root: the back. Knowledge passes from the other side of the back to the front; support comes about by giving back to back.

I record this link as an observation; I do not present it as a link established in this way in the classical commentaries.

This is the most striking verbal feature of the verse and it is generally lost entirely in translation.

The root ن ب أ occurs four times in this single verse:

1. نَبَّأَتْ بِهِ (nabbaʾat bihī) — "she gave news of it" (the woman) 2. نَبَّأَهَا بِهِ (nabbaʾahā bihī) — "he gave her news of it" (the Prophet) 3. أَنبَأَكَ (anbaʾaka) — "who told you?" (the question) 4. نَبَّأَنِيَ (nabbaʾaniya) — "He told me" (the answer)

And one of the last two words of the verse is الْخَبِير — that is, from the root of khabar. So the verse uses both of Arabic's words for "news": four times the root nabaʾ, once the root khabar.

In al-Nabaʾ the difference between these two words was taken up and I do not repeat it. I recall the finding there in a single sentence: khabar — information; nabaʾ — a tiding to be delivered. One is relayed, the other conveyed; and what determines a nabaʾ is the consequence of the news — a nabaʾ is news upon hearing which something must change.

What the verse does is this: it follows the passage of a piece of news through four hands, with the same word.

And the subject of each of the four is different:

The chain has four links and at the very end the source is reached. The verse follows a circulation of speech and comes out at last at a point outside that circulation.

And note the form of the last sentence: "The Knowing, the All-Aware told me." The answer does not say "God"; it gives two names. And both names have to do with knowledge:

• الْعَلِيم — the knowing. Root ع ل م.
• الْخَبِير — the one who is aware; who has news of the inner side of a thing. Root خ ب ر.

The concrete sense of the root خ ب ر is striking: khabara'l-arḍa — he ploughed the earth deeply, he turned it over. A khabīr is one who knows a thing not from its surface but from within. From the same root come khubr (inward knowledge of a thing) and ikhtibār (learning by trial).

That is, the answer says this: this knowledge did not come from a piece of gossip; it came from the party who knows what is under the soil.

This is the most fitting answer to the question ("who told you this?"). For the question was investigating the source — from whom did the information leak? The answer does not follow the chain of leakage; it steps outside the chain.

This is the finest sentence of the sūra and it is worth pausing over.

The situation is this: a person has told someone else a word entrusted to her. The owner of the word has learned of it. And the other party does not know that he has learned it.

This is one of the most asymmetrical positions that can arise in human relations. You have in hand a piece of information the person facing you does not know you have. And that information is against her.

Three behaviours are possible in this position:

The verse records the third.

And one must see what the third does. To say none of it looks at first sight like the most delicate option, but it is not: information not said continues to stand within the relationship as an asymmetry and can be used at any moment. To say all of it, on the other hand, looks at first sight like the most honest option, but it does more than is needed: it leaves the other party defenceless.

To say a part of it does this: it removes the asymmetry but does not expose. The other party now knows that it is known — so the relationship is levelled. But what is known and how much is not opened — so no one is driven into a corner.

أَعْرَضَ — root ع ر ض. This root was treated at al-Jinn 72/17: the concrete sense of the root is the breadth, the width, the side of a thing; aʿraḍa ʿan — to turn one's side to someone. There the word recounted a negative attitude: turning away from the remembrance of one's Lord.

Here the same verb recounts a positive attitude. And this shows that the word is neutral: turning away takes its value from what is turned away from. To turn away from remembrance is a loss; to turn away from the detail of a fault is a virtue.

Setting the two verses side by side makes this neutrality of the root visible:

عَرَّفَ — root ع ر ف (ʿ-r-f): to know, to recognise. But it differs from ʿilm: maʿrifa is recognition — knowing again what has been met with before. From the same root comes maʿrūf (known, recognised; and from there "good" — because what is good is what everyone recognises).

ʿArrafa is of the tafʿīl form: he made known, he acquainted. That is, the Prophet does not give the information; he makes known that he knows.

The difference is fine but real: to say "you did such and such" is to pour out the information; to say "I know this" is only to state the position. The verb in the verse is closer to the second.

1. Beginning with وَإِذْ. The verse begins with idh — an adverb pointing to a past time. In the Qurʾān this particle is generally used to recall a scene, and an implied verb "remember" is most often assumed before it (udhkur idh).

Its effect is this: the verse does not recount an incident, it points to an incident. To recount is to set the incident before the reader; to point is to turn back to something already known. And for the first addressees the incident was known.

This partly explains why the detail is not given: for the first addressees the detail was already unnecessary. But the Qurʾān, in putting on record a known incident, chose not to record the detail — and what we have today is the result of that choice.

2. The changes of agent made without names. There are four different subjects of verbs in the verse (the Prophet, one of the wives, God, and again the Prophet) and none of them is mentioned by a name. The subjects are followed only by the conjugation of the verb.

The richness of Arabic conjugation makes this possible: between nabbaʾat (she gave news) and nabbaʾa (he gave news) there is a difference of one letter, and this suffices to state without a name whom the sentence is speaking of. English has no such resource; in translation we are forced to write the subjects out explicitly and the cover the verse provides is partly removed.

3. The question relayed directly. Qālat man anbaʾaka hādhā — the question is given not in indirect speech but as a direct quotation. This is common in the Qurʾān's technique of narration and its effect is to bring the scene to life.

But here there is an interesting balance: the scene is brought to life while the identities are concealed. That is, the verse gives the texture of the incident and does not give its identity.

One. The definition of a secret and the speed of its dispersal.

The verse shows in three words how a secret disperses: asarra (he said in secret) → nabbaʾat bihī (she gave news of it). There is nothing in between.

What this says to the present is direct: a secret, from the moment it is shared, is no longer a thing that can be controlled. In giving his word a person thinks he is giving it to one person; but what he gives is a piece of information, and information has a logic of spreading of its own.

This does not amount to blaming the other party — nor does the verse blame; it merely records. What is said is plainer: the sharing of a secret is an operation with a risk, and the one who bears the risk is the one who shares.

Two. The ethics of the party who holds the information.

This is the most transferable part of the verse.

Today it is far easier to know something about someone that he does not know. Messages are forwarded, screenshots circulate, mutual acquaintances relay. And the person who has this information is in exactly the position the verse describes: he has in hand something the other party does not know he has.

The measure the verse shows is this: make known that it is known, do not open the detail.

It is not hard to see why this works. Opening the detail does two things: it puts the other party into a defensive position (and a person on the defensive does not think about the matter, he saves himself), and it opens up how you obtained the information for discussion as well (so that the subject changes). Merely making known that it is known, by contrast, leaves the other party the possibility of making his own reckoning himself.

This verse is of the same family as the architecture of criticism recorded in ʿAbasa: there too the criticism was said and left, and it was not pressed.

Three. When information is held as a power.

One of the two options the verse leaves out was "to say none of it". This deserves separate attention, because it is often taken for delicacy.

To know a fault of the other party and never to say it makes the relationship permanently unequal. Information is a power even while it goes unused: the uncertainty of when it will be used enters the texture of the relationship. And most often it is used one day.

The road the verse chooses restores equality by spending the information: the person who says what he knows has also ended that information's character as a power.

Four. Asking after the source.

The question in the verse — "who told you this?" — is a thoroughly familiar reaction. The first reflex of a person whose fault has been learned is most often not to discuss the fault but to find the source of the leak.

One must see what this reflex does: it shifts the subject from what was done to how it was heard of. And this shift is most often made unawares.

The answer in the verse does not accept this shift. Instead of following the chain of the source it steps outside the chain. The measure lying beneath this is: how a thing was heard of does not change the fact that it was done.

In tatūbā ila'llāhi fa-qad ṣaghat qulūbukumā, wa in taẓāharā ʿalayhi fa-inna'llāha huwa mawlāhu wa Jibrīlu wa ṣāliḥu'l-muʾminīn, wa'l-malāʾikatu baʿda dhālika ẓahīr

"If you two repent to God — your hearts have inclined. But if you back one another against him, then God is his mawlā; and Jibrīl, and the righteous among the believers. And after that the angels too are a backer to him."

The first word of the verse gives a piece of grammatical information, and that information is the key to reading the verse.

تَتُوبَا — the dual form. In Arabic the singular, the dual and the plural are separate conjugations; the dual is used only for two persons. The same form continues throughout the verse: qulūbukumā (the hearts of you two), taẓāharā (if you two back one another).

That is, the verse is said to exactly two persons.

This is information lost in translation. English "you" serves for two people and for twenty alike; Arabic makes the distinction. That the verse's addressee is two is written in the verse itself.

And this is the narrowest point in the sūra's map of addressees. In the previous verse one person was spoken of (baʿḍi azwājih); here two persons are addressed; two verses later it will pass to all the believers. The text first narrows as far as it can, then widens as far as it can.

Names are given in the reports as to who is meant. For the grounds I recorded in the introductory section I write no name here: there are differences among the reports, and the verse itself gives no name.

This is the most disputed grammatical point in the verse.

The sentence is built thus: "If you two repent — your hearts have inclined."

This is an incomplete sentence in Arabic as it is in English. The condition "if you repent" must have an answer: if you repent, what happens? The verse does not say. What it says stands rather like the ground for repenting.

The explanations given for this gap in the classical sources on grammar:

All three are defensible and in practice they come out at the same place. I do not impose a preference.

But there is something that follows from the form of the sentence and it must be recorded: the verse does not command repentance. It does not say tūbū (repent); it says in tatūbā (if you two repent). The conditional form does not compel the one addressed — it shows the door.

And this becomes meaningful when compared with the eighth verse. There the same root will come in the imperative: tūbū ila'llāhi tawbatan naṣūḥā. That is:

Where it is particular, an offer; where it is general, a command. I record this ordering as an observation: the text does not give a command to two persons whose identity is known, it shows the possibility; when it says the same thing to everyone it uses the imperative.

Root ت و ب (t-w-b). This root was taken up at al-Baqara 2/37 and the findings there are not repeated. I recall it in one sentence: the sense of the root is to turn, and the same root is used both of the servant and of God — the servant tāba ila'llāh (turns towards God), God tāba ʿalayh (turns towards the servant). The main point recorded there was this: God's turning comes before the servant's turning; the servant can turn because he has already been turned towards.

What is new here: the preposition. In the verse tatūbā ila'llāh — "towards God". That is, the servant's direction is what is recounted, not God's turning. In the eighth verse the same preposition will come: tūbū ila'llāh.

This consistency is meaningful: in recounting the side that falls to the servant, the sūra always uses ilā. The servant's work is to change direction.

The most pictorial expression in the verse.

Root ص غ و (ṣ-gh-w). The concrete sense of the root: to bend, to lean to one side, to slip.

The liveliest use of the root has to do with the ear: aṣghā ilayhi — he gave ear to him. And this describes exactly a bodily movement: bending the head to that side in order to hear a sound better. From the same root comes iṣghāʾ (giving ear). The Qurʾān uses it in this sense too: "so that the hearts may incline towards it" (al-Anʿām 6/113) — there too the same root recounts the heart's bending to one side.

The other concrete usages strengthen the same picture: a vessel tilting to one side, the sun leaning to the west, a wall slipping to one side.

The function of the word in the verse. The verse does not say "your hearts have gone bad", it does not say "your hearts have darkened", it does not say "your hearts have become unbelieving". It says "they have inclined".

This is a thoroughly measured word and its choice is meaningful. To bend is not to break. A bent thing can be straightened; a broken thing cannot. In making the diagnosis the word leaves the door open.

And one further nicety: bending is a movement that is hard to notice. If a vessel overturns you see it; if it tilts slightly you do not. The word says that a heart can change direction by a small slippage — and that this slippage becomes visible only through an intervention from outside. This is what the verse does: it states the inclination.

An observation, as my own reading: the word's link with the movement of the ear falls into the context of the sūra. The third verse was about the hearing and the relaying of a word. The fourth verse speaks of the heart's inclining — and the same root is the root that recounts bending the head to hear a sound. I do not present this as a claim of wordplay; I only record that the picture the root carries falls into the context.

There is a mismatch in the verse that goes unnoticed: the subject is dual, but qulūb (hearts) is plural.

The expected form would be qalbākumā (the heart of you two — dual). Two persons, two hearts. But the verse says qulūbukumā: "the hearts of you two", plural.

The explanation the classical grammatical sources give is this: in Arabic, for organs whose possessor is dual and of which there is one apiece in each person, the plural is preferred instead of the dual. The reason the lexicographers give is the heaviness that arises from duals coming one after another — in the form qalbākumā two dual markers come one upon the other and this falls heavily on the Arab ear; the plural is lighter.

The same usage is found elsewhere in the Qurʾān as well; for example for the hand in the verse on theft (al-Māʾida 5/38) and in similar structures. This is recorded as a rule in the books of grammar.

I do not turn this into a point of meaning. That is, I make no interpretation such as "the plural points to the dividedness of the heart" — such an interpretation has been made in the classical sources too, but since the rule of grammar already requires this form, drawing a meaning from it remains without support. I record the rule and add nothing to it.

Root ظ ه ر — treated at the third verse: back, outer face, backing. I do not repeat it.

The pattern: تَفَاعُل (tafāʿul). And here we come again to a pattern that has met us many times in this commentary.

In al-Takāthur the function of this form was recorded in detail: the tafāʿul form states reciprocity (mushāraka) — it shows the act going back and forth between two or more parties. Fakhara (he boasted) → tafākhara (they boasted against one another); kaththara (he multiplied) → takāthara (they vied with one another in multitude). I do not repeat that analysis.

What the pattern adds here: ẓahara is to back someone; taẓāhara is to back one another. That is, the verse speaks not of two persons each taking a position, but of their forming an alliance.

And the difference is large. That two persons hold the same view is one thing; that they stand in that view by leaning on one another is another. In the second, each draws strength from the existence of the other. The correctness of the view is not separately tested — because the other party already thinks the same thing, and this works like a confirmation.

This is a mechanism observed in social psychology: that a conviction is shared by someone else increases the felt confidence in the correctness of that conviction — even if nothing at all has changed in the conviction itself. Even a group of two is enough to produce this effect.

I record this not as a claim of a "scientific miracle" but as the coinciding of the mechanism the pattern of the word describes with a modern observation. The pattern already says it: the verb goes back and forth between the parties.

And the place of the tafāʿul form in this sūra takes on meaning together with the other examples in the corpus:

All four are on the same pattern. And the pattern itself is neutral: reciprocity is neither good nor bad, it takes its value from its subject. The finding recorded at al-Baqara 2/148 holds good here too: the Qurʾān does not forbid competition or reciprocity; it changes their subject.

The second half of the verse gives a list and the form of the list is striking.

"God is his mawlā; and Jibrīl, and the righteous among the believers. And after that the angels too are a backer to him."

The structure of the list sets up an ascent:

1. الله — and first huwa mawlāh is said: He is the one who is mawlā. This is a repetition of the word that occurred in the second verse; there "your mawlā", here "his mawlā". 2. جبريل — the bearer of the revelation. 3. صالح المؤمنين — the righteous among the believers. 4. الملائكة — and after these, the angels.

And the effect of the ascent is this: on one side two persons, on the other side this list. The verse states the meaninglessness of the alliance by showing the disproportion.

The expression صَالِحُ الْمُؤْمِنِينَ is disputed and here I do not speak with certainty. The construct is built with a singular word (ṣāliḥ); against that, its sense may be plural. The classical explanations:

The first explanation requires fewer assumptions linguistically. But I do not impose a preference and it is not binding.

And the real point of grammar is at the end: ظَهِير is singular.

The sentence says "wa'l-malāʾikatu baʿda dhālika ẓahīr". The subject is plural (al-malāʾika — the angels), the predicate singular (ẓahīr — a backer). The expected form would have been ẓuharāʾ or some similar plural.

The explanations given in the classical sources:

All three have been defended in the classical sources. I adopt no preference.

But the third explanation has a beautiful side in point of meaning and I content myself with recording it: if the company has been taken as a single whole, the sentence says this — on one side there are two persons and the two back one another; on the other side there are countless beings and all of them are a single backing. Multitude of number does not guarantee being one; the side that is one is named with a singular word. I write this as my own reading; it is not transmitted.

One. That an alliance does not produce truth.

The mechanism the verse records with the tafāʿul pattern touches one of the commonest errors of daily life: taking the finding of someone who shares a view to be the confirming of that view.

That two persons think the same thing gives no new information about the truth of that thing — both may be looking from the same place, with the same incomplete information, with the same interest. But the felt certainty rises. And because it rises, the matter is not thought through again.

This is a relative of the finding recorded in al-Bayyina: there it was said that knowledge produces division, and that the reason for this is not a lack of knowledge but the using of knowledge to enlarge one's own position. Here the mechanism is on a smaller scale: a ring of confirmation made of two people.

The measure the verse gives is simple and comes from outside: how large is the list facing you? That is, a conviction is weighed not by the number of those who share it but by the weight of what stands against it.

Two. The diagnosis "it has inclined".

The verse's choice of word — not broken but inclined — shows a method.

The word used in describing a fault also determines the fault's correctability. A diagnosis that says "you are like this" leaves the other party no possibility of change; because being is not a thing that can be changed. A diagnosis that says "you have inclined in this direction" describes a movement, and every movement can be undone.

This is what the verse does: it says the situation exactly, but it gives no adjective, it describes a movement.

Its practical counterpart holds both in warning someone else and in assessing one's own state. A person's saying about himself "I am just that sort of person" is a sentence that looks like a diagnosis but in fact makes change impossible. To say "I have inclined in that direction" says the same truth without covering it, and makes something doable.

Three. That repentance is not commanded.

The verse does not command repentance to the most particular addressee; it offers it in the conditional. And this is of the same family as the attitude seen in al-Muzzammil and ʿAbasa: a correction is made, then the one addressed is set free.

What this says to the present has to do with how a warning is made. To turn a warning into a command places the one warning in the position of a supervisor — and from that moment the matter is no longer the wrong that was done but the relation of authority between two people. The conditional form prevents this transformation: the door is shown, no one is pushed.

ʿAsā rabbuhū in ṭallaqakunna an yubdilahū azwājan khayran minkunna muslimātin muʾminātin qānitātin tāʾibātin ʿābidātin sāʾiḥātin thayyibātin wa abkārā

"If he divorces you, it may be that his Lord will give him in exchange wives better than you: submitting, believing, devoutly obedient, repentant, worshipping, travelling (or fasting) — previously married and virgin."

This verse is one of those most open to being misread when torn from its context. Two grammatical signs prevent this and both stand at the head of the sentence.

The first sign: عَسَىٰ. This is a verb in Arabic stating hope and expectation — "it may be that", "perhaps". It states no certainty, it sets up no ruling, it gives no promise.

The Qurʾān uses this word in contexts of threat as well as in contexts of promise; what they have in common is that the outcome is left open. Indeed the eighth verse of this sūra will come with the same word: ʿasā rabbukum an yukaffira ʿankum — "it may be that your Lord will cover over your evil deeds." There the addressees are all the believers and the word again states not a certainty but a possibility.

That is, the verse does not announce a decision; it voices a possibility.

The second sign: إِن طَلَّقَكُنَّ. The conditional form. "If he divorces." And this condition did not come about.

When the two signs are joined the logical structure of the sentence is this: an uncertain consequence of a condition that did not come about. This is the form of sentence with the weakest bindingness there can be.

I stress this because the verse is sometimes relayed as though it were a ruling or a threat. The grammar does not permit that.

The place the verse really works is the list of attributes, and what the list is built upon is the most important information in the verse.

Eight attributes are counted. They must be seen one by one, and then all together.

Now look at the list as a whole.

All of the first six attributes recount the woman's relation with God: submitting, trusting, obedience, turning, worship, and the sixth, in whichever sense it is taken, again a form of worship.

There is not a single attribute in the list to do with the husband. Neither obedience, nor service, nor agreeableness, nor beauty, nor lineage, nor wealth. Not one.

This is the real thing the verse says and it is generally skipped. The khayr in the expression "better wives" is not defined from the side of the husband; it is defined from the side of the relation established with God. That is, the measure comes not from inside the marriage but from outside it.

And this agrees exactly with the account of the four women at the end of the sūra: there too the women will be assessed not by their husbands but by their own choices. The sūra will apply from the tenth verse on the measure it lays down in the fifth.

The last two attributes are of a different kind. Thayyibāt and abkār state not a moral quality but a status: those who have married before and those who have not. And by their being mentioned together something is done: both categories are counted as equal.

This has a consequence: that a woman has married before subtracts none of the attributes in the list. The list counts moral qualities, then sets two marital statuses side by side. No distinction is made; both are counted.

Moreover the root of the word أَبْكَار (ب ك ر) means "to be early, to be first" — from the same root come bukra (the early time of morning) and bākir (first, early). English shows the same in virgin soil and virgin forest, where the word keeps the root's original sense: untouched, in its first state.

Root س ي ح (s-y-ḥ). The concrete sense of the root is the flowing of water over the ground. Sāḥa'l-māʾ — the water flowed, it spread. From this come sayḥ (flowing water, a stream) and siyāḥa (going about a place without stopping, roaming). The Turkish words seyahat (travel) and seyyah (traveller) are borrowings from this root.

The meaning of the word in the verse is disputed in the classical sources:

I adopt no preference. The first view is the most widespread in the sources; but widespread currency is not linguistic proof. The second view is closer to the sense of the root but is strained by historical practice. I do not have enough data in hand to speak with certainty.

The context of the verse is, for the reasons I recorded in the introductory section, a context in which the detail should not be entered upon. The verse speaks of the consequence of a condition, and that condition did not come about.

I content myself with saying this much: the verse names no person, it issues no ruling about any person, and the measure it gives — eight attributes — looks not at persons but at qualities. This means that the rule of method recorded in al-Māʿūn is at work here too: the verse describes an attribute; whoever bears that attribute is the one the verse is speaking of.

Where the measure of "better" is taken from.

What the verse does in a single sentence is to set up a measure of comparison. And where the measure is taken from is more important than the measure itself.

Whether a person is "better" has an answer that changes according to whom and to what the question is put. The verse answers this question not from inside the relationship but from outside it: six of the eight attributes counted are qualities the person carries on her own. These are not produced in a relationship with someone else, and they do not disappear when that relationship ends.

The side of this that faces the present is direct: to measure a person's worth by how far she suits someone else makes that person dependent on the relationship. When the relationship changes the measure changes too; and the person is obliged to negotiate her own worth over again each time.

The list in the verse does not use this measure. And its not using it lays down a principle valid for men and women alike: a person's quality comes before her relationships.

Yā ayyuha'lladhīna āmanū qū anfusakum wa ahlīkum nāran waqūduha'n-nāsu wa'l-ḥijāratu ʿalayhā malāʾikatun ghilāẓun shidādun lā yaʿṣūna'llāha mā amarahum wa yafʿalūna mā yuʾmarūn

"O you who believe! Guard yourselves and your families against a fire whose fuel is people and stones. Over it are angels, thick and hard; they do not disobey God in what He commands them and they do what they are commanded."

This verse is the sharpest turn of the sūra.

For five verses the subject was a single house. The addressee was one person, then two persons, then a group of wives. In this verse it suddenly says "O you who believe" and the command is given to all the believers.

There is no conjunction in between. There is no transitional sentence. The sūra builds no sentence such as "here is the lesson that follows from this". It merely changes the address.

And this transition shows the method of the sūra. The text draws a rule out of a particular incident; but it does not tie the rule to the incident. The incident is left behind, the rule is carried forward.

One must see what is left behind: in the first five verses something had gone wrong inside a house. The sixth verse lays a responsibility upon every house. No link is made but the link is plain: what goes on inside a house is something that enters the reckoning.

Root و ق ي (w-q-y). This root was taken up at al-Baqara 2/2 in connection with al-muttaqīn and treated once more at al-Muzzammil 73/17. I do not repeat those analyses; I recall them in two sentences:

At the centre of the root lies not fear but being guarded — to guard, to be a shield, to put a barrier before a thing. From the same root come wiqāya (protection, a shield), taqwā (being guarded), muttaqī (one who is guarded). And the definition recorded in al-Baqara was this: "The muttaqī is not a coward; he is one who is careful. Like a person who watches where he treads as he walks."

There are three new things here.

First: the form of the word. The imperative of this root is one of the shortest imperative forms in Arabic. The original of the verb is waqā; in the imperative conjugation the initial wāw and the final yāʾ drop away and a single letter (qi) remains; in the plural imperative the plural wāw is added to it and قُوا results.

That is, the root that produces one of Arabic's most heavily laden notions comes down in the imperative to two letters. This is a coincidence, the natural result of grammar; but the contrast between the shortness of the sound and the weight of the command is heard as one reads the verse.

Second, and the principal one: that the verb is transitive.

The most familiar word derived from this root is taqwā, and taqwā is reflexive: a person guards himself. As was recorded in al-Layl, in Form VIII (ittaqā) the verb becomes reflexive.

Here the verb is not reflexive but transitive. And it has two objects:

قُوا أَنفُسَكُمْ وَأَهْلِيكُمْ "guard yourselves and your families"

That is, the root that produces taqwā here takes someone else as its object. Being guarded ceases to be a shield a person draws around himself and becomes a shield set before someone else.

This stands in a place of its own among the uses of و ق ي in the corpus:

Third: what is guarded against. The verse does not say "guard yourselves against God" — that is a formula frequent in the Qurʾān (ittaqu'llāh) and there the object of ittiqāʾ is God, that is, it means "be careful towards Him".

Here the object is fire. Qū … nāran — "guard against a fire". That is, what is to be guarded against is named concretely.

Its effect is this: the command does not ask for an abstract state of carefulness; it asks for a definite shield against a definite danger. And what the danger is is described immediately afterwards.

The verse gives the two objects in a definite order: first yourselves, then your families.

This order could have been reversed, and had it been reversed the sense would have changed. What the order says is this: one who is not guarded himself can guard no one.

This is a principle the Qurʾān sets up openly elsewhere. The question recorded at al-Baqara 2/44 — "Do you enjoin good upon people and forget yourselves?" — is the interrogative form of the same principle. The distinction between "hearing or reasoning" treated in the discussion of al-Mulk passed through a similar place.

And it has a practical consequence: responsibility towards the family does not take the place of a person's own reckoning. The two are counted separately.

The root أَهْل was taken up in al-Bayyina in connection with the expression ahl al-kitāb. I do not repeat the finding there; I recall it in one sentence: the root أ ه ل means to belong to a place or a thing, to form closeness and familiarity with it; ahl are a person's household, and at the centre of the word lie belonging and familiarity. It was recorded there that the expression "People of the Book" means "the household of the Book".

What is new here: the word is used in its original sense, that is, for an actual household. And the "familiarity" dimension of ahl makes the command meaningful — the people you are charged with guarding are those who are familiar with you. Those who know you, who live in the same house as you, who see how you behave.

This explains why the command is given with this word: the way to guard a household passes first of all through being the person they see. Familiarity runs both ways.

Here a difficulty must not be ignored.

The sixth verse commands you to guard your family. The tenth verse will recount that two prophets could not guard their families: the wives of Noah (Nūḥ) and of Lot (Lūṭ) are not saved, for all that they were in the houses of two prophets. And the verse records this not as a fault but as a rule: lam yughniyā ʿanhumā mina'llāhi shayʾā — "they could not ward off anything from them."

There are four verses between the two, and between them there is a real tension:

• 66/6: Guard your family against the fire.
• 66/10: Two prophets could not guard their families.

I do not close this tension with a cheap formula. It is easy to say "so the command asks for effort, not for a result", and it is probably true — but to say this and pass on lightens the real weight the verse leaves upon the reader.

What can be said is this, and I write it as my own reading:

The command in the sixth verse is a command of an act. That acts are commanded does not mean that results are guaranteed — no command in the Qurʾān contains the certainty of its result. Indeed the finding recorded in Nūḥ was exactly this: that sūra gave twenty verses to effort and a single word to the result; and the sentence of the verdict did not come out of Noah's mouth. The boundary between the messenger and the verdict had been engraved there into the grammar of the sūra.

The same boundary works here too. The verse does not say "save your family"; it says "guard". Guarding is an act, saving is a result.

But this does not fully resolve the tension and I do not claim that it does. The sūra on one side lays a responsibility and on the other shows the limit of that responsibility. To carry both at once may be what the text wants of the reader. I leave this not as a solution but as a reading.

وَقُود (waqūd) — fuel, that is, the substance on which the fire feeds.

And here I must enter a caution. The root of this word is و ق د (w-q-d): to catch fire, to burn. From the same root come īqād (kindling), mawqid (a hearth), waqada (it burned).

This is NOT the root of the verb قُوا at the head of the verse. The root of that verb is و ق ي (w-q-y). The two roots share their first two letters and occur side by side in the same verse, so they are extremely liable to be confused. But their third letters differ and their meanings have nothing to do with one another:

I make here the same distinction that was made in Nūḥ between andhir (ن-ذ-ر) and tadhar (و-ذ-ر), and on the same ground: no meaning is to be drawn from a resemblance of sound. That the two words are in the same verse may be a literary choice — I can say nothing certain about that — but no claim of kinship of root can be built.

النَّاسُ وَالْحِجَارَةُ — people and stones.

This expression occurs in one other place in the Qurʾān, in almost the same wording: "So guard yourselves against the fire whose fuel is people and stones" (al-Baqara 2/24). There the context is a challenge; here the context is a command to guard. The same description, with two different functions.

That the fuel is people is worth pausing over. In an ordinary fire the fuel and the thing that burns are separate: the wood is the fuel, what is warmed is something else. Here the fuel and what suffers the punishment are the same. The substance on which the fire feeds is not separated from the substance thrown into the fire.

This has a consequence and I record it as an observation: in this description the fire is not something brought from outside and lit; it is something fed from within. The descriptions the Qurʾān uses elsewhere look in the same direction — the nāru'llāhi'l-mūqada (the kindled fire of God, al-Humaza 104/6) treated in al-Humaza and its "rising over the hearts" are the closest example. Note: the word mūqada there is, like waqūd in this verse, from the root و ق د.

For the mentioning of stones several explanations are given in the classical commentaries: that the idols were of stone, that stone holds fire for a long time, or that it heightens the severity of the description. I have no evidence in hand for a firm preference; I content myself with relaying the explanations.

The question of the keepers of hell was treated at al-Muddaththir 74/30-31 and that section had already mentioned this verse within itself. I do not repeat the findings there — in particular the setting of the number as a fitna and the note "none but He knows the hosts of your Lord". The analysis made in al-ʿAlaq in connection with al-zabāniya attaches to the same place.

What is new here is the two adjectives.

غِلَاظ (ghilāẓ) — root غ ل ظ (gh-l-ẓ). The concrete sense of the root is thickness: a body's not being thin, being bulky and dense. Ghalīẓ — thick, dense, heavy.

The moral sense of the word (harshness, coarseness) is derived from this concrete sense: what is thick does not flex, does not easily take a shape. The English expressions thick-headed and a delicate touch carry the same logic — fineness means flexibility, thickness means resistance.

The Qurʾān uses the same root for the weight of an oath as well: mīthāqan ghalīẓā — "a firm/heavy covenant". There the word is not negative; it states unloosability.

شِدَاد (shidād) — root ش د د (sh-d-d). The concrete sense of the root is to bind tight, to draw taut. Shadda — he bound tightly. The Arabic shadda sign (the mark showing that a letter is read twice) is from this root too: the letter has been "tightened".

From this come shidda (tightness, density, force) and shadīd (tight, strong). The usual English rendering, violence, has narrowed in the same way and stays almost only within "the use of force"; the Arabic field is far wider and its centre is tightness.

The picture the two adjectives set up together: ghilāẓ states volume and density, shidād states boundness and tension. That is, what is described is large and tight — a structure that neither flexes nor slackens.

And note: neither adjective states a moral fault. Rendered into English as "harsh and severe", the words acquire an association of cruelty; what they carry in Arabic is a description of substance and structure.

The next clause confirms this.

The sentence describes the quality of the angels in two halves:

لَّا يَعْصُونَ اللَّهَ مَا أَمَرَهُمْ وَيَفْعَلُونَ مَا يُؤْمَرُونَ "They do not disobey God in what He commands them, and they do what they are commanded."

At first sight this looks like a repetition: is not failing to disobey a command the same as doing the command?

It is not. And the difference gives the definition of the notion of obedience:

That is, obedience consists of two separate works: not doing and doing. And these do not stand in for one another. To have abandoned a thing does not mean to have done what was commanded.

This distinction is another form of a finding that has met us many times in this commentary. The finding recorded at al-Fajr 89/17 was this: the verse does not say "you do not give to the orphan", it says "you do not honour" — the threshold is shifted from not-doing to doing. In al-Māʿūn the same measure was formulated as "not pushing away would have been counted enough; not pushing away is not enough, honouring is required."

Here the same measure is laid down by way of the description of the angels: flawless obedience is the doing of both halves at once.

Root ع ص ي (ʿ-ṣ-y). The concrete sense of the root is interesting: ʿaṣā — a staff, a stick. ʿAṣā also means "to part from a body, to break away" — shaqqa'l-ʿaṣā, "he broke the staff", that is, he parted from the community.

From this comes ʿiṣyān (breaking away from a command or from a body). That is, disobedience is by its root not an "opposing" but a parting.

One detail shows the weave of the sūra.

In the sixth verse the angels are qualified as غِلَاظ. Three verses later, in the command given to the Prophet in the ninth verse, the same root comes:

وَاغْلُظْ عَلَيْهِمْ (66/9) — "be hard towards them"

That is, what is a quality of the angels becomes a command given to the Prophet. The same root, in one verse an attribute, in another an instruction.

I record this as an observation. And I enter this note as well: I draw from it no conclusion such as "man ought to be like an angel" — there is no such conclusion in the text. What I record is only the root's circulation within the sūra.

But one more thing may be said and I write it as my own reading: ghilẓa is in the sixth verse the quality of a keeper of punishment and carries there no negative association. In the ninth verse it is commanded as a stance, and there too — in the context of jihād — it must be read not as a command to wrong anyone but as a command not to slacken. The original sense of the root (thickness, not flexing) is preserved in both places.

One. That responsibility crosses the boundary of the house.

This is the most direct thing the verse says and one of the places most resisted today.

At the centre of the modern moral understanding stands individual responsibility: everyone is responsible for himself, no one is called to account for another's choice. The Qurʾān itself sets up this principle openly — the sentence recorded at al-Baqara 2/134, "what they earned is theirs, what you earn is yours", and the principle at al-Anʿām 6/164, "no bearer of burdens bears another's burden", must be read together with this verse.

The verse does not break this principle. Had it broken it, it would have said "you are responsible for your family's reckoning". It does not. It says "guard" — that is, it commands an act, it does not transfer a reckoning.

The difference between the two is this: a person is not responsible for the choices of his household; but he is responsible for what he does towards them. What he said, what example he set, what he ignored is written to his own account.

This distinction prevents both of the errors at the two extremes:

• The first error: doing nothing at all towards one's household on the ground that "everyone is responsible for himself". The verse closes this off.
• The second error: counting the household's choices as one's own reckoning and putting pressure on them. al-Baqara 2/124 and 2/134 close this off.

Two. What guarding is not.

The verse says qū (guard); it does not say ajbirū (compel), it does not say uḥkumū (rule).

And the sense of the root carries this difference: wiqāya is holding a shield. One who holds a shield does not take the place of the one guarded; he comes between him and the danger. This is a different work from controlling.

In practice the difference shows here: control narrows the field of choice of the one guarded; a shield meets the danger. The first weakens the person — the one guarded cannot learn to make his own decision; the second strengthens him.

And the definition of muttaqī recorded at al-Baqara 2/2 serves once more here: "a person who watches where he treads as he walks." The same holds for the family: not to walk in their place, but to enable them to see where they are treading.

Three. The practical counterpart of the order of the command.

First yourselves, then your families. This order carries an ancient observation about how education inside a house works: a household takes in what is seen far more than what is said.

The side of this that faces the present is direct and uncomfortable: in a house it is not the words most repeated but the behaviours most repeated that are transmitted. Children learn not what is said to them but what the parent says to himself.

The order of the verse records this by grammar. And it attaches to the same place as the finding "the proof is a person" recorded in al-Bayyina: "whoever carries an idea is a part of the proof of that idea." A house is the place where this proof is presented on the smallest scale.

Yā ayyuha'lladhīna kafarū lā taʿtadhiru'l-yawm, innamā tujzawna mā kuntum taʿmalūn

"O you who deny! Do not offer excuses today. You are only being repaid for what you were doing."

This verse is the shortest and the harshest verse of the sūra. And in point of word order it is striking.

The previous verse was said in this world: "O you who believe, guard." This verse is said on the day of resurrection: "do not offer excuses today."

That is, in passing from one verse to another the text skips over time. The reader is carried from a command in this world to the moment in which the consequence of that command is seen. The whole distance in between — a lifetime, a resurrection — is compressed into the gap of a single verse.

This leap is common in the Qurʾān's technique of narration and its effect is this: the distance between the warning and the subject of the warning is removed. A finding recorded in the discussion of al-Maʿārij was close to this: the text carries the reader over to the other side of the happening.

And one detail: the formula of address stays the same. The sixth verse yā ayyuha'lladhīna āmanū, the seventh yā ayyuha'lladhīna kafarū. The same frame, a single word changed. Two groups are called in the same manner — and the difference is only in the verb.

Root ع ذ ر (ʿ-dh-r). The field of meaning of the root is wide and it has an interesting centre: to wipe out the trace of a thing, to cover a fault, to make excusable.

• عُذْر (ʿudhr) — an excuse, a reason that lightens a fault. The Turkish özür is a direct borrowing from this.
• مَعْذِرَة (maʿdhira) — the offering of an excuse.
• اعْتَذَرَ (iʿtadhara) — the iftiʿāl form: he produced an excuse for himself, he tried to make himself excusable.

What the pattern adds is decisive here. The iftiʿāl form, as was treated at the first verse in connection with ibtighāʾ, states reflexivity and effort: a person's doing something for himself, with exertion.

That is, iʿtidhār is not a passive "having an excuse"; it is an effort spent in producing an excuse. The verse forbids not the existence of an excuse but the search for one.

And this difference is needed in order to understand the verse exactly. The sentence does not say "you have no excuse" — that would be a finding. It says "do not look for an excuse" — that is a command and it aims at a behaviour.

The parallel usage in the Qurʾān. The same verb occurs once more in a similar scene: expressions to the effect that "that day their excuses will not profit them" and the formula "do not offer excuses today" are repeated in the Qurʾān. What they have in common is that the time for excuses has passed.

What the verse says is not the invalidity of the excuse but the closing of its time. An excuse is meaningful when it can change something.

إِنَّمَا — in Arabic the particle of ḥaṣr (restriction): "only, solely, nothing else." In al-Bayyina at 98/4 the restriction built with the mā … illā structure was treated; innamā does the same work in a single word.

The sentence says: what you see is nothing else at all, only the return of what you did.

This restriction cuts off the search for an excuse at the root. For an excuse is an attempt to put a distance between the outcome and the act: "I did it, but for such a reason", "I did it, but others did too", "I did it, but I did not know". The sentence closes this distance — there is nothing at all between the outcome and the act.

Root ج ز ي (j-z-y): to give in return, to pay what is equivalent. The usual English rendering, penalty, is used only for a negative return; in Arabic the word is neutral and is used for the return of good as well. This distinction is needed in order to read most of the uses of jazāʾ in the Qurʾān correctly.

مَا كُنتُمْ تَعْمَلُونَ — note the combination of forms: kuntum (perfect) + taʿmalūn (imperfect). In Arabic this combination states a continuous past: "what you were doing", "what you kept doing".

That is, what is repaid is not single events but a behaviour kept up. In al-Māʿūn the imperfect's stating of continuity was dwelt on several times; here the same principle comes in its backward-looking form.

The verse's place within the sūra is strange at first sight. The sixth verse addresses the believers, the eighth verse addresses the believers again; between them this verse addresses those who deny.

In the classical commentaries it is noted that this is an iʿtirāḍ (an interposition). But its function can be seen and it is twofold:

First: it shows the other side of the sixth verse. The sixth verse commanded guarding against a fire. The seventh shows what one who has not been guarded says beside that fire. That is, between the command and the ground of the command, the living scene of the ground is set.

Second: it prepares the ground for the eighth verse. The eighth verse will command repentance. And the opposite of repentance is what is shown in the seventh verse: the excuse.

This opposition explains the verse's presence here and is worth pausing over:

At al-Baqara 2/37, where Adam (Ādam) and Iblīs are compared, exactly this distinction was recorded: "The two beings were in the same situation: they broke a command. What made the difference is what they did afterwards." Adam took the fault upon himself, Iblīs gave grounds for it.

The seventh and eighth verses of Sūrat al-Taḥrīm place the same distinction in two neighbouring verses.

The function of the excuse.

The verse does not describe an excuse as a lie. This is important: excuses are most often true. The person really was tired, really did not know, others really were doing it too.

What the verse touches is not the truth of the excuse but its function. An excuse puts a distance between the act and the person. And when that distance is set, the act becomes impossible to work on — because the act is no longer the person's.

The sense of effort added by the iftiʿāl form falls exactly into place here: an excuse is not found, it is produced. And its production requires labour; the human mind most often spends this labour without noticing.

Its practical consequence is this: the first thing a person who wants to change a behaviour has to do is to be able to see the behaviour itself and not its explanation. So long as the explanation stands in front of the behaviour, the behaviour stays where it is.

And the close of the verse says this in a single sentence: innamā tujzawna mā kuntum taʿmalūn. What is measured is not the explanation but what was done.

Yā ayyuha'lladhīna āmanū tūbū ila'llāhi tawbatan naṣūḥā, ʿasā rabbukum an yukaffira ʿankum sayyiʾātikum wa yudkhilakum jannātin tajrī min taḥtiha'l-anhār, yawma lā yukhzi'llāhu'n-nabiyya wa'lladhīna āmanū maʿah, nūruhum yasʿā bayna aydīhim wa bi-aymānihim yaqūlūna rabbanā atmim lanā nūranā wa'ghfir lanā, innaka ʿalā kulli shayʾin qadīr

"O you who believe! Turn to God with a sincere, unmixed repentance. It may be that your Lord will cover over your evil deeds and admit you into gardens beneath which rivers flow — on a day when God will not disgrace the Prophet and those who believed with him. Their light runs before them and on their right. They say: 'Our Lord! Complete for us our light and forgive us. Surely You have power over all things.'"

The root ت و ب (t-w-b) was taken up at al-Baqara 2/37; I do not repeat that analysis. In one sentence: the sense of the root is to turn, and the same root is used both of the servant and of God — the servant tāba ila'llāh, God tāba ʿalayh; and God's turning comes before the servant's turning.

In the fourth verse this root came in the conditional (in tatūbā); here it comes in the imperative (tūbū), and the addressees are all the believers.

What is new here is not the root itself but the adjective attached to it. The word tawba occurs in many places in the Qurʾān; this is the only place where it is qualified by نَصُوح. This calls for the word to be examined separately.

Root ن ص ح (n-ṣ-ḥ). This root has two separate branches in Arabic. Both are recorded in the classical dictionaries, both fit this verse, and the picture that emerges when the two are read together explains the word completely.

The first branch: purity, unmixedness.

The clearest use of this branch of the root is for honey: نَاصِح العَسَل (nāṣiḥu'l-ʿasal) — pure honey, entirely separated from its wax, its comb, its admixture.

From this comes the sense nuṣḥ (sincere counsel). And the way it comes out is this: counsel counts as naṣīḥa only when it has been purified of the counsellor's own interest. If there is a reckoning inside it, that honey is adulterated.

This shows the original sense of a word that has narrowed. The usual English rendering, admonition or preaching, has taken on the colour of a long and tiresome lecture; the centre of the Arabic word is purity. Counsel is speech from which the giver has taken out his own share.

The Qurʾān uses this sense plainly: the prophets, addressing their peoples, say "I am giving you counsel" (anṣaḥu lakum), and immediately afterwards there most often comes the note "I ask of you no wage". The two sentences stand side by side, because the second is the condition of the first.

The second branch: to sew, to mend.

The second and less known branch of the root has to do with tailoring:

• نِصَاح (niṣāḥ) — sewing thread.
• مِنْصَحَة (minṣaḥa) — a sewing needle.
• نَصَحَ الثَّوْبَ — he sewed the garment, he closed its tear.

This branch gives the real picture the word adds to the verse, and it is worth pausing over.

Repentance is a repentance that sews the tear.

Let us open out this image, because by itself it gives a description of repentance.

There is a difference between a tear and a stain. A stain is on the surface; it is wiped, washed, covered over. A tear, however, is a break in the cloth itself — covering it does not remove it, because the problem is not on the surface but in the fabric.

And note what a stitch does: a stitch does not abolish the tear. In sewn cloth the mark of the stitching remains; the tear is where it was, but it no longer opens. The cloth is whole again and can be used again.

This says two things about repentance and both are worth noting:

First: repentance does not make what happened not to have happened. A mark remains. The Qurʾān says this already — it says "He will cover over your evil deeds" (yukaffira ʿankum), it does not say "He will abolish them". To cover over is to close what happened; not to make it not to have happened.

Second: but wholeness comes back. Sewn cloth is not torn cloth. It serves.

And a stitch requires labour: a needle, thread, time, attention. A tear does not close by itself. The weight the adjective naṣūḥ adds lies here: the word adds to repentance a quality of effort.

The joining of the two branches. I record this as an observation: the two branches, though they look unrelated at first sight, come out at the same place. Pure honey is what has had the admixture separated out; sewn cloth is what has had the separated joined together. Both are an operation of separating and joining. In the classical dictionaries the two branches are given under separate entries and no necessary link is drawn between them; for that reason I write this not as an etymological claim but as the falling of two senses together into the verse.

And the root has one further branch: the word nuṣḥ is used in Arabic also in the sense of "wanting someone's good, wishing well for him". On this reading tawbatan naṣūḥā becomes "a repentance a person makes wanting his own good". This explanation too occurs in the classical sources.

The grammatical dispute of the verse is here.

The word naṣūḥ is of the faʿūl pattern. And this pattern in Arabic generally carries an active sense (an intensive active participle): one who does much, who does continually. Example: ghafūr (much forgiving), shakūr (much thanking).

The problem is this: tawba is a verbal noun, and verbal nouns cannot be "counsel-giving". That is, the adjective looks incongruous with what it describes.

The classical explanations:

All four have been defended in the classical sources. I do not impose a preference.

But I content myself with recording a beautiful side of the fourth explanation in point of meaning: if repentance is a thing that counsels its owner, then repentance becomes not a closing but a beginning — not something that turns back and closes the matter, but something that goes on speaking forward. I write this as a possibility of reading; not as a preference.

Root ك ف ر (k-f-r). This root was taken up at al-Insān 76/3 and in al-Baqara; I do not repeat it. In one sentence: the sense of the root is to cover, to close over — the farmer is called kāfir, because he covers the seed with earth.

What is striking here is the use of the same root in two opposite directions.

• Kufr: the servant's covering over what has been given to him.
• Takfīr (as in this verse): God's covering over the servant's evil deeds.

The same verb, the same form (tafʿīl), two different subjects — and entirely opposite values. Covering becomes, according to who covers and what is covered, either the greatest ingratitude or the greatest mercy.

I record this as an observation. It is an example of a fact that has met us many times regarding the Qurʾān's economy of words: the same root changes value according to its position.

عَسَىٰ — treated at the fourth and fifth verses: it states hope, not certainty.

And here there is a striking point. The verse commands repentance (tūbū — imperative), but it does not make the outcome certain (ʿasā — it may be that). That is, the obligation is certain, the recompense hopeful.

This asymmetry is meaningful and says something: repentance is not a transaction that buys the outcome. The servant does what falls to him; the outcome is in the hands of the other party. This looks to the same place as the question of riḍā treated at the last verse of al-Bayyina — a deed is not a payment but an attempt at approach.

Root خ ز ي (kh-z-y): to be disgraced, to be left in shame, to be brought low. Khizy — disgrace, abasement.

What the word carries is a state fallen into before others. That is, not merely a pain; a pain undergone while being seen.

And the choice of this word brings out the social dimension of the scene. The resurrection is often described in the Qurʾān as a place where everyone is together; and part of what is undergone there is being seen.

The sentence of the verse is built negatively: lā yukhzī — He will not disgrace. The expected form could have been positive: "He will exalt", "He will honour". The negative form has been chosen.

This has an effect: the sentence names the thing feared and abolishes it. That is, the assurance given to the one addressed is not something he will have; it is something that will not happen to him. This is a form of consolation closer to human psychology — because a thing feared is not consoled until it is named.

The most pictorial scene of the sūra, and it deserves careful reading.

يَسْعَىٰ — root س ع ي (s-ʿ-y): to walk fast, to run, to run about after a task. The saʿy of the pilgrimage is from this root, and the Turkish mesai (working hours) is a borrowing from it.

The verb has been chosen strikingly. The verse does not say "their light is before them", it does not say "their light lights them". It says "their light runs".

That is, the light is not a standing source of illumination; it is something that moves. And the direction of its movement is in front of its owner.

بَيْنَ أَيْدِيهِمْ وَبِأَيْمَانِهِمْ — "before them and on their right."

Bayna aydīhim means in Arabic "between their hands" and idiomatically it means "before them".

Bi-aymānihim — "on their right side". And here the root ي م ن treated at the second verse comes back: there aymān meant "oaths", here the same plural form means "right sides". In the same sūra, six verses apart, the same word in two different senses.

I record this not as a claim of wordplay but as an indication of the root's liveliness: Arabic keeps the link between "the right hand" and "a word given" inside the word, and this sūra uses both.

The mention of the right side comes from the same symbolic field as the distinction between aṣḥāb al-yamīn and aṣḥāb al-shimāl treated in al-Ḥāqqa; I do not repeat that discussion.

Note: a scene like this one is recounted in Sūrat al-Ḥadīd as well, and there the hypocrites' asking for light and not receiving it is added. The scene here is shorter and the other party is not mentioned.

The finest place in the verse — perhaps in the sūra — is here.

The situation is this: the light has already been given. The verse says so: nūruhum yasʿā — their light runs. The scene is a scene in which the light is present.

And at exactly that moment they pray: "Our Lord, complete for us our light."

That is, they ask that what is in their hands be completed. They make a request although they possess it.

This must be dwelt upon, because it describes an attitude.

أَتْمِمْ — root ت م م (t-m-m): to be complete, to be without lack. Itmām — completion. The word asks that the remaining part of something incomplete be given.

That is, inside the prayer there is an admission: the light given is incomplete. And this admission is the most striking side of the scene — a confession of lack even at the moment of salvation.

In the classical commentaries several explanations are given for this prayer: the light's diminishing as one advances on the bridge, or each believer's light being in the measure of his deeds and his asking for more. These explanations rest on reports and there are differences among them; I prefer none of them.

But what follows from the structure of the prayer stands independently of the explanations, and I record it:

The light has been given and is still asked for. This is a form of asking and its counterpart in this world is direct: not counting what one has as enough, but not denying that one has it either. The prayer does not say "give us light" — that would be to treat what was given as nothing. It says "complete our light" — it owns what was given and asks for its continuation.

And immediately afterwards: وَاغْفِرْ لَنَا — "forgive us". A request for forgiveness from the mouths of people who have entered the garden. This too is part of the same attitude.

A link: the finding recorded in al-Bayyina — that at the summit of the reward there is placed not an object but a relation — appears here in another form. The scene gives the moment of salvation as a moment of speech. Those who are saved do not fall silent; they ask. The relation continues.

The length and structure of the verse. This is the longest verse of the sūra and it houses more than one scene within it: a command (repent), a promise (He covers, He admits), a note of time (on that day), a description (the light runs), and a prayer (complete).

And the arrangement works like a lens widening backwards:

1. Command — now, to the one addressed. 2. Hope — the near future, the outcome. 3. That day — the far future, the scene. 4. Description — the inside of the scene. 5. Prayer — the mouths of the people in the scene.

That is, the verse begins with a command and carries the one addressed step by step to the moment in which the consequence of that command is lived, and at the very end speaks from his mouth. At the end of the verse the one addressed hears his own voice.

This technique was seen in the discussions of al-Ḥāqqa and al-Maʿārij too: the reading out to the one addressed of his own sentence in the scenes of the resurrection.

One. That repentance is a work of repair.

The sewing sense of naṣūḥ takes repentance out of being a feeling and makes it a work.

The common conception of repentance is emotional: to feel remorse, to grieve, to resolve not to do it again. These are necessary but they do not sew a cloth.

The image of sewing adds this: if there is a tear, there is a particular place that must be repaired. A general remorse does not close a particular tear. Whoever's whatever you tore, that is what is to be sewn.

Its practical counterpart is already recorded in the classical conditions of repentance — that where there is a claim of another person it must be restored is a measure widely repeated in the tradition. The word itself already carries this measure.

Two. The measure of unmixedness.

The first branch of the root — pure honey — lays down another measure for repentance: what has got mixed into it?

The things that can get mixed into a repentance are known: the discomfort of having been caught, fear of consequences, concern for reputation, the continuation of blaming someone else. None of these invalidates repentance — a person rarely acts with a pure intention. But the word shows a direction: to pick out what has got mixed in.

And it is of the same family as the measure recorded in the discussion of ostentation in al-Māʿūn: the value of an act is measured less by whether it was done than by what it was done with an eye to.

Three. Asking while owning what was given.

The last prayer of the verse shows a form of praying, and this form is rare today.

There are two common extremes: either not seeing what one has and always speaking of the lack, or letting go of asking by saying "one ought to give thanks". What the prayer does is to pass between the two: it names what was given ("our light"), and it asks for its continuation ("complete").

This shows that thanks and request can stand in the same sentence. The finding recorded in al-Fajr — that a person counts abundance an honouring and straitness a humiliation, and that both are rejected — is the opposite of this prayer: there a person was drawing a verdict from what he received, here he counts what he received and asks for its continuation.

Yā ayyuha'n-nabiyyu jāhidi'l-kuffāra wa'l-munāfiqīna wa'ghluẓ ʿalayhim, wa maʾwāhum jahannamu wa biʾsa'l-maṣīr

"O Prophet! Strive against those who deny and against the hypocrites, and be hard towards them. Their abode is hell; and what an evil destination!"

At first sight this is the verse that stands most out of place in the sūra. Its subject is neither a family matter, nor repentance, nor the four women.

The same verse occurs in another place in the Qurʾān as well, in almost identical wording (al-Tawba 9/73). That is, the expression is an established formula.

Then what is it doing here? Several readings are possible and I present none of them as certain:

The third reading is the most functional in point of word order: the ninth verse says al-kuffār, the tenth gives an example for li'lladhīna kafarū; the eleventh gives an example for li'lladhīna āmanū. That is, this verse sets up the parties of the two coming examples.

I record this not as a preference but as a possibility of reading.

Root ج ه د (j-h-d). The concrete sense of the root is to expend force, to spend one's strength, to be put to effort. Jahd — strength, effort. Juhd — capacity. Ijtihād — trying to reach a conclusion in a question by expending all one's strength on it. The Turkish içtihat (independent legal reasoning) and cihat (striving) are from this family.

The pattern: مُفَاعَلَة (mufāʿala). This form, as was recorded in the discussion of al-Maʿārij too, states a two-sided effort: the act is done reciprocally, or an exertion one party keeps up against another is recounted.

That is, jihād is not a one-directional act of attack; it is the name of a reciprocal struggle. The pattern says so.

And I must enter a note about the scope of the word. The word jihād is used in the Qurʾān over a wide field: with wealth, with life, with the tongue, with the Qurʾān. In Sūrat al-Furqān it says "strive against them with it — that is, with the Qurʾān — with a great striving" (al-Furqān 25/52), and there the instrument is explicitly the text itself.

In this verse no instrument is stated. There is only the verb and its object. I draw no instrument out of the verse itself; I content myself with recording the general scope of the word.

And as the method of this text requires: the legal framework of jihād, its conditions, against whom and how it is to be done, are the subject of the literature of law and I issue no ruling on this. Nor do I enter into current political disputes — this is one of the rules of method to which this commentary is bound.

A detail of word order: the verse mentions two groups together — al-kuffār and al-munāfiqīn. But with a single verb. That is, the same verb is directed at two different groups, and the situation of the two groups is not the same: one is openly opposite, the other inside and hidden.

This shows of itself that the word cannot be reduced to a single method: an open method of conflict cannot in any case be applied against a hidden opposition. I record this as an observation.

Root غ ل ظ — treated at the sixth verse: thickness, density, not flexing. I do not repeat it.

What must be noticed here is what the word does not say. The verse does not say uqtul (kill), it does not say iẓlim (wrong), it does not say ubghuḍ (hate). It says "be thick" — that is, do not flex, do not slacken, do not soften.

And the original sense of the root supports this reading: ghilẓa is not a way of behaving but a consistency. What is thick does not change shape under pressure.

The context of the verse requires this reading. What had the first verse of the sūra corrected? A softening. The Prophet had changed a boundary God had set, in order to please the other side.

The ninth verse sets up the other face of the same subject: not slackening in matters of boundaries.

Setting the two verses side by side shows that the sūra sets up an axis:

At first sight they look contradictory: one seems to ask for slackening, the other for hardness. But the common point is this: in both, the measure is set not by the person's own preference but from outside. In the first the person narrowed of his own accord; in the second he is asked not to slacken of his own accord. In both verses what is corrected is adjusting the setting oneself.

I write this as my own reading; it is not binding.

Root أ و ي (ʾ-w-y): to take refuge in a place, to shelter, to come back. Maʾwā — a place of refuge, a shelter. From the same root comes īwāʾ (to shelter someone).

The interesting side of the word: it is a positive word. A maʾwā is the place a person returns to at the end of the day — home, nest, refuge. The Qurʾān uses the same word for the garden as well (jannatu'l-maʾwā).

Its use here for hell produces a contrast: fire as a place of refuge. And this contrast is a technique repeated several times in the Qurʾān's descriptions of punishment — the giving of a comfortable word to an uncomfortable object.

وَبِئْسَ الْمَصِيرُ — "what an evil destination!" مَصِير, root ص ي ر: to turn into a state, to arrive at an outcome. Maṣīr — the place arrived at, the state turned into.

This formula is repeated in the neighbouring sūras too; Sūrat al-Taghābun will use the same expression (64/10). One of the verbal kinships between the two sūras.

The adjusting of the measure by oneself.

The axis I recorded above — narrowing in the first verse, a warning against slackening in the ninth — says something practical.

A person constantly adjusts the measure according to his relation with the other party. He slackens it before someone he loves, he tightens it before someone he does not. And this adjustment is most often made without being noticed; the person is not even aware that he has changed the measure — he counts his own setting as "what the situation required".

The two verses at the two ends of the sūra show the same error in two different directions in the same person: in one place an extra narrowing, in another the risk of slackening.

The common lesson is this: the measure must be independent of the relationship. Whom you are facing does not change what is right. This holds both towards those near and towards those far — and in both it is hard.

The last three verses of the sūra are one of the most striking structures in the Qurʾān. Before entering the verses one by one one must see the structure itself, because the structure says more than the verses.

The Qurʾān gives two mathals — one for those who deny, one for those who believe. And the subject of both examples is women.

Look at the table. The condition of the husbands and the fate of the wives are in exact inverse proportion. Two wives of prophets lose, the wife of a tyrant gains. And the fourth example breaks the system from outside by leaving the marriage column entirely empty.

This cannot be an accidental arrangement. What the structure sets up is a single principle which I shall treat in detail below: salvation neither comes with marriage nor goes with marriage.

And one thing more: the verses mention the own acts of all four. For the first two khānatā (they betrayed), for the third qālat (she said — a prayer), for the fourth aḥṣanat, ṣaddaqat, kānat. That is, each is mentioned by what she did; not by the house she was in.

Root م ث ل (m-th-l): to resemble, to be equal. Mithl — like, equal. Mathal — an example, a comparison, a likeness. Timthāl — a statue, an image. The Turkish misal, temsil, emsal, mesela are from this family.

ضَرَبَ مَثَلًا — word for word "he struck an example". In Arabic the verb ḍaraba is used over a very wide field (to set out on a road, to pitch a tent, to mint coin, to give an example) and the usage here is idiomatic.

The explanation the lexicographers give is this: to give an example is like striking a seal or a coin — an abstract sense is made visible by being stamped into a concrete mould. I relay this explanation as a possibility.

And note: the "mathal" given here is not an invented comparison. In some of the places where the Qurʾān uses the word mathal there are descriptive comparisons (a spider's web, a fire going out); here, however, real persons are mentioned. That is, a mathal is not an imagined story; it is an example case chosen from history.

Ḍaraba'llāhu mathalan li'lladhīna kafaru'mraʾata Nūḥin wa'mraʾata Lūṭ, kānatā taḥta ʿabdayni min ʿibādinā ṣāliḥayni fa-khānatāhumā fa-lam yughniyā ʿanhumā mina'llāhi shayʾan wa qīla'dkhula'n-nāra maʿa'd-dākhilīn

"God set forth, for those who deny, the example of the wife of Noah and the wife of Lot. Both were under two righteous servants from among Our servants; but they betrayed them. Those two could ward off from them nothing that came from God. And it was said: 'Enter the fire, along with those who enter.'"

In the discussion of Nūḥ a list was made of what Sūrat Nūḥ does not recount, and in that list stood this item: "There is no wife of Noah (she occurs at al-Taḥrīm 66/10)."

The finding recorded in that section was this: the omissions in Sūrat Nūḥ are not an omission but a choice — that sūra took not the event side of the story but the effort side.

Now the person not mentioned in that sūra is mentioned here. And the manner of the mention is striking: Sūrat Nūḥ recounted a prophet's effort over twenty-eight verses; this verse says in a single sentence that a person from that same prophet's house took no share in that effort.

Setting the two texts side by side gives the sharpest consequence of the boundary recorded in Nūḥ — the boundary between the messenger and the verdict: the messenger is not the owner of the verdict even in his own house.

Root م ر أ (m-r-ʾ): a human being, a person. Marʾ — a person (a man); marʾa — a woman. From the same root come murūʾa (manliness, humanity) and imruʾ (a person).

The striking side of the word is this: Arabic has two words for "spouse", and the verse uses one here and the other elsewhere.

• زَوْج (zawj) — a pair, one of a couple. Root ز و ج: two things forming a pair. This word is used both for the man and for the woman and is common in the Qurʾān. The form azwāj occurred in the first, third and fifth verses of this sūra too.
• امْرَأَة (imraʾa) — a woman; and in the construct state, "the wife of so-and-so".

The difference between the two words has been treated in the classical commentaries. The distinction commonly recorded is this: zawj is used in cases where the two parties really form a pair; imraʾa is preferred in cases where the pairhood remains deficient in some respect — such as a difference of belief or the absence of a child.

As proof of this distinction the Qurʾān's own usages are adduced: imraʾa for the wives of Noah and Lot, imraʾa again for the wife of Pharaoh, imraʾa again for the wife of Abraham (Ibrāhīm) — in the context of the tidings of a child; whereas zawj in the contexts of the garden and of ordinary marriage.

I do not impose this as a rule and I enter this note: there are also places in the Qurʾān where the word imraʾa is used without implying any deficiency. To claim that the distinction works everywhere would be forced. But the usage in this verse is one of the clearest examples supporting the distinction — because the whole matter of the verse is precisely the pairhood's not having worked.

Kānatā taḥta ʿabdayni — word for word "they were under two servants".

This is an idiomatic expression used in Arabic for marriage and it means "to be under someone's marriage contract". Turkish has the same idiom (nikâhı altında, "under his marriage contract") and it probably passed over from Arabic.

The historical background of the idiom has to do with the legal structure of the period, and I write this not as an endorsement but as a note of history: this use of the word is the trace of a language in which marriage was thought of as a relation of protection.

And by using precisely this idiom the verse refutes what the idiom implies. "To be under" implies a protection; the verse says that that protection did not take place. The choice of word and the verdict of the sentence work against one another.

I record this as my own reading.

The most striking choice of word in the verse is here and it is generally skipped.

The verse speaks of two prophets. And it does not mention them as prophets. It does not say nabiyyayn, it does not say rasūlayn. It says عَبْدَيْنِ: two servants.

Moreover it reinforces the construct twice over: ʿabdayni min ʿibādinā — "two servants from among Our servants". The same root twice, one after the other.

The root ع ب د (ʿ-b-d) was taken up at al-Fātiḥa 1/5 and mentioned once more in Nūḥ; I do not repeat it. In one sentence: the concrete image of the root is a road's being levelled by being trodden underfoot (ṭarīqun muʿabbad), and at its centre lies walking in submission.

But why has this word been chosen? The logic of the verse requires it.

What the verse says is this: the position of these two persons was not enough to save their wives. Had the verse mentioned them as "two prophets", the sentence would have been a comparison of ranks. By saying "two servants" it brings the position down from the start to the plane of servanthood.

And the consequence of this is: the reason for their not being able to save was not their insufficiency; it was no one's being a saviour. A prophet too is a servant; and a servant cannot take on the reckoning of another servant.

The finding recorded at al-Baqara 2/124 holds good here once more: there too the rule was said to a prophet who was asking, and it was not bent for anyone.

صَالِحَيْنِ — root ص ل ح (ṣ-l-ḥ): a thing's not being spoiled, being in working order, being sound. Ṣalāḥ (soundness), iṣlāḥ (setting right), maṣlaḥa (what is to the benefit of a matter), ṣulḥ (the repairing of a broken relation) are from the same root.

At the centre of the word lies not a moral loftiness but the not having broken down of workability. For a tool to be ṣāliḥ is for it to be working.

And note: this adjective occurred in the fourth verse too — ṣāliḥu'l-muʾminīn. The same word, twice in the sūra, in both places in a context of support or of position.

The word in this verse most open to being misunderstood is here and it must be treated carefully.

Root خ و ن (kh-w-n). The central sense of the root in Arabic is this: to diminish something entrusted to one, not to hand back the trust in full.

Khiyāna is defined in Arabic as the opposite of amāna. And this definition gives exactly what the word is:

• أَمَانَة (amāna) — root أ م ن: trust. To leave a thing with someone in the confidence that it will be asked back.
• خِيَانَة (khiyāna) — not to give back in full what was left; to diminish it from within.

That is, betrayal is by its root a diminishing. Not the denial of the whole trust — that is another thing; the withholding of a part of the trust.

The lexicographers record the "hiddenness" dimension of the root as well: khāʾinatu'l-aʿyun — "the treachery of the eyes" (it occurs in Sūrat al-Muʾmin), that is, what the glance steals away in secret. Betrayal is not done openly; the diminishing is done so as not to be noticed.

Now, the question of what this betrayal was.

In English the sentence "his wife betrayed him" comes to almost a single meaning. In Arabic and in this verse the case is not so, and this must be written plainly.

In the classical tradition of exegesis there is an extremely strong agreement about what this betrayal was: the betrayal is in the religious field, not in the field of marital fidelity. That is, both worked from inside their own houses for the other side, against the cause their husbands carried.

This is not a minority view; it is the position of the overwhelming majority of the exegetes, and its ground is sound as well:

First: the Qurʾān's own word. The verse does not use the word zinā; it uses khiyāna. And khiyāna, as was seen above, is a diminishing done against a trust. What is a trust in a prophet's house is the message carried.

Second: the general acceptance regarding the prophets. It is widely recorded in the tradition that no prophet's wife is charged with unchastity; it is noted that this has to do with an acceptance concerning the standing of prophethood.

Third and strongest: the Qurʾān's own narrative. In the places where the story of Lot is recounted, it is stated plainly that Lot's wife was among those who stayed behind and that she was destroyed together with her people (e.g. al-Aʿrāf 7/83; Hūd 11/81; at al-Ḥijr 15/60 al-ghābirīn — those who stayed behind). That is, her position is described by her being on the side of her people.

Fourth: the verse's own logic. The verse gives these two women as "an example for those who deny". That is, what is at issue is not a moral fault but a choice of side. The example is built upon denial.

When I combine these four grounds, the reading that the betrayal is in the religious field seems to me clearly the stronger. I write this as a preference; but it is a preference on which the classical exegesis is almost unanimous.

And one further note: the verse does not say what the betrayal was. The attitude the sūra has kept up from the beginning works here too: the act is named, its detail is not given.

The verdict of the verse is here.

Root غ ن ي (gh-n-y): to have no need, to suffice for oneself. Ghinā — wealth; but the central sense in Arabic is not "to have much property" but not to be in need of another. Ghanī — self-sufficient. And this is one of the names of God; indeed it will occur in Sūrat al-Taghābun too (wa'llāhu ghaniyyun ḥamīd, 64/6).

The formula أَغْنَىٰ عَنْهُ — "he warded (a thing) off from him, he made him free of need, he profited him." The formula recounts one person's standing in another's place and meeting his need.

The construction of the sentence sets up a zero: lam yughniyā … shayʾan — "nothing at all". Indefinite and in a negative context; in Arabic this combination states complete negation. It leaves not even a partial benefit.

And this structure is a formula repeated in the Qurʾān. The same root, the same negation, different subjects:

• For wealth and children: "Neither their wealth nor their children will profit them anything against God" (Āl ʿImrān 3/10, and the like).
• For Abū Lahab: "his wealth and what he earned did not profit him" (Tabbat 111/2) treated in the discussion of al-Masad — the same root, the same structure.
• And here: for prophet husbands.

Setting these three examples side by side completes the list the Qurʾān builds: wealth does not profit, children do not profit, a prophet husband does not profit.

That is, the verse adds not an exception but the most unexpected item on the list.

The last three words of the sentence are worth pausing over.

وَقِيلَ — a passive verb: "it was said." The speaker is not stated. This is common in the Qurʾān's scenes of the resurrection and its effect is that the verdict is tied not to a person but directly to the situation.

مَعَ الدَّاخِلِينَ — "along with those who enter."

This expression may look superfluous. "Enter the fire" would have sufficed. What does it add?

It adds this: there is no privilege. The two women do not enter the fire with a special standing — neither a lighter nor a heavier one. Along with everyone, an ordinary entry.

And this is the final stress of the whole matter of the verse. From the beginning the verse had been recounting that a position was of no use; the last expression says that the position was of no use in any direction. Being the wife of a prophet neither saves nor brings a separate treatment.

The sentence recorded at al-Baqara 2/134 — "what they earned is theirs, what you earn is yours" — appears here turned into a scene.

Wa ḍaraba'llāhu mathalan li'lladhīna āmanu'mraʾata Firʿawn, idh qālat rabbi'bni lī ʿindaka baytan fi'l-jannati wa najjinī min Firʿawna wa ʿamalihī wa najjinī mina'l-qawmi'ẓ-ẓālimīn

"And God set forth, for those who believe, the example of the wife of Pharaoh. When she said: 'My Lord! Build for me, with You, a house in the garden. And save me from Pharaoh and from his work; and save me from the wrongdoing people.'"

The previous verse recounted the worst outcome in the best house. This verse does exactly the reverse.

Pharaoh is the most conspicuous historical example the Qurʾān uses for tyranny. The analysis recorded at al-Baqara 2/49 was this: the verb yasūmūnakum (root s-w-m: to put a good up for sale, to press insistently) stated that the torment was not a single occurrence but a state continually pressed upon them; and the commercial root of the word made one sense that people were being disposed of like a good haggled over.

There was one further finding there and it serves here: that the male children were killed and the women left is "not a slaughter born of rage but a calculated policy" and "the most dangerous form of oppression is not the angry but the planned."

That is, the woman recounted in this verse lives at the centre of a planned system of oppression. And there she offers a prayer.

The word order of the prayer is the finest detail of the sūra and it is almost always lost in translation.

The Arabic order is this:

رَبِّ — my Lord ابْنِ لِي — build for me عِندَكَ — with You بَيْتًا — a house فِي الْجَنَّةِ — in the garden

Note: عِندَكَ comes before بَيْتًا.

That is, the sentence does not say "build me a house in the garden, let it be with You". It says this: "with You — a house — in the garden."

In Arabic the bringing forward of an element (taqdīm) generally states specification and importance. What is brought forward here is nearness.

The meaning of this is: the woman says first where she wants to be, then what she wants. What is asked for is not a place; it is a position. The house comes second, within that position.

And in third place stands the garden. So the order is: nearness → house → garden. The garden is at the very end of the list — as the address of what is asked for, not as the thing itself.

To see how out of the ordinary this is one must compare. In most of the examples of prayer in the Qurʾān what is asked for is named directly: forgiveness, mercy, patience, offspring, the garden. Here what is asked for is described by way of the place of what is asked for.

The finding treated at the last verse of al-Bayyina — that at the summit of the reward there is placed not an object but a relation, and that "a deed becomes not a payment but an attempt at approach" — is the exact counterpart of this prayer. There the Qurʾān set good pleasure above the description of the garden. Here a servant asks for nearness before the garden.

There is no verbal link between the two texts; but the hierarchy they set up is the same. I record this as an observation.

Root ب ن ي (b-n-y): to build, to raise a structure. Bināʾ (a building), bannāʾ (a builder, a mason), bunyān (a firm structure) are from the same root. The Turkish bina (building) and bünye (constitution) come from here.

The concrete sense of the word is putting stone upon stone — handwork, taking time, done by stages.

And here I record an observation as my own reading; I do not present it as transmitted and I do not load too much upon it:

Pharaoh is mentioned once in the Qurʾān as building a structure: it is recounted that he ordered a high tower to be built for him (al-Qaṣaṣ 28/38; al-Muʾmin 40/36-37). That is, in history his name is that of a ruler mentioned together with vast structures.

And in the same house a woman asks for a structure — but not of stone, and not from himself.

I do not want to force the difference between the two structures; the verse sets up no such comparison and I do not pretend that I do. I record only this: the same family of verbs looks in two different directions within the same household.

A linguistic note: the imperative ibni has come down to a single syllable because the final letter of the root (yāʾ) has dropped — an ordinary conjugation in Arabic's defective verbs. Like qū in the verse at 66/6, this too is a shortened imperative.

The second half of the prayer is as striking as the first.

نَجِّنِي — root ن ج و (n-j-w): to be saved, to slip clear of a danger. Najāt (salvation), nājiya (one who is saved) are from the same root. One of the concrete senses of the root is also "a high place, a hill the flood does not reach" — salvation is staying above the water.

In the tafʿīl form (najjā): he saved. And this form, as was recorded in al-Bayyina, states intensification and repetition.

The verb occurs twice and this repetition is meaningful:

1. Najjinī min Firʿawna wa ʿamalih — "save me from Pharaoh and from his work." 2. Najjinī mina'l-qawmi'ẓ-ẓālimīn — "save me from the wrongdoing people."

And the real matter is in the first request: عَمَلِهِ.

The woman asks to be saved from her husband. And she asks besides to be saved from his work. The two are counted separately.

This distinction is worth pausing over. To move away from a person and to move away from what that person does are not the same thing. A person living inside a system, even if he puts a distance between himself and the owner of that system, goes on living inside what the system produces: he eats at his table, he benefits from his order, he stands in the security he provides.

What the prayer does is to name this second bond separately.

And there is a third layer: al-qawmi'ẓ-ẓālimīn — the wrongdoing people. That is: the person, the work, and the community. Three separate rings:

The prayer counts all three rings one by one. And note the order: it goes from the nearest to the furthest — husband, order, society.

This is the most concentrated expression of a question that has met us many times in al-Baqara: the question of how a person can separate himself from the structure he is inside. And the answer the prayer gives is that the separation does not happen in a single step.

I record one further observation.

The verse does not recount what this woman did. No escape, no revolt, no struggle is recounted. The only thing recounted is the sentence she said.

Idh qālat — "when she said". And after it, the prayer.

That is, the ground for a person's being presented as "an example for those who believe" is a sentence that came out of her mouth. Not an act, but a word. And the word is not a claim either, but a request.

One must think about what this means. A person is in a position where she can change nothing — she is in the house of the most powerful man, she cannot flee, she cannot resist, she cannot bring down the system. The only thing left in her hands is what she wants.

And the verse says that this is enough.

I write this as my own reading; but the structure of the verse supports this reading: the example is built not upon an act but upon a prayer.

Wa Maryama'bnata ʿImrāna'llatī aḥṣanat farjahā fa-nafakhnā fīhi min rūḥinā wa ṣaddaqat bi-kalimāti rabbihā wa kutubihī wa kānat mina'l-qānitīn

"And Mary, the daughter of ʿImrān. She guarded her chastity; and We breathed into it of Our spirit. She confirmed the words of her Lord and His books, and she was of those devoutly obedient."

The first three examples were built upon an axis of marriage: two women with good husbands, one woman with a bad husband.

The fourth example abandons this axis entirely. Mary has no husband. And the verse mentions her not by reference to a man but by reference to her father: ibnata ʿImrān — the daughter of ʿImrān.

This is important in point of the structure of the example. Three examples tested the "husband factor": a good husband does not save, a bad husband does not prevent. The fourth example removes the factor altogether and shows what remains.

That is, the structure proceeds in three steps:

1. Good husband + bad choice = loss (66/10) 2. Bad husband + good choice = gain (66/11) 3. No husband + good choice = gain (66/12)

The third step makes certain the conclusion that follows from the second: the variable has nothing whatever to do with the husband.

Root ح ص ن (ḥ-ṣ-n). The concrete sense of the root is a fortress: ḥiṣn — a place surrounded by walls, which cannot be taken. And all the derivatives carry this picture:

• حِصْن (ḥiṣn) — a fortress, a fortified place.
• تَحَصَّنَ (taḥaṣṣana) — he withdrew into the fortress, he became sheltered.
• حَصِين (ḥaṣīn) — sound, well protected.
• مُحْصَنَة (muḥṣana) — a woman protected, taken into shelter.
• And for a horse: فَرَسٌ حِصَان — ḥiṣān, a war horse. A horse that protects its owner and cannot be captured.

The finding that comes out of this is important: Arabic names chastity not with a word of deprivation but with a word of fortification.

In English and in many languages chastity is described by way of "not doing" — keeping away from something. The Arabic word turns this round: chastity is drawing a wall around oneself. That is, not a subtraction but a building.

And note: the verb is active and its subject is Mary: aḥṣanat — "she guarded". Not passive. That is, she was not guarded by someone else; she guarded herself.

This distinction becomes clear when compared with the other uses of the word in the Qurʾān. The same root is used passively too, in the form muḥṣanāt. Here the active form has been chosen.

A note: the same root recounts a fortress in this verse, while in the previous verse the word for house (bayt) recounted a structure. The last two verses of the sūra close with two different words for structure: one a house asked for, the other a fortress already built. I record this as an observation, not as a claim.

فَنَفَخْنَا فِيهِ مِن رُّوحِنَا — "We breathed into it of Our spirit."

This expression is used in the Qurʾān for the creation of Adam as well (nafakhtu fīhi min rūḥī — al-Ḥijr 15/29; Ṣād 38/72). The same verb, the same construct.

Note the particle مِن: min rūḥinā — "of Our spirit". This states partitiveness; and at al-Baqara 2/3, in connection with mimmā razaqnāhum, the function of this particle was recorded. Here too what is recounted is a taking of a share from a whole, not a transfer.

The wide disputes about Mary and Jesus (ʿĪsā) are not the subject of this sūra and I do not carry them here; the verse passes over this matter in a single sentence and I pass over it too.

Root ص د ق (ṣ-d-q): to be true, the word's matching reality. Ṣidq (truthfulness), ṣadīq (a friend — one whose word is trustworthy), ṣadaqa (what is given; the act that confirms the giver's claim), ṣiddīq (one who confirms much) are from the same root.

In the tafʿīl form (ṣaddaqa): he confirmed, he counted true. And this is the exact opposite of the verb kadhdhaba (he counted false) treated in al-Māʿūn — the same pattern, the opposite root.

That is, the verse uses the opposite of the verb at the opening of Sūrat al-Māʿūn: there yukadhdhibu bi'd-dīn (the one who counts the religion false), here ṣaddaqat bi-kalimāti rabbihā (she who counted true the words of her Lord). Both take their object with the particle bi.

بِكَلِمَاتِ رَبِّهَا وَكُتُبِهِ — "the words of her Lord and His books."

Two things at once: words and books. The two are counted separately.

Kalimāt — the words of God; in the Qurʾān it is used in the sense of revelation, of ruling, and of the decree that an event should come about. Kutub — the books sent down. At al-Bayyina 98/3 the root ك ت ب was treated in connection with kutubun qayyima; I do not repeat it. In one sentence: the concrete sense of the root is to join, to sew, and "to write is to bind."

A note on variant readings: it is recorded in the sources that this word is read both as kutubihī (plural) and as kitābihī (singular). The difference between the two readings is whether the confirming is directed at all the books or at a particular book. Because I am not sure to which imam of the variant readings it belongs I write no name, and I make no preference between the two readings; both are meaningful.

The sūra closes with a grammatical surprise.

قَانِتِين — masculine plural. The expected form would have been qānitāt: the feminine plural, just as it occurred in the fifth verse (qānitātin).

That is, the same sūra used the same word in the feminine seven verses earlier; at the end it uses the masculine. This is data enough to draw attention.

Root ق ن ت (q-n-t): the sense of the root is continuous and silent obedience; it also carries the sense of standing long in the presence. Qunūt — standing in prayer and supplicating. At the centre of the word there is not a noise or a display but continuity and silence.

The classical explanations for the grammatical question:

All three have been defended in the classical sources. I do not impose a preference.

But there is an observation that stands independently of the explanations and I record it.

The first explanation — that it states a genus — is the normal working of Arabic grammar and is probably correct. However, in the fifth verse of the sūra the feminine form of the same word had been used; that is, the text knows how to use the feminine form and does use it. That the masculine form comes at the end is not something grammar makes necessary but the choosing of one of two possible forms.

And the consequence of the form chosen is this: Mary is entered into a category not defined by sex. The last word of the sūra, in recounting a woman, positions her not "among the women" but "among those who obey".

I write this not as a ruling but as a record of a work the text does.

And a support from within the Qurʾān: in the address directed to Mary in Sūrat Āl ʿImrān there is a similar structure: "O Mary! Be devoutly obedient to your Lord (uqnutī), prostrate, and bow with those who bow" (Āl ʿImrān 3/43). The maʿa'r-rākiʿīn at the end of the sentence is masculine plural too — "with those who bow", not the feminine form.

That is, about the same person, around the same family of verbs, the same grammatical choice has been made in two different sūras. This repetition takes the choice out of being accidental. I impose no conclusion; I record the data.

We have read the three verses one by one. Now they must be read together, because what the structure says is more than what the verses say one by one.

Let us set up the table once more, this time separating out the variables:

There are two variables: the husband's condition and the woman's choice. The outcome agrees exactly with the second and not at all with the first.

It is set up like the design of an experiment: the first variable takes three different values across four examples (prophet, prophet, tyrant, none), the second variable takes two values, and the outcome follows only the second variable.

That is, the examples have been chosen not by way of sex but by way of choice. And that the selection is this clean shows that the structure is deliberate.

There is one further detail and it deserves to be recorded in a calm language.

The Qurʾān gives two examples here: one for those who deny, one for those who believe. The example in both is a woman.

And the addressees are mixed: the sixth verse of the sūra had called out yā ayyuha'lladhīna āmanū — to all the believers. That is, while addressing all believers, women and men alike, the verse shows women as both the negative and the positive example.

I do not turn this into material for a modern polemic; the verse conducts no such debate. I record only what stands in the text: the position of being an example is, in both directions, not limited by sex. The male reader too is expected to think by way of these examples, because the mathal is given to him.

And the last word of the sūra — al-qānitīn, masculine plural — agrees with this reading: the category is not sexed in any case.

This is the strongest contribution the sūra makes to the corpus, and it is the third leg of two findings established earlier.

In al-Baqara, in two separate places, two different axes of the same question were treated:

The first axis — lineage. al-Baqara 2/124: Abraham asks that the leadership continue through his offspring as well. The answer: "My covenant does not reach the wrongdoers." The finding recorded there was this: "Lineage does not produce authority… And this was said to a prophet who was asking. That is, the rule was bent for no one."

The second axis — belonging. al-Baqara 2/134 (and 2/141): after a chain of lineage is counted out, it comes down: "They were a community that has passed away. What they earned is theirs, and what you earn is yours." The finding recorded there: "forebears leave an inheritance, they do not leave a record", and "no lineage here enters the reckoning, neither for nor against."

The third axis — marriage. al-Taḥrīm 66/10-12: and this sūra adds the axis al-Baqara did not treat.

Putting the three together makes the principle the Qurʾān sets up fully visible:

And all three build the same sentence: no bond transfers the reckoning.

This triad is one of the cleanest cross-findings in the corpus. For three separate sūras, in three separate contexts, take up three separate kinds of bond and all three arrive at the same conclusion. And each time the strongest example is chosen: for lineage the line of Abraham, for belonging the chain of the prophets, for marriage the houses of two prophets.

That is, the principle is tested not with marginal cases but with the very places where, if it were a rule, one would most expect it to hold.

And al-Taḥrīm adds one thing more: the reverse is also true.

The two verses in al-Baqara said that the bond does not work in one's favour. al-Taḥrīm shows at the same time that it does not work against one either: being in Pharaoh's house does not prevent a woman's salvation.

Setting up this symmetry is important, because read in one direction only the principle turns into a sentence of hopelessness. What the sūra does is this:

A good bond does not save; a bad bond does not condemn either.

And the second half of the sentence recorded at al-Baqara 2/134 was already saying this: "you will not be asked about what they did" — that is, the fault of one's forebears is not loaded onto anyone's back either.

The sixth verse said "guard yourselves and your families". The tenth verse shows that two prophets could not accomplish this.

I left this tension open in the discussion of 66/6. Here, in the light of the table of the four women, one thing more may be added — and I write it as my own reading:

The sūra sets up two things at once. On one side responsibility: there is a work to be done towards one's household and that work is commanded. On the other side a limit: the result of that work is not guaranteed, because the other party is a subject too and makes his own choice.

The two do not cancel one another. On the contrary, without one the other is spoiled:

• Without responsibility, the limit turns into indifference.
• Without the limit, responsibility turns into a pressure that counts another's life as one's own reckoning.

The sūra sets the two side by side six verses apart and builds no reconciling sentence between them. It appears that the reader is expected to carry both at once.

The verbal texture of the sūra must be gathered up. A few roots circulate within the text and their circulations are meaningful.

The root recounts in one verse the coming out of knowledge, in the next two kinds of backing. The common image is the back: knowledge passes from the other side of the back to the front, while support comes about by giving back to back.

The root occurs eight times in the sūra and four of them in a single verse:

The one the sūra addresses is called by a name from this root (al-nabiyy), and the third verse of the sūra follows the circulation of a piece of news by using the same root four times. The finding recorded in al-Nabaʾ — that a nabaʾ is news upon hearing which something must change — is here tested in a domestic scene.

In the sixth verse قُوا (guard, و ق ي) and وَقُودُهَا (its fuel, و ق د) stand side by side. The two roots are different. I made this distinction plainly above and I repeat it here, because to record the distinction in places where confusion is likely is the method of this text.

The same root, in one place a quality, in another an instruction.

The root occurs three times in the sūra and in three different forms: offer → quality → command. And the addressee widens each time.

The sūra calls out five times with يَا أَيُّهَا:

Five calls, and the arrangement is symmetrical: the Prophet → the believers → those who deny → the believers → the Prophet. The middle of the sūra (verse 7) stands alone, with the calls to the believers on either side of it and the calls to the Prophet outermost.

I record this symmetry as an observation. And I add this: the verse standing at the centre of the symmetry is the shortest and the harshest verse of the sūra — the verse in which the excuse is refused.

What the sūra follows from beginning to end is a place.

The sūra opens inside one house, spreads to all houses, goes to three houses in history, and at the very end leaves the word "house" and closes with a word of fortification.

One. Falling silent where the text conceals.

The most transferable lesson of this sūra lies less in its content than in its method.

The Qurʾān chooses not to record all the defining details of an incident it has in hand. It does not give the names, it does not give the subject, it does not give the secret. And in the third verse it recounts the same behaviour as a virtue: ʿarrafa baʿḍahū wa aʿraḍa ʿan baʿḍ.

Today its counterpart is direct. Relaying everything known about an incident is most often done under the name of "honesty". The measure the sūra shows is different: what is needed for a matter to be understood and what curiosity wants are separate things. The first is recounted, the second is not.

And the measure works independently of the intention of the one recounting. The person who relays the detail may not be ill-intentioned; but the detail relayed lives its own life from the moment it is relayed.

Two. The position of the one who holds information.

The three options treated in the third verse — to say all of it, to say none of it, to say a part of it — describe a situation met with every day today.

The measure was: make known that it is known, do not open the detail. This prevents both putting the other party into a defensive position and holding information as a power.

Three. Extra prohibition.

What the first verse corrects — closing off to oneself what is lawful — takes aim at a form of piety: the piety measured by the number of things not done.

And the form of the correction is important: no prohibition is laid down, a question is asked; and the sentence closes with forgiveness. That is, to the person who takes on extra weight, not a punishment but a measure is recalled.

The picture that emerges when read together with al-Muzzammil 73/20 is this: in correcting a person who loads too much upon himself, the Qurʾān always uses the same method — it relaxes the measure, and presents this not as a concession but as a direction.

Four. Responsibility and limit standing together in the family.

The tension between the sixth verse and the tenth touches the most real question of parenthood and of close relationships.

The errors at the two extremes are known:

• Indifference: doing nothing at all by saying "everyone is responsible for himself".
• Appropriation: counting a close one's choice as one's own reckoning and putting pressure on him.

The sūra closes off both. And the way it closes them is in the word itself: wiqāya is holding a shield, not control. One who holds a shield does not take the place of the one guarded.

Its practical measure may be this: what can be done is to arrange the thing the other party will see — one's own behaviour. The decision remains his.

Five. That repentance is a work of sewing.

The second sense of naṣūḥ — to sew, to mend — takes repentance out of being an emotional state.

And it says two things at once: a mark remains, but wholeness comes back. This prevents both behaving "as if nothing had happened" and saying "nothing can be done now".

What the sewing requires is also definite: knowing where the tear is. A general remorse does not close a particular tear.

Six. The psychology of looking for a guarantee.

What the last three verses of the sūra say takes aim at one of the commonest tendencies of religious life: looking for security by way of a bond.

The bond may be lineage, it may be marriage, it may be a congregation, it may be attachment to a teacher or to a community. The common structure is the same: instead of carrying his own reckoning directly, a person hopes to carry it by attaching himself to someone he trusts.

The attraction of this tendency is understandable. To carry one's own reckoning is a lonely business; to hold on to a bond gives the feeling of sharing the load.

The sūra removes this feeling — and in removing it, it uses the strongest example. No bond stronger than living in a prophet's house can be imagined. The verse says that even that is no guarantee.

But the same structure says the reverse too — and this is the second half of the principle: a bad bond does not condemn either. A woman in Pharaoh's house, while continuing to remain in that house, builds a sentence and is remembered by that sentence.

What emerges when the two are read together is not a hopelessness but an equalising: no one has won or lost in advance because of where he was born or where he stands.

Seven. That the narrowest possibility is counted enough.

Of the woman recounted in the eleventh verse the sūra records a single thing: the sentence she said. No act, no escape, no struggle is recounted.

The side of this that faces the present has to do with how a person assesses his own possibility. A person who cannot change a situation most often thinks he has nothing at all in his hands — and the moment he thinks it, he lets go of the one thing he does have.

What the verse records is that the smallest thing left in hand — what one wants — is worth recording.

Per the method (Method), I list plainly the places where, throughout the text, a definite language was deliberately not used and a note of caution was entered:

• The occasion of revelation. The existence of the reports was recorded, their content was not. No name was given, the incident was not recounted, no preference was adopted as to which branch is the sounder. The ground was explained within the text in three items; the most important is that the Qurʾān itself did not give these details.
• The two persons meant at 66/4. Names are given in the reports; no name was written here. The verse itself gives no name.
• The finding "the text applies to the reader the behaviour it describes" (the coincidence between ʿarrafa baʿḍahū wa aʿraḍa ʿan baʿḍ at 66/3 and the sūra's own attitude) is my own reading, it is not a classical transmission and it is not binding.
• The distinction يَا أَيُّهَا النَّبِيّ / يَا أَيُّهَا الرَّسُول. It was presented as an observable tendency, not as a rule; it was stated that there are crossing usages.
• That the word مَرْضَاة has been chosen as a lofty word and the reading of criticism drawn from this are my own reading; they are not transmitted.
• Which branch of meaning the verb فَرَضَ works in at 66/2. Two branches were given and the first branch (determination) was preferred on the ground that the second (obligation) makes the sentence meaningless. This is a preference; it is not binding.
• The expiation for an oath and its legal details. al-Māʾida 5/89 was pointed to; no ruling was issued. Whether making something forbidden to oneself counts legally as an oath, and the order and details of the expiation, are the subject of the literature of law and were not settled here.
• The link between the two branches of the root س ر ر (secrecy and joy). It was relayed as an explanation given by the lexicographers, without a claim of certainty.
• Where the answer to the condition is at 66/4. Three explanations were given, no preference was adopted.
• Whom the expression صَالِحُ الْمُؤْمِنِينَ means. Two explanations were given; it was stated that the first requires fewer assumptions, but no preference was imposed. The names occurring in the reports were not written, because there are differences among them.
• That the word ظَهِير comes in the singular. Three explanations were given, no preference was adopted. The beautiful side of the third explanation in point of meaning was noted separately with the note "my own reading".
• The meaning of the word سَائِحَات. Three views were given in a table; no preference was adopted. It was recorded that the first view is the most widespread in the sources, but that widespread currency is not linguistic proof.
• The mentioning of stones as fuel at 66/6. The classical explanations were relayed; it was stated that there is no evidence for a firm preference.
• No link of root was drawn between قُوا and وَقُودُهَا and it was recorded plainly that the two roots are different (و ق ي / و ق د). Nothing certain was said as to whether the two words being in the same verse is a literary choice.
• The circulation of the root غ ل ظ between 66/6 and 66/9 was recorded as an observation; no conclusion of the sort "man ought to be like an angel" was drawn from it.
• The legal framework, conditions and method of jihād. The scope of the word was recorded; no ruling was issued and no current political disputes were entered upon. It was written plainly that the verse states no instrument.
• The place of 66/9 within the sūra. Three readings were given; it was stated that the third is the most functional in point of word order, but it was not presented as a preference.
• To whom the adjective نَصُوح belongs (to the repentance or to the repenter). Four explanations were given, no preference was imposed. The beautiful side of the fourth explanation in point of meaning was recorded as a "possibility of reading".
• The link between the two branches of the root ن ص ح (purity and sewing). Since the two branches are given under separate entries in the classical dictionaries and no necessary link is drawn between them, the joining was presented not as an etymological claim but as an observation.
• The explanations concerning the prayer for the completion of the light at 66/8. They were relayed with the note that they rest on reports and that there are differences among them; no preference was adopted.
• The difference of usage between امْرَأَة and زَوْج. It was relayed as a distinction treated in the classical commentaries; it was not imposed as a rule and it was recorded plainly that the distinction does not work everywhere.
• The nature of the betrayal at 66/10. It was recorded that in the classical tradition there is a strong agreement that the betrayal is in the religious field, and four grounds were given. This is a preference; but it was stated that it is a preference on which there is near unanimity. That the verse does not say what the betrayal was was also recorded separately.
• The connecting of the verb بَنَىٰ with the structures of Pharaoh was presented plainly as my own reading, it was stated that the verse sets up no such comparison, and no weight was loaded upon it.
• The variant reading كُتُبِهِ / كِتَابِهِ at 66/12. The existence of the difference was recorded; because I am not sure to which imam of the variant readings it belongs, no name was written, and no preference was adopted between the two readings.
• That the word الْقَانِتِين comes in the masculine plural. Three explanations were given in a table, no preference was imposed. It was stated that the first explanation (that it states a genus) is the normal working of grammar and is probably correct. The observation drawn from it — that the category is not sexed — was presented not as a ruling but as a record of a work the text does.
• The tension between 66/6 and 66/10 was deliberately left open. It was partly explained by the distinction between act and result; but it was written plainly that the tension is not fully resolved.
• The sūra's symmetry of address and the circulations of roots under the seven weave headings were presented as observations; they were not attributed to the classical sources.
• The whole of the section "What this says to the present" and the three findings drawn from the table of the four women are my own reading; they are not transmitted and they are not binding. The joining of the triad lineage-belonging-marriage is this commentary's own cross-reading.