Differences Between the Qur'ans

The crime that earns the punishment is different. In Hafs and Shu'bah (yakdhibun) the hypocrites are punished for lying. This picks up their claim two verses earlier: "We believe in God and the Last Day" (2:8). In the other six (yukadhdhibun) they are punished for takdhib, calling the message a lie. That is the charge the Qur'an normally brings against open unbelievers, not against people who outwardly profess belief.

Wa'adna is a one-sided promise made by God. Wa'adna, with the long ā, is form III (muwa'adah): an appointment agreed between two parties. Al-Razi states the problem plainly: the reading without alif is straightforward "because the promise came from God, whereas muwa'adah is a reciprocal form and requires two parties".

Several early authorities preferred wa'adna precisely to keep God out of a bilateral arrangement. Abu 'Ubayd (d. 224/838) chose it and rejected wa'adna, arguing that "mutual appointment comes only from human beings; God alone makes promises and threats". Abu Hatim al-Sijistani (d. 255/869) agreed: "mutual appointment is mostly between created beings who are equals, each promising the other". Makki ibn Abi Talib (d. 437/1045) added that "the plain wording contains a promise from God to Moses, and none from Moses", so the verb should be taken as one-sided. Others defended wa'adna. Al-Nahhas (d. 338/950) called it "better and finer", and Ibn 'Atiyyah replied that Moses' acceptance of God's promise "resembles a mutual appointment". Al-Tabari accepted both readings for the same reason. Abu Hayyan, writing after the canon was fixed, closes the discussion by saying neither may be preferred "because both are mutawatir".

The question was never a matter of style. It was whether the Qur'an may describe God as a party to a bilateral arrangement with a man. Among the King Fahd mushafs, al-Duri and al-Susi carry the reading that avoids it, here and at 7:142.

Sources al-Razi, Mafatih al-Ghayb, on 2:51; al-Qurtubi, al-Jami' li-Ahkam al-Qur'an, on 2:51 (citing Abu 'Ubayd, Makki, Abu Hatim and al-Nahhas); Abu Hayyan, al-Bahr al-Muhit, on 2:51; Ibn 'Atiyyah, al-Muharrar al-Wajiz, on 2:51; al-Tabari, Jami' al-Bayan, on 2:51.

In six mushafs God names Himself as the one who forgives ("We"). In Qalun and Warsh the verb is passive and the forgiver goes unnamed. Qalun and Warsh have the same passive at the parallel passage 7:161, where the other six again have "We will forgive".

This is the Qur'an's key text on abrogation (naskh), and the two readings give it different content. Nunsiha (Hafs, Shu'bah, Qalun, Warsh) is from insa', "to cause to forget". Nansa'ha (al-Duri, al-Susi, al-Bazzi, Qunbul) is from nas', "to defer". Al-Tabari glosses it as "We postpone it, leaving it in force and unchanged".

The difference matters doctrinally. The "forgetting" reading is the natural Qur'anic support for the idea that verses were removed from the Prophet's and the Companions' memory. Al-Tabari, under this very verse, cites Anas ibn Malik's report about the Companions killed at Bi'r Ma'unah: "We recited a Qur'an about them … then it was lifted." Al-Tabari himself prefers nunsiha, but takes it to mean "We leave it", not "We make you forget it". He reports that others rejected the "forgetting" sense outright, on the ground that the Prophet could not have forgotten Qur'an that had not been abrogated. In the al-Duri, al-Susi, al-Bazzi and Qunbul mushafs the verse does not mention forgetting at all.

Sources al-Tabari, Jami' al-Bayan, on 2:106.

In six mushafs the clause is a statement: Muhammad bears no responsibility for those who reject him. In Qalun and Warsh it is a prohibition: "do not ask". Al-Tabari records the reports that the exegetes attached to the prohibition reading. Muhammad ibn Ka'b al-Qurazi and Dawud ibn Abi 'Asim say the Prophet said, "Would that I knew what became of my parents!", and that "Do not ask about the companions of Hellfire" was revealed in response.

Al-Tabari prefers the statement reading. Part of his reason for rejecting the report is not that the Prophet's parents were saved. It is that the Prophet could not have doubted that idolaters, his parents included, belong to Hell ("…the impossibility of the Messenger doubting … that his parents were among them"). The editor of al-Tabari notes that both reports are mursal and weak. Even so, in the mushaf that most of North and West Africa reads, the verse is the one tradition tied to the fate of Muhammad's mother and father.

Sources al-Tabari, Jami' al-Bayan, on 2:119 (reports nos. 1875–1877 and al-Tabari's comment).

With the imperative (ittakhidhu) the clause is a command to the believers. It is the basis of the prayer performed behind the Maqam Ibrahim after circumambulating the Ka'bah. It also underlies the famous report of 'Umar in Sahih al-Bukhari: "I agreed with my Lord in three things. I said, 'Messenger of God, if only we took the Station of Abraham as a place of prayer', and then 'And take the Station of Abraham as a place of prayer' was revealed."

With the past tense of Qalun and Warsh (ittakhadhu) the clause is a narrative about what people did in Abraham's time. It commands nothing, and 'Umar's celebrated "agreement" with a divine command has no command to agree with.

Sources al-Bukhari, al-Sahih, Kitab al-Salat, bab ma ja'a fi al-qiblah (report of Anas from 'Umar).

The meaning is almost the same: wassa (form II, enjoining insistently) against awsa (form IV). It is included because the written text differs. The Qalun and Warsh mushafs carry an extra letter (alif) that the other six do not. Ibn al-Jazari states that this is how it is written "in the codices of the people of Medina and Syria", while the others are written without it. In other words, the Uthmanic codices sent to the different cities did not have the same consonants.

Sources Ibn al-Jazari, al-Nashr fi al-Qira'at al-'Ashr, on 2:132.

Hafs makes the question a direct address to the Jews and Christians, continuing "Do you argue with us about God?" in the previous verse. The other seven turn it into a fresh question about them in the third person.

Al-Tabari did not treat this as a free choice. He wrote: "The correct reading in our view is 'Or do you say', with ta' … and it is not permissible to read it with ya', because it is anomalous (shudhudh) against the reading of the readers." The reading he ruled out is the one printed in the Shu'bah, Qalun, Warsh, al-Duri, al-Susi, al-Bazzi and Qunbul mushafs.

Sources al-Tabari, Jami' al-Bayan, on 2:140; Ibn al-Jazari, al-Nashr, on 2:140.

In Qalun and Warsh the one who is asked to imagine the scene is the Prophet (or the listener). In the other six it is the wrongdoers themselves. Al-Tabari prefers the Qalun–Warsh reading, and his reason is an objection to the other one: "once they have seen the punishment, they are certain that all power belongs to God, so there is no sense in saying 'if they could see'. One only says 'if you could see' to someone who has not seen." In other words, he finds the reading of Hafs and five others semantically weak.

Sources al-Tabari, Jami' al-Bayan, on 2:165.

Qalun and Warsh differ in two words. They make fidya a construct ("a ransom of feeding" rather than "a ransom, namely feeding"), and they make the poor plural. The question is how much is owed for each day not fasted. Ibn 'Ashur states that "there is consensus that what is due is the feeding of one poor person", and he explains the plural as matching the plural of those who are able to fast ("a plural set against a plural"), while the singular "looks to what is due from each individual". Al-Tabari explains the singular as "one poor person for each day he breaks the fast", and cites al-Hasan al-Basri for the plural: "poor people, for the whole month". On the construct, al-Tabari sides with the Qalun–Warsh reading and calls the opposing argument "clearly wrong".

The verse is also one of the classic abrogation cases. Al-Tabari collects reports from Mu'adh ibn Jabal, Ibn 'Umar, 'Alqamah and others that the option to pay instead of fasting was later cancelled by "whoever witnesses the month, let him fast it" (2:185).

Sources Ibn 'Ashur, al-Tahrir wa-l-Tanwir, on 2:184; al-Tabari, Jami' al-Bayan, on 2:184.

Yathurna (seven of the eight) refers to a state: the end of bleeding. Al-Tabari cites Mujahid and 'Ikrimah: "until the blood stops". Yattahharna (Shu'bah) refers to an act: purifying oneself by washing. The verse governs when intercourse becomes lawful again. The majority (Malik, al-Shafi'i and others) require the ritual bath first. The Hanafis permit intercourse in some cases once bleeding stops, even before washing.

Al-Qurtubi reports how the Hanafis defended their position: "the two readings are like two verses, so both must be acted upon". They assign one reading to bleeding that stops at the maximum length of a period (lawful without washing) and the other to bleeding that stops earlier (lawful only after washing). Ibn al-'Arabi rejects this as unworthy of eloquent speech. The legal disagreement is open. Its textual root is a doubled consonant present in one printed mushaf and absent from the other seven.

Sources al-Tabari, Jami' al-Bayan, on 2:222; al-Qurtubi, al-Jami' li-Ahkam al-Qur'an, on 2:222 (ninth issue).

One dot separates nunshizuha (with z: to raise up, set one on another) from nunshiruha (with r: to resurrect). Hafs and Shu'bah describe the reassembly of the skeleton. The other six describe the bones being brought back to life before they are clothed with flesh. The undotted early script could not tell the two apart. Only the later reading tradition decides which one a given mushaf contains.

Hafs speaks of God in the third person ("He"). The other seven have God speak in the first person ("We"). Qalun and Warsh go further with the jussive nukaffir. That makes the expiation grammatically part of the consequence of giving charity secretly. In the indicative of the others it is a separate promise.

Al-Tabari explains the Qalun–Warsh reading as a direct address to the Jews: "you, O Jews, saw the idolaters as twice the Muslims". In the other six it is a statement about the two armies at Badr. He prefers the latter, which he takes to mean that the Muslims saw the unbelievers as twice the Muslims' number.

Either way the verse sits uneasily with 8:44, where God makes each army look few in the other's eyes. Al-Tabari reconciles them by positing successive reductions: God shrank the enemy in the Muslims' eyes once, then again, then a third time. Who did the seeing, and who was seen, depends on the mushaf.

Sources al-Tabari, Jami' al-Bayan, on 3:13.

With wada'at ("she delivered") the clause is God's own aside, interrupting the mother's prayer. With Shu'bah's wada'tu ("I delivered") it is still the mother speaking. The same words are God's comment in seven mushafs and a human being's words in the eighth.

With the doubled kaffalaha (Hafs, Shu'bah) God is the subject and Zechariah the object: God assigned him as Mary's guardian. With kafalaha and the nominative Zakariyya'u (the other six) Zechariah is the subject who took her into his care. The Qur'an's own account (3:44) says the guardians cast lots for her.

Only Hafs has the third person here. The other seven continue the divine first person of the previous verse ("I will punish them", 3:56). The difference is one of voice, not doctrine. It is also the kind of shift the tradition elsewhere treats as meaningful, as the next entries show.

Hafs keeps both verbs in the third person, as a report about the People of the Book. Shu'bah, Qalun, Warsh, al-Bazzi and Qunbul make both a direct rebuke. Al-Duri and al-Susi mix them: "they seek" … "you will be returned".

With ya' (Hafs alone) the promise concerns the upright group among the People of the Book described in the preceding verses (3:113–114). With ta' (the other seven) it is addressed to the believers generally. The group that is promised reward for its good deeds is different.

The context is the defeat at Uhud and the rumour that Muhammad had been killed ("if he dies or is killed, will you turn on your heels?", 3:144). In Hafs and Shu'bah the verse holds up earlier prophets who fought alongside many followers. In the other six it speaks of prophets who were killed, which bears directly on the rumour. It also opens an ambiguity the exegetes discuss: whether the "many devoted men" were killed too, or survived and did not lose heart.

Hafs alone turns from the believers being addressed to a third party: the hypocrites and unbelievers mentioned in the previous verse. In the other seven the comparison is addressed to the believers about their own wealth.

Yaghulla (six mushafs) says a prophet would never embezzle spoils. Al-Tabari, citing Ibn 'Abbas, says it was revealed when a red velvet cloth went missing from the booty of Badr and some said, "Perhaps the Messenger of God took it." Yughalla (Qalun, Warsh) is passive: a prophet is not to be defrauded by his followers, or not to be accused of fraud.

Al-Tabari preserves a Successor's verdict on the passive reading. Sa'id ibn Jubayr was asked, "How do you read this verse: yaghull or yughall?" He answered: "No, rather yaghull. By God, a prophet is indeed defrauded and killed." In other words, he rejected yughall because, read that way, the verse says something false. That reading is now printed in the Qalun and Warsh mushafs.

Sources al-Tabari, Jami' al-Bayan, on 3:161 (reports nos. 8136–8140).

Two verbs, three combinations. In Hafs and Shu'bah both prohibitions are addressed to the Prophet (or the listener). In al-Duri, al-Susi, al-Bazzi and Qunbul both are addressed to the boastful themselves. Qalun and Warsh begin with the boastful and end with the Prophet.

Qalun and Warsh switch to the divine first person ("We") here and in the matching promise of Paradise in the previous verse (4:13). The other six keep the third person throughout.

This is a penal verse, and the trigger for the punishment depends on the vowelling. Al-Tabari explains Shu'bah's ahsanna (active) as "when they become Muslim", and reports that this is how its readers understood it. The passive uhsinna of the other seven was understood by others as "when they are married off". So whether an unmarried slave woman who commits zina is liable to the half-punishment at all turns on a vowel. The jurists divided over exactly this question.

Sources al-Tabari, Jami' al-Bayan, on 4:25.

Takun (feminine) and yakun (masculine) differ only in agreement with mawaddah, "affection". There is no difference in meaning. Hafs, al-Bazzi and Qunbul read takun; the other five read yakun.

The verse sets the standard for sparing a man met on a raid. Al-Tabari reports it was revealed after al-Miqdad killed a herdsman who had said, "I am a Muslim". With al-salam (Qalun, Warsh) the protected act is surrender or submission. With al-salām, with the long ā (the other six), it is the greeting of peace. Al-Tabari judged the first to be the correct reading, the one now printed only in the Qalun and Warsh mushafs. The Hafs mushaf therefore carries the reading he considered second best for a rule about who may be killed.

Sources al-Tabari, Jami' al-Bayan, on 4:94 (report no. 10224 and al-Tabari's verdict).

Active voice with God as the stated agent (Hafs, Shu'bah, Qalun, Warsh) against the passive with no agent (al-Duri, al-Susi, al-Bazzi, Qunbul).

The same alternation, here confined to Hafs and Shu'bah against the other six.

Al-Tabari ties the anna ("because") reading to an event that has already happened: the Quraysh barred the Prophet and his Companions at al-Hudaybiyah, and Surat al-Ma'idah was revealed afterwards. The in ("if") reading speaks of a future, hypothetical barring. He links it to the Quraysh's attempt to bar the Muslims at the conquest of Mecca. So the two printed texts place the verse at different points in the Prophet's career: after the event in one, before a possible event in the other.

Sources al-Tabari, Jami' al-Bayan, on 5:2.

With the accusative arjulakum the feet are joined to the faces and hands, which are washed. With the genitive arjulikum they are joined to the heads, which are wiped. Al-Razi records who read which: "Ibn Kathir, Hamzah, Abu 'Amr and 'Asim in the transmission of Abu Bakr [Shu'bah] read it in the genitive, and Nafi', Ibn 'Amir and 'Asim in the transmission of Hafs read it in the accusative." That is exactly the split in the King Fahd mushafs.

The jurists did not settle the matter. Al-Razi lists the positions. The majority require washing, and explain the genitive away either as a case ending "by proximity" to the neighbouring word (jarr al-jiwar) or as referring to wiping over leather socks. Al-Razi, setting out the case for wiping, calls the "proximity" explanation untenable. Al-Hasan al-Basri and al-Tabari held that "the worshipper may choose between wiping and washing". The Zaydi imam al-Nasir li-l-Haqq required both. Twelver Shi'i law requires wiping, on the plain sense of the genitive reading, which is printed in five of the eight mushafs.

Sources al-Razi, Mafatih al-Ghayb, on 5:6; al-Tabari, Jami' al-Bayan, on 5:6.

With the accusative al-juruha, "wounds" is part of what God prescribed for the Israelites in the Torah. With the nominative al-juruhu (al-Duri, al-Susi, al-Bazzi, Qunbul) it can be read as a new sentence, a direct statement of law rather than a report of earlier scripture. Whether Muslims take the law of wounds from Islamic legislation proper or from "the law of those before us" therefore has a different textual footing in different mushafs.

With the accusative al-kuffara the unbelievers are a second, separate object of the prohibition: no alliance with them at all. With al-Duri and al-Susi's genitive al-kuffari they become a sub-group of "those who take your religion in mockery". The prohibition then covers only unbelievers who mock. Al-Tabari gives the genitive reading, says it matches the codex of Ubayy ibn Ka'b ("…and from the unbelievers"), and explains it as "from those given the Book … and from the unbelievers". A general rule on alliances is narrowed or broadened by one case ending.

Sources al-Tabari, Jami' al-Bayan, on 5:57.

In Shu'bah's active yasrif the subject is God: "whomever He averts it from". In the passive yusraf of the other seven the punishment "is averted" with no agent named.

Hafs alone puts the two verbs in the subjunctive. That places them inside the wish: they wish they could return and not deny. The other seven put them in the indicative, which starts a separate declaration: we will not deny and we will be believers.

The difference matters because of the next verse: "…and indeed they are liars" (6:28). Al-Razi spells out the problem. If the words are part of a wish, how can God call them liars, when "a person making a wish is not described as lying"? Read with the indicative, the words are a promise and the charge of lying fits. Read with Hafs, al-Razi has to argue that "a wish is, in effect, a promise".

Sources al-Razi, Mafatih al-Ghayb, on 6:27.

The Hafs reading, "they do not call you a liar", conflicts with the Qur'an's own record of the Meccans doing exactly that ("…a sorcerer, a liar", 38:4; compare 6:66 and the repeated kadhdhabu). The Qalun–Warsh reading, yukdhibunaka (form IV), means "they do not find you to be lying" or "they cannot prove you a liar". That avoids the conflict. Al-Tabari records that the lighter form was read by some Kufans. In the King Fahd mushafs it survives only in Qalun and Warsh.

Sources al-Tabari, Jami' al-Bayan, on 6:33.

In Shu'bah (masculine verb) and in Hafs, al-Duri, al-Susi, al-Bazzi and Qunbul (feminine verb) the way of the criminals is the subject: it "becomes clear". In Qalun and Warsh, sabila is accusative and the verb is second person: the Prophet is the one who discerns it.

A difference of a single dot in the later script: yaqussu (with sad, "He narrates") against yaqdi (with dad, "He decides"). Al-Duri and al-Susi make God's activity juridical, which fits "judgement belongs only to God" and "best of judges". The other six make it narrative.

Hafs and Shu'bah put the sailors' vow in the third person ("if He saves us"), as words spoken about God. The other six put it in the second person ("if You save us"), as a prayer addressed to God. Shu'bah also reads khifyatan for khufyatan ("in secret"), two forms of the same word.

All three verbs change, not only the first. In Hafs, Shu'bah, Qalun and Warsh the Jews are addressed directly and accused of concealing parts of the Torah. In the other four they are spoken about. Since the verse opens with a reply to those who said "God has sent down nothing to any human being", the exegetes disputed who is meant: the Jews, or the Meccan idolaters.

In Shu'bah the Book itself is the warner. In the other seven the Prophet is.

The verse records what Muhammad's opponents would say about the Qur'an. Darasta: "you have studied [it]". Dārasta, with the long ā (form III): "you have studied it with someone", that is, read and learned from the People of the Book. Al-Tabari attributes dārasta to Ibn 'Abbas (variously reported), other early authorities and some Basrans, and glosses it "you read with them and learned from the People of the Book". He himself prefers darasta, which he also takes to mean "you learned from the People of the Book", citing 16:103 ("it is only a human being who teaches him"). Half of the eight mushafs put the accusation of collaborative study into the Qur'an's own wording.

Sources al-Tabari, Jami' al-Bayan, on 6:105.

With the kasra of al-Duri, al-Susi, al-Bazzi and Qunbul (innaha) the final clause is a fresh, emphatic statement: they will not believe. With the fatha of the other four (annaha) it becomes the object of "what will make you perceive". Taken at face value, that asks "what makes you think they will not believe?" and so implies that they would. That is the opposite of the point.

Al-Tabari records how the problem was handled. Some who read annaha took it to mean la'allaha, "perhaps it", citing the codex of Ubayy ibn Ka'b and an Arab idiom. Others treated the negative la as redundant, as in 7:12. The kasra reading needs neither device.

Sources al-Tabari, Jami' al-Bayan, on 6:109.

Bushran (with b: heralds of good news) against nushuran (with n: spread out, dispersing). The difference is again a matter of dots on one letter. Al-Bazzi and Qunbul also make "winds" singular. Since the exegetes often associate the plural riyāḥ with mercy and the singular rīḥ with punishment, the change is not trivial.

A direct rebuke to the Israelites (Hafs, Qalun, Warsh) against a description of them (the other five).

In the scene where God draws Adam's descendants from their loins and makes them testify, al-Duri and al-Susi report the excuse in the third person. The other six address it directly to the audience. Both verbs change, in 7:172 and 7:173. In 7:172 the same narrators, with Qalun and Warsh, also read dhurriyyatihim ("their offspring", plural) for dhurriyyatahum.

The passage (7:189–190) speaks of the first couple: "He created you from a single soul and made from it its mate". Al-Tabari reports the exegetes' disagreement over the "partners" this couple assigned, beginning with the view that it was "partnership in the name". This refers to the widely transmitted story that Adam and Eve named their child 'Abd al-Harith at Iblis's suggestion. In Shu'bah, Qalun and Warsh the noun is shirkan: the couple's act is named with the very word for the one unforgivable sin. Later theologians who held that prophets are protected from sin reread the verse as being about their descendants. The text they were rereading differs by mushaf.

Sources al-Tabari, Jami' al-Bayan, on 7:190.

With ya' the warning is addressed to the unbelievers about themselves. With ta' it is addressed to the Prophet about them. Among the eight, only Hafs reads ya' here.

The variants here are small. Yakun and takun differ only in the grammatical gender of the verb agreeing with mi'ah, "hundred". Da'fan and du'fan are two forms of the word "weakness". In 8:65 the split is different: Hafs, Shu'bah, al-Duri and al-Susi read yakun, and Qalun, Warsh, al-Bazzi and Qunbul read takun.

The theological dispute attached to this verse comes from wording that all eight mushafs share: "Now God has lightened … and He knows that there is weakness in you." Al-Razi reports that "Hisham" argued from it that "God does not know particulars until they occur": "the meaning is, 'Now God knows that there is weakness in you', which requires that His knowledge of their weakness came about only at that time". The mainstream reply is that God knew beforehand that the weakness would occur and knows it, when it occurs, as having occurred. Al-Razi does not say which Hisham he means.

Sources al-Razi, Mafatih al-Ghayb, on 8:66 (second issue).

With the passive yudallu (Hafs alone among the eight) the unbelievers are acted upon. Al-Razi glosses it: "their leaders lead them astray by imposing this postponement on them". With the intransitive yadillu (the other seven) they go astray themselves. Al-Razi calls this "the reading of the majority, and it is good because it attributes the straying to the unbelievers". Whether the verse places the cause of straying in the unbelievers or outside them depends on the mushaf.

Sources al-Razi, Mafatih al-Ghayb, on 9:37; Abu Hayyan, al-Bahr al-Muhit, on 9:37.

Hafs and Shu'bah name God as the one who pardons and punishes. The other six put both verbs in the passive and leave the agent unnamed, which also makes the second ta'ifah nominative.

Al-Bazzi and Qunbul have a whole word, min, that the other six lack. This is a consonantal difference that matches the codex sent to Mecca. The image changes slightly: rivers flowing beneath the gardens, or springing from beneath them.

There is no change in meaning, but the consonantal text differs: the Qalun and Warsh mushafs lack the conjunction wa. The same thing happens elsewhere in these surahs. At 3:133 Qalun and Warsh read sari'u without wa, and at 5:53 Qalun, Warsh, al-Bazzi and Qunbul read yaqulu without it. At 5:54 Qalun and Warsh read yartadid, with two dals, where the others read yartadda.

Hafs's taqatta'a (intransitive, "break apart") describes something happening in their hearts, through death or remorse. The passive tuqatta'a of the other seven implies an agent who cuts them, through death or punishment.

Al-Razi explains that the subject of kada may be "hearts" itself, or an implied "pronoun of the matter" (damir al-sha'n) that the following clause then explains. The masculine yazighu of Hafs fits the second analysis more naturally. The meaning of the clause hardly changes.

Sources al-Razi, Mafatih al-Ghayb, on 9:117.

In one group the unbelievers accuse the man (sāḥir, a sorcerer). In the other they accuse the message (siḥr, sorcery). The Uthmanic consonants allow both, because the long a is not written. The Hafs mushaf adds it as a small superscript alif.

A change of person (He / We), not of voice: both verbs are active.

The same split recurs at 10:96 ("those against whom the word / words of your Lord has come true will not believe"). The singular points to a single decree of damnation. The plural can be read as a plurality of decrees or warnings. In both verses the decree concerns people's failure to believe, and the verses are a standard text in debates over predestination.

Another change of person (He / We), not of voice. Hafs alone has the third person, as at 6:128.

Liyudillu (Hafs, Shu'bah) is transitive: Pharaoh's wealth was given so that they would mislead others. Liyadillu (the other six) is intransitive: so that they themselves would stray. On either reading the particle li- ("so that") makes God's gift serve their misguidance, and al-Razi records how the verse divided the theologians.

The Ash'aris ("our companions", in al-Razi's words) "used this verse as proof that God misleads people and wills their misguidance", taking the li- as the lam of purpose. The Mu'tazilis held that willing unbelief would be evil, so the words had to be reinterpreted. They offered several readings. The li- is the lam of outcome (lam al-'aqibah): this is what it led to, not why it was given. The phrase is an implied question. The li- is a lam of supplication: "let them stray". Al-Razi rejects these as "strained and very weak", observing that Moses did not know outcomes. The transitive Kufan reading sharpens the problem, because God's gift then serves the misleading of others.

Sources al-Razi, Mafatih al-Ghayb, on 10:88.

Hafs alone reads 'ummiyat (form II passive): someone has made the mercy obscure to them. The other seven read 'amiyat (form I): it was simply hidden from them, or they were blind to it. Al-Tabari prefers Hafs's reading for a theological reason: the same verse attributes the mercy to God, "so likewise the obscuring of it from the others" is God's doing. In the Hafs mushaf the people's blindness is something done to them; in the other seven it is a state they are in.

Sources al-Tabari, Jami' al-Bayan, on 11:28.

Hafs's su'idu is passive: their blessedness is conferred on them. The other seven read sa'idu, an ordinary intransitive verb. The verse belongs to the passage dividing mankind into "wretched and happy" (11:105), a standard text in debates on predestination. Al-Tabari treats both as correct.

Sources al-Tabari, Jami' al-Bayan, on 11:108.

Four mushafs, four sentences. In Hafs, Shu'bah, Qalun and Warsh the brothers propose an outing for Joseph ("he"); in al-Duri, al-Susi, al-Bazzi and Qunbul they propose it for themselves ("we"). Independently, Qalun, Warsh, al-Bazzi and Qunbul read yarta'i / narta'i, from ra'a, "to graze, to tend", where the others read yarta' / narta', from rata'a, "to enjoy oneself freely". Who is playing, and whether they are playing or working, differ by mushaf.

In six mushafs Joseph settles wherever he himself chooses. In al-Bazzi and Qunbul he settles wherever God chooses. The agent of the choice is a man in one text and God in the other.

Al-Tabari reports the Kufan reading liyudillu, "so that they mislead people from God's way", and the Basran liyadillu, "so that they themselves stray". The same split between Hafs, Shu'bah, Qalun and Warsh on one side and al-Duri, al-Susi, al-Bazzi and Qunbul on the other recurs at 22:9 and 39:8. It is the variant of 10:88, on which al-Razi records the Ash'ari–Mu'tazili dispute.

Sources al-Tabari, Jami' al-Bayan, on 14:30.

Verses 23 to 37 list commands (worship God alone, honour parents, give relatives their due, keep covenants) and prohibitions. With Hafs and Shu'bah's sayyi'uhu, only the evil part of "all that" is hateful. With sayyi'atan, read by the other six, the whole list, commands included, is called "an evil, hateful to your Lord". Al-Tabari prefers sayyi'uhu for exactly this reason: the passage contains "things forbidden and things commanded".

Sources al-Tabari, Jami' al-Bayan, on 17:38.

Hami'ah (seven mushafs) is "full of black mud"; Shu'bah's hamiyah is "hot". Al-Tabari records that the Companions disputed it: "Mu'awiyah recited this verse and said 'a hot spring', and Ibn 'Abbas said: it is a muddy spring." Al-Tabari accepts both, reasoning that the sun could set in a spring that is both hot and muddy.

Sources al-Tabari, Jami' al-Bayan, on 18:86.

After inna, Arabic grammar requires the accusative hadhayni. Shu'bah, Qalun and Warsh nevertheless read inna hadhani. Al-Duri and al-Susi read the grammatical inna hadhayni, which adds a letter (ya') the consonantal text does not have. Hafs, al-Bazzi and Qunbul lighten inna to in, which removes the problem.

Al-Razi collects the reports on this verse. 'A'ishah, asked about it and two similar places (4:162, 5:69), said: "Nephew, this is an error of the scribe." 'Uthman is reported to have looked at the codex and said: "I see in it a solecism, and the Arabs will put it right with their tongues." Abu 'Amr himself said: "I am ashamed to read inna hadhani." Al-Tabari chooses inna hadhani, the reading of three of the eight, because it follows "the script of the codex". The four printed forms are four different answers to a sentence the tradition itself regarded as a problem.

Sources al-Razi, Mafatih al-Ghayb, on 20:63; al-Tabari, Jami' al-Bayan, on 20:63.

Tahjuruna (six mushafs) is "you abandon [the Qur'an]" or "you talk nonsense". Qalun and Warsh read tuhjiruna, from ihjar, "to speak obscenely", which the exegetes took as the Meccans' abuse of the Prophet. Al-Tabari did not accept the second reading: "The more correct of the two readings is the one the readers of the cities follow … because of the consensus of the authoritative readers."

Sources al-Tabari, Jami' al-Bayan, on 23:67.

The scene is the Judgement: the false gods disown their worshippers. In Hafs the idolaters themselves are told "you" cannot avert the punishment. In the other seven it is the false gods who "cannot" avert it or help. The powerless party is a different set of beings.

Hud's people answer him. With khuluq (Hafs, Shu'bah, Qalun, Warsh) they defend their way of life: "this is the custom of our forefathers". With khalq (the other four) they attack his message: "this is nothing but the invention of the ancients". Al-Tabari prefers khuluq, citing Ibn 'Abbas.

Sources al-Tabari, Jami' al-Bayan, on 26:137.

In six mushafs the limitation lies with the Prophet: he cannot make them hear. In al-Bazzi and Qunbul it lies with the hearers: the deaf do not hear. The same split recurs, with the same narrators, at 30:52.

With anna (Hafs, Shu'bah) the last clause is the content of the Beast's speech: it will tell them that people did not believe. With inna (the other six) the Beast simply "speaks to them", and the final sentence is a fresh statement in God's voice, or the Beast's words quoted directly.

Al-Tabari judges the address form (ta') correct, because if the verse spoke about the destroyed nations it would have to say "what they used to call upon", whereas the Meccans being addressed were still idolaters. The reading he argues against is the one printed in the Hafs, Shu'bah, al-Duri and al-Susi mushafs.

Sources al-Tabari, Jami' al-Bayan, on 29:42.

Al-Tabari explains why Abu 'Amr read yuda''af (al-Duri, al-Susi). Abu 'Amr held that yuda'af, the reading of the others, means "to make a thing three times its size", so that the wives would be punished threefold. He chose yuda''af so that the punishment is only twice that of other women. Al-Tabari rejects Abu 'Amr's interpretation, saying no scholar held it except him and Abu 'Ubaydah. Al-Bazzi and Qunbul make God the explicit agent ("We will double"). The size of a divinely prescribed punishment was argued from the form of a verb.

Sources al-Tabari, Jami' al-Bayan, on 33:30.

The consonantal text differs: Shu'bah's mushaf has no final ha' ("it"), matching the Kufan codex; the others have it. With 'amilathu, ma is naturally relative: people eat the fruit and what they make from it. Without the pronoun, ma can equally be read as a negative: "their hands did not make it", so that the crops are God's work alone. Al-Tabari reads ma as relative. The verse can thus speak for or against human agency in what the earth produces, depending on the mushaf.

Sources al-Tabari, Jami' al-Bayan, on 36:35.

In six mushafs the Qur'an is the warner; in Qalun and Warsh the Prophet is addressed as the warner. At 46:12 the same split falls differently: Qalun, Warsh and al-Bazzi read "that you may warn", while Qunbul reads "that it may warn" with the rest. The two transmitters of Ibn Kathir part company here.

Two independent variants combine into four texts. Aw an ("or that") against wa-an ("and that") is a consonantal difference between the Kufan codex and the others. Yuzhira l-fasada ("he will cause corruption to appear") against yazhara l-fasadu ("corruption will appear") changes the subject of the verb.

Pharaoh asks Haman to build a tower. In Hafs fa-attali'a is subjunctive ("so that I may look"); in the other seven it is indicative, joined to "that I may reach the ways". Al-Tabari knew the accusative only from the non-canonical Humayd al-A'raj, and wrote: "The reading I do not permit any alternative to is the nominative, because of the consensus of the authoritative readers." The reading he would not permit is the one in the Hafs mushaf.

The second variant changes the meaning. With Hafs and Shu'bah's sudda Pharaoh "was barred" from the way, acted on by another. With sadda (the other six) he "barred others" from it.

Sources al-Tabari, Jami' al-Bayan, on 40:37.

'Ibad ("servants") and 'inda ("with, in the presence of") are different words with different consonants. Al-Tabari notes that the Medinans read 'inda, taking it from "those who are with your Lord are not too proud" (7:206), and accepts both, "since the angels are God's servants and are with Him". In one group of mushafs the verse states the angels' status, in the other their location.

Sources al-Tabari, Jami' al-Bayan, on 43:19.

In Hafs, al-Duri and al-Susi the promise is for the dead: those killed in battle. In the other five it is for all who fight, whether they die or survive.

Qablahu (six mushafs): the nations that preceded Pharaoh. Qibalahu (al-Duri, al-Susi): his followers and armies. Al-Tabari accepts both.

Sources al-Tabari, Jami' al-Bayan, on 69:9.

Wat'an (six mushafs) speaks of weight and effort. Al-Duri and al-Susi's wita'an (form III, muwata'ah) speaks of agreement, the heart agreeing with the tongue. Al-Tabari accepts both. These are two different reasons for night prayer.

Sources al-Tabari, Jami' al-Bayan, on 73:6.

Two letters, two meanings. Danin (with dad) is "stingy": the Prophet does not hold back what he receives. Zanin (with za') is "suspect, open to accusation": he is not to be doubted about it. Al-Tabari: with dad, "he is not stingy towards them in teaching them what God taught him"; with za', "he is not suspected" in what he conveys. The readers of Medina and Kufa read dad, some Meccans and Basrans za'. Among the eight, al-Duri, al-Susi, al-Bazzi and Qunbul read za'.

Sources al-Tabari, Jami' al-Bayan, on 81:24.

Latarkabunna (plural) addresses mankind. Latarkabanna (singular), the reading of al-Bazzi and Qunbul, addresses one person, explained by exegetes as the Prophet (or as the individual human being). The addressee of the prophecy changes.