What do Islam's own sources teach about violence?

Summary

• Both kinds of verse are really there. The Quran forbids compulsion in religion,¹ tells Muslims to incline to peace when the enemy does,² exempts treaty-keepers by name,³ and orders kindness to non-Muslims who do not fight.⁴ It also makes fighting a legal obligation,⁵ tells believers to kill polytheists wherever they find them,⁶ and tells them to fight the People of the Book until they pay a tax "while they are humbled."⁷ A case built on only one half of that list is not a case.
• The martial verses are late. On the traditional classification and on Theodor Noldeke's (d. 1930), the restraint-under-persecution material is Meccan and the command-to-fight material is Medinan, from the last ten years of Muhammad's life.⁸ The direction of travel is from patience to war, not the other way.
• Abrogation is the load-bearing question, and the numbers are wild. How many Quran verses did classical scholars say were cancelled? Forty-two in the treatise attributed to al-Zuhri (d. 124/742), 138 in al-Nahhas (d. 338/949), 238 in Ibn Salama (d. 410/1020), 20 in al-Suyuti (d. 911/1505), 5 in Shah Wali Allah (d. 1176/1762).⁹ The count moved by a factor of nearly fifty, and the reason is that the word naskh was stretched to cover things that were not abrogation at all.⁹
• Yes, the tradition does say Quran 9:5 cancels the peaceful verses — and the tradition also says it does not. Ibn Salama credits the "verse of the sword" with abrogating 124 other verses.⁹ Al-Husayn ibn al-Fadl (d. 282/895) says it cancelled every verse about patience under injury.¹⁰ Against that, al-Tabari (d. 310/923) — the most authoritative early commentator the tradition has — refused the abrogation claim in writing at Quran 2:190, 2:256, 8:61 and 60:8, and called it "arbitrary imposition."¹¹
• The verse of the sword is almost never quoted with what surrounds it. Quran 9:4 exempts treaty-keepers.³ Quran 9:6 grants protection to any polytheist who asks to hear the Quran, and then an escort to safety.¹² Quran 9:7 says "as long as they are upright toward you, be upright toward them."¹³ A polemic that prints 9:5 alone has destroyed its own credibility, because the next verse is one click away.
• Muhammad's Medinan career was not purely defensive. Al-Waqidi's own summary has him leading twenty-seven expeditions and fighting in nine.¹⁴ Tabuk in year 9 was a march toward Byzantine territory by an army no one had attacked. And the killing of the Banu Qurayza men is not a late invention: the arbitration that produced it is in Sahih al-Bukhari and Sahih Muslim.¹⁵
• "Islam spread by the sword" fails as a demographic claim. Islam became the majority religion of Iraq and Khurasan only in the tenth century, and of Egypt only around the fourteenth — three to seven centuries after the conquest.¹⁶
• Jews fared better under Islam than under Latin Christendom, and the historian who says so is not an apologist. Mark Cohen's comparative study finds anti-Jewish violence far less prevalent in the Islamic world, and gives five structural reasons for it.¹⁷ He rejects the "interfaith utopia" myth in the same essay.¹⁷
• Apostasy is where the tension is sharpest. "If somebody (a Muslim) discards his religion, kill him" is in Sahih al-Bukhari.¹⁸ "There shall be no compulsion in [acceptance of] the religion" is in the Quran.¹ Al-Tabari reconciled them by ruling that 2:256 never covered apostates in the first place.¹¹
• The comparison with the Bible cannot be dodged. Deuteronomy 20:16-18 and 1 Samuel 15:3 are more severe than anything in the Quran: no surrender is offered, and infants are named. The defensible Christian argument is about scope and permanence, not about severity — and it costs something to make.
• The modern numbers are usable only with their caveats. Of 452 suicide attacks worldwide in 2015, exactly two were carried out by non-jihadist organisations.¹⁹ That is a fact about one tactic in twenty-two mostly war-torn countries, not a measure of religious violence in general. Three of the nine most active terrorist organisations on earth in 2019 were Maoist.²⁰
• No single sentence about "what Muslims believe" is true. Pew asked Muslims in twenty-one countries whether suicide bombing in defence of Islam is ever justified. The answers ran from 1 per cent in Azerbaijan to 40 per cent in the Palestinian territories.²¹

Verdict.  Islam's foundational texts authorise offensive war to establish Muslim political supremacy, and classical Islamic law turned that into two permanent institutions that no Christian text creates:  jihad as a communal obligation binding until the end of the age, and dhimma, a permanent subordinate legal status for conquered non-Muslims.  That is the honest finding, and it is a claim about law rather than about body counts.  But the same sources restrain warfare in ways that popular Christian polemic systematically hides;  the doctrine that supposedly cancels those restraints was argued against inside the tradition by its own leading commentator;  and the historical record shows conquest producing tax revenue far more often than it produced converts.  A reader who wants either "the Quran is a book of peace" or "the Quran is a manual for terrorism" will find this article unsatisfying, because neither claim survives contact with the texts.

• Where this article's own case is weak. Four places, stated up front. First, most of what anyone knows about Muhammad's campaigns comes from the maghazi and sira literature — the raid-narratives and the biography — written 150 to 200 years after the events by authors who were not present. This article uses that material, and therefore cannot dismiss it when a Muslim quotes it back. Second, the strongest Christian structural argument is that the New Testament supersedes the conquest; that is itself an abrogation-shaped move, which is the very doctrine the article criticises. Third, the Old Testament conquest texts are harder to defend than anything examined here, and Christians disagree among themselves about how to read them. Fourth, two books this article needed could not be obtained: Richard Bulliet's Conversion to Islam in the Medieval Period (1979), whose percentages are therefore not printed, and Robert Pape's original 2003 paper, which is reported here only through the quotations his critics give.

The claim being tested

Three separate propositions get blurred into one another, and they have to come apart before anything can be tested.

Proposition one:  Islam is a religion of peace.  As usually stated by Muslim spokesmen, this means that the Quran permits fighting only in self-defence, that jihad — literally "striving" — primarily denotes inner moral struggle rather than war, and that Quran 2:256, "there shall be no compulsion in [acceptance of] the religion," is a general and permanent principle.

Proposition two:  Islam is inherently violent.  As usually stated by Christian polemicists, this means that the peaceful verses were abrogated by Quran 9:5, that Muhammad's later career was one of unprovoked expansion, and that classical Islamic law obliges Muslims to fight non-Muslims until Islam rules.

Proposition three:  Muslims are violent.  This is a claim about 1.9 billion living people, and it is not the subject of this article.  What a body of text authorises and what its adherents actually do are different questions with different evidence.  Where the two are run together, the argument becomes unfalsifiable in both directions:  a peaceful Muslim proves the religion is peaceful, and a violent one proves it is not.

This article tests propositions one and two against the Quran, the two hadith collections Sunni Muslims regard as soundest, the classical commentaries, the abrogation manuals, the earliest biographies of Muhammad, and the modern historical and survey literature.  It reaches a verdict on proposition three only to say that no single verdict is available.

Three commitments govern the method.  Every source is named and dated, so a reader can check it.  Where a report comes from the sira — the biographical tradition — rather than from a hadith collection, the article says so every time, because those are different pools that received different levels of scrutiny.  And the same standard is applied to the Bible that is applied to the Quran;  a section that embarrasses a Christian is not omitted.

The strongest case for the Muslim claim

Stated at full strength, without evaluation.  A believing Muslim should recognise this section as fair.  Every text named here is quoted in full further down.

The Quran's first permission to fight is explicitly defensive, and it says so.  Quran 22:39 gives permission to fight "to those who are being fought, because they were wronged."  Abdel Haleem renders it "Those who have been attacked are permitted to take up arms."²²  The next verse identifies them as people evicted from their homes for saying "Our Lord is Allāh," and gives the purpose of divinely sanctioned force in words Christian polemic almost never quotes — preventing the destruction of "monasteries, churches, synagogues, and mosques."²³  Christian and Jewish buildings, named, and named first.

The core fighting passage is framed as reciprocity, three times in five verses.  "Fight in the way of Allāh those who fight against you but do not transgress" — Pickthall renders the second clause "but begin not hostilities."²⁴  "And if they cease, then indeed, Allāh is Forgiving and Merciful."²⁵  "So whoever has assaulted you, then assault him in the same way that he has assaulted you."²⁶

Peace offers must be accepted.  "And if they incline to peace, then incline to it [also]."²  Ibn Kathir (d. 774/1373) ties the verse directly to the truce of Hudaybiyya:  "when the pagans inclined to peace in the year of Hudaybiyah and sought cessation of hostilities for nine years… he accepted this from them."²⁷

The verse critics quote most is the verse that most needs its neighbours.  Quran 9:5 sits inside a passage that exempts by name every polytheist tribe that kept its treaty (9:4), grants safe passage and an escort home to any polytheist who asks to hear the Quran (9:6), and states the ground of the fighting as treaty-breaking and a first attack (9:12-13) rather than unbelief.²⁸

The rules of engagement are in the soundest collections.  Muhammad forbade the killing of women and children.²⁹  Sahih Muslim has him instructing every commander not to embezzle spoils, break pledges, mutilate bodies or kill children, and to offer an enemy three options in order:  Islam, the poll tax, and only then fighting.³⁰

The hadith most often quoted against Islam was first used as an argument for restraint.  "I have been ordered (by Allah) to fight against the people until they testify that none has the right to be worshipped but Allah…" is in Sahih al-Bukhari.³¹  Its earliest recorded use, in Sahih Muslim, is in the caliph Umar's mouth, objecting to Abu Bakr's war on the tribes that stopped paying the alms tax.³²  The saying's first job in the record was to stop a war, not to start one.

"No compulsion in religion" was defended by the classical tradition, not abandoned by it.  Ibn Kathir glosses Quran 2:256 "Do not force anyone to become Muslim, for Islam is plain and clear," and records no abrogation of it.²⁷  Al-Tabari (d. 310/923) records the abrogation claim and rejects it:  "there is no sense in the statement of one who claims the verse's ruling is abrogated by the permission to make war."¹¹

Abrogation was contested inside the tradition, and the maximal counts were falling for centuries.  Al-Suyuti recognised twenty genuine cases;  Shah Wali Allah reduced it to five.⁹  The sword verse is itself said to be abrogated — by the conditional clause with which it ends.⁹

The conquests were not conversion campaigns, and the best witness is a Christian one.  Robert Hoyland notes that later Muslim historians claimed the conquerors offered conversion before fighting, but "this is never mentioned in earlier sources."  He quotes John bar Penkaye, a Syriac Christian writing about 687:  "Of each person they required only tribute, allowing him to remain in whatever faith he wished."¹⁶

And the demography confirms it.  If conquest converted populations, Egypt would have been Muslim in the seventh century.  It became majority Muslim around the fourteenth.¹⁶

That is the Muslim case at its best.  Most of it is true.  The rest of this article is about what it leaves out.

The texts, printed whole

Everything below is quoted from quran.com's text of the 1924 Cairo edition, in the Sahih International translation unless another translator is named.  Sahih International's square brackets are the translator's additions, not the Arabic, and they are kept so the reader can see them.  Where an argument turns on a word, a second translation and the Arabic term are given.

Where the fighting verses come in the story

The Quran's chapters are not in chronological order.  Both Muslim tradition and Western scholarship divide them into Meccan — before the emigration to Medina in 622 CE — and Medinan, after it.  Gerhard Bowering states the rule in the Encyclopaedia of the Qur'an:  the division follows "whether they were revealed before or after the emigration."⁸  It is a boundary in time, not in geography.

Every verse below that commands fighting is Medinan.  Every verse that commands patience under persecution is Meccan.  Both sides accept that pattern, and it is why abrogation matters so much:  if the later cancels the earlier, the Quran's settled position on war is its most militant one.  One caution, though.  The relative order within the Medinan material is not secure.  Behnam Sadeghi's stylometric study — the strongest quantitative work on Quranic chronology — confirmed the earliest phases and explicitly failed to confirm the internal ordering of the Medinan block.³³  So this article says "Meccan" and "Medinan" freely, and never ranks two Medinan verses against each other on the strength of a traditional list.

The permission to fight — Quran 22:39-40

"Permission [to fight] has been given to those who are being fought, because they were wronged.  And indeed, Allāh is competent to give them victory."²²

"[They are] those who have been evicted from their homes without right - only because they say, 'Our Lord is Allāh.'  And were it not that Allāh checks the people, some by means of others, there would have been demolished monasteries, churches, synagogues, and mosques in which the name of Allāh is much mentioned [i.e., praised]."²³

The tradition treats this as the first permission to fight.  Two features are worth noticing before anything else, because a Christian polemic normally passes over both.  The stated ground is being attacked and expelled, not the other party's unbelief.  And the stated purpose of divinely sanctioned force includes protecting monasteries, churches and synagogues — by name, and listed before mosques.

Reciprocity and its limit — Quran 2:190-194

2:190 — "Fight in the way of Allāh those who fight against you but do not transgress.  Indeed, Allāh does not like transgressors."  Pickthall:  "Fight in the way of Allah against those who fight against you, but begin not hostilities.  Lo!  Allah loveth not aggressors."²⁴

2:191 — "And kill them [in battle] wherever you overtake them and expel them from wherever they have expelled you, and fitnah is worse than killing.  And do not fight them at al-Masjid al-Ḥarām until they fight you there.  But if they fight you, then kill them."³⁴

2:192 — "And if they cease, then indeed, Allāh is Forgiving and Merciful."²⁵

2:193 — "Fight them until there is no [more] fitnah and [until] religion [i.e., worship] is [acknowledged to be] for Allāh.  But if they cease, then there is to be no aggression [i.e., assault] except against the oppressors."³⁵

2:194 — "So whoever has assaulted you, then assault him in the same way that he has assaulted you."²⁶

The whole meaning of 2:193 turns on one Arabic word, fitnah, and the translators do not agree about it.  Pickthall renders it "persecution."  Yusuf Ali gives "Tumult or oppression."  Abdel Haleem gives "persecution."  Sahih International refuses to translate it and leaves it transliterated.³⁵  If fitnah means persecution, 2:193 says "fight until the persecution stops."  If it means unbelief or idolatry, it says "fight until Islam is the only religion."  Neither side may pick a gloss silently, and this article does not.

Al-Tabari on 2:190 rejects the claim that it was abrogated, and gives his reason.  He reports the abrogation view on the authority of Ibn Zayd, and the opposite view — that "do not transgress" means "do not kill women, children and monks" — on the authority of the caliph Umar ibn Abd al-Aziz (d. 101/720).  Then he rules:

"Abu Ja'far said:  the more correct of these two positions is the one Umar ibn Abd al-Aziz stated.  Because a claimant's claim that a verse is abrogated, when it may possibly be un-abrogated, without evidence for the soundness of his claim, is arbitrary imposition (tahakkum) — and arbitrary imposition is beyond no one's power."¹¹

That is a methodological objection, and a good one:  anyone can assert that a verse has been cancelled, so the burden of proof has to sit with the person asserting it.

And al-Tabari's non-abrogation reading is not a pacifist reading.  In the same passage he glosses the verse's positive command as fighting the unbelievers until they return to God's obedience "or give you the jizya in humiliation if they are People of the Book."¹¹  What he denies is that warfare has no limits.  He does not deny that warfare against unbelievers is commanded.

Fighting as an obligation — Quran 2:216

"Battle has been enjoined upon you while it is hateful to you.  But perhaps you hate a thing and it is good for you;  and perhaps you love a thing and it is bad for you.  And Allāh knows, while you know not."⁵

Abdel Haleem: "Fighting is ordained for you, though you dislike it."⁵  The Arabic is kutiba alaykum al-qital — literally "fighting has been written upon you."  It is the same construction the Quran uses for the fast of Ramadan at 2:183 and for the law of retaliation at 2:178.  That is the language of legal obligation, not of permission, and it is the plainest textual answer to the claim that jihad means only inner struggle.

No compulsion in religion — Quran 2:256, and what it actually covered

"There shall be no compulsion in [acceptance of] the religion.  The right course has become distinct from the wrong.  So whoever disbelieves in ṭāghūt and believes in Allāh has grasped the most trustworthy handhold with no break in it."¹

This is the most-quoted verse in the entire debate, and both sides misreport what the classical commentators said about it.

Ibn Kathir does not say it is abrogated.  He glosses it "Do not force anyone to become Muslim, for Islam is plain and clear," gives the occasion — Medinan parents whose children had been raised as Jews among the Banu Nadir — and handles a hadith in which Muhammad tells a reluctant man to embrace Islam "even if you dislike it," concluding that "the Prophet did not force that man to become Muslim."²⁷

Al-Tabari says it is not abrogated, and then restricts it.  He surveys the views and rules that the verse "was revealed concerning a specific class of people," namely "the People of the two Books and the Magians and everyone whose confirmation in his religion contrary to the religion of truth is accepted and the jizya taken from him."  He closes:  "there is no sense in the statement of one who claims the verse's ruling is abrogated by the permission to make war."¹¹

And in the same passage he states the exclusions:

"All the Muslims have transmitted from their Prophet … that he compelled a people to Islam and refused to accept from them anything but Islam, and ruled that they be killed if they refused it — such as the idol-worshippers among the Arab polytheists, and the apostate from his religion, the religion of truth, to unbelief."¹¹

Al-Qurtubi (d. 671/1273) lists six views, and puts abrogation first.  "The scholars differed over the meaning of this verse in six views.  (First) It is said that it is abrogated, because the Prophet compelled the Arabs to the religion of Islam and fought them and accepted from them nothing but Islam.  Sulayman ibn Musa said this;  he said:  Quran 9:73 abrogated it.  This is related from Ibn Mas'ud and many of the commentators."  His second view is al-Tabari's.¹⁰

So the honest verdict is uncomfortable for everyone.  "No compulsion in religion" is real, it is Quranic, and the tradition's greatest commentator defended it against those who wanted it cancelled.  But the classical mainstream reading is not "nobody may ever be coerced in religion."  It is "those from whom the poll tax may be accepted may not be coerced" — which by al-Tabari's own statement excludes Arab pagans and excludes apostates.  Saying either more or less than that is a misreport.

The waverers — Quran 4:89-91

4:89 — "So do not take from among them allies until they emigrate for the cause of Allāh.  But if they turn away [i.e., refuse], then seize them and kill them [for their betrayal] wherever you find them."³⁶

4:90 — "Except for those who take refuge with a people between yourselves and whom is a treaty or those who come to you, their hearts strained at [the prospect of] fighting you… So if they remove themselves from you and do not fight you and offer you peace, then Allāh has not made for you a cause [for fighting] against them."³⁷

4:91 — Abdel Haleem:  "So if they neither withdraw, nor offer you peace, nor restrain themselves from fighting you, seize and kill them wherever you encounter them."³⁸

Quoting 4:89 without 4:90 and 4:91 is the single commonest citation abuse in this area, and it is checkable in ten seconds.  The passage as a unit is about people who had left Medina, professed loyalty to both sides, and could rejoin the fighting.  It exempts by name three groups:  those who reach a people Medina has a treaty with, those who come over, and those who offer peace.

The Badr passages — Quran 8

Surah 8 is tied to the battle of Badr in 624 CE.

8:12 — "[Remember] when your Lord inspired to the angels, 'I am with you, so strengthen those who have believed.  I will cast terror into the hearts of those who disbelieved, so strike [them] upon the necks and strike from them every fingertip.'"³⁹

This verse appears constantly in polemical lists as a command to Muslims.  It is not.  The Quran's own opening clause identifies the addressee as the angels.  A polemic that presents 8:12 as a standing order to believers has misread the sentence, and a Muslim who points that out is right.

8:39 — "And fight against them until there is no fitnah and [until] the religion [i.e., worship], all of it, is for Allāh.  And if they cease - then indeed, Allāh is Seeing of what they do."⁴⁰

8:57 — "So if you, [O Muḥammad], gain dominance over them in war, disperse by [means of] them those behind them that perhaps they will be reminded."  Abdel Haleem:  "make a fearsome example of them to those who come after them."⁴¹

8:60 — "And prepare against them whatever you are able of power and of steeds of war by which you may terrify the enemy of Allāh and your enemy."⁴²

8:61 — "And if they incline to peace, then incline to it [also] and rely upon Allāh."²

8:67 — "It is not for a prophet to have captives [of war] until he inflicts a massacre [upon Allāh's enemies] in the land.  You [i.e., some Muslims] desire the commodities of this world, but Allāh desires [for you] the Hereafter."  Abdel Haleem:  "It is not right for a prophet to take captives before he has conquered the battlefield."⁴³

Verse 67 rebukes the Muslims for having taken ransom at Badr rather than killing.  The verb is yuthkhina, to make heavy slaughter, to subdue thoroughly.  It is the hardest verse in surah 8 and it is rarely cited by either side, because it embarrasses the militant reading (which prefers verses addressed to all Muslims for all time) and the pacific reading equally.

Al-Tabari on 8:61 refuses the abrogation claim in the strongest words he uses anywhere in this material.  Qatada and others had said 8:61 was cancelled by the sword verse.  Al-Tabari answers:

"As for what Qatada and those who said the like of his statement said, that this verse is abrogated — that is a statement for which there is no evidence from Book, nor sunna, nor the native constitution of reason."¹¹

His reason is a distinction, not a sentiment:  8:61 concerns the Banu Qurayza, who were People of the Book and from whom the poll tax may be taken;  9:5 concerns the Arab idol-worshippers, from whom it may not.  "Neither of the two verses negates the ruling of the other;  rather each of them is operative in what it was revealed about."¹¹

The verse of the sword, with its neighbours — Quran 9:1-15

Surah 9 is the one Muslim tradition most often names as the last revealed.  Verse 5 is the single most quoted verse in the whole argument.  Here it is, bare:

"And when the inviolable months have passed, then kill the polytheists wherever you find them and capture them and besiege them and sit in wait for them at every place of ambush.  But if they should repent, establish prayer, and give zakāh, let them [go] on their way.  Indeed, Allāh is Forgiving and Merciful."

Abdel Haleem:  "When the [four] forbidden months are over, wherever you encounter the idolaters, kill them, seize them, besiege them, wait for them at every lookout post;  but if they turn [to God], maintain the prayer, and pay the prescribed alms, let them go on their way, for God is most forgiving and merciful."⁶

Now the verses that surround it, which are almost never quoted with it.

• 9:1-2 — a declared release from treaty obligations, with four months of safe movement first.
• 9:4 — "Excepted are those with whom you made a treaty among the polytheists and then they have not been deficient toward you in anything or supported anyone against you; so complete for them their treaty until their term [has ended]."³
• 9:6 — "And if any one of the polytheists seeks your protection, then grant him protection so that he may hear the words of Allāh. Then deliver him to his place of safety."¹²
• 9:7 — "So as long as they are upright toward you, be upright toward them."¹³
• 9:11 — "But if they repent, establish prayer, and give zakāh, then they are your brothers in religion."⁴⁴
• 9:12-13 — the stated reason for war: "if they break their oaths after their treaty and defame your religion" / "a people who broke their oaths and determined to expel the Messenger, and they had begun [the attack upon] you the first time."²⁸

That is the case for the restrained reading, and it is a strong one.  Verse 5 stands between an exemption for treaty-keepers and a safe-conduct clause, in a passage that states treaty-breaking and a first attack as its ground.

And here is the other half of the evidence, which a Muslim apologetic presentation leaves out just as reliably.

Ibn Kathir heads his comment on 9:5 "This is the Ayah of the Sword."  He glosses "besiege them, and lie in wait for them in each and every ambush" as:  "do not wait until you find them.  Rather, seek and besiege them in their areas and forts… This way, they will have no choice, but to die or embrace Islam."  And he reports two early authorities:

"This honorable Ayah (9:5) was called the Ayah of the Sword, about which Ad-Dahhak bin Muzahim said, 'It abrogated every agreement of peace between the Prophet and any idolator, every treaty, and every term.'  Al-'Awfi said that Ibn 'Abbas commented:  'No idolator had any more treaty or promise of safety ever since Surah Bara'ah was revealed.'"²⁷

Ibn Kathir on 9:6 is equally in the record, and equally binding:  the person granted protection is safe "until he goes back to his land, his home, and area of safety," and the ruling covers people who come "for business transactions, to negotiate a peace treaty, to pay the Jizyah, to offer an end to hostilities."²⁷

The honest summary.  Quran 9:5 is directed at Arab polytheists who had broken treaties, inside a passage that repeatedly exempts treaty-keepers and grants safe conduct to enquirers.  And a major classical commentator reads it as cancelling every treaty with every idolater and leaving them "no choice, but to die or embrace Islam."  Both of those things are in the sources.  An article that gives only the first is apologetics.  An article that gives only the second is polemic.

The poll-tax verse — Quran 9:29

"Fight against those who do not believe in Allāh or in the Last Day and who do not consider unlawful what Allāh and His Messenger have made unlawful and who do not adopt the religion of truth [i.e., Islām] from those who were given the Scripture - [fight] until they give the jizyah willingly while they are humbled."⁷

The last clause is hatta yu'tu al-jizyata 'an yadin wa-hum saghirun, and the four standard English translations diverge sharply on it.

Abdel Haleem's "agree to submit" is the softest reading in the set, and it is not what al-Tabari says the word means.  Al-Tabari glosses wa-hum saghirun:  "its meaning is:  while they are abased and subdued.  The lowly and contemptible person is called saghir."¹¹  On 'an yadin he says the Arabs use the phrase "of every giver who gives something to one who has power over him, whether willingly or unwillingly."  On what the abasement consists of, he reports disagreement:  Ikrima said "you take it while you are seated and he is standing";  others said the payers must carry it over themselves, unwilling;  others that the payment itself is the abasement.¹¹

This verse is more important to the argument than 9:5, and it gets far less attention.  Quran 9:5 concerns Arab polytheists, a category that ceased to exist within a generation.  Quran 9:29 concerns Jews and Christians, and it is the textual foundation of dhimma — the permanent protected-but-subordinate status of non-Muslims under Islamic rule.  It is not an emergency measure.  It is a constitution.

Ibn Kathir connects it to the economics of the moment:  when idolaters were barred from the Meccan sanctuary the Muslims feared the loss of trade, and "Allah compensated them for the losses they incurred because they severed ties with idolators, by the Jizyah they earned from the People of the Book."²⁷

The remaining commands — 9:73, 9:111, 9:123, 47:4

9:73 — "O Prophet, fight against the disbelievers and the hypocrites and be harsh upon them."⁴⁵

Two qualifications belong with this verse.  The verb is jahid, which covers argument and moral effort as well as war.  And "the hypocrites" were Medinan Muslims of doubtful loyalty against whom no fighting is recorded in the sources.  Al-Qurtubi nevertheless reports it as the verse that abrogated 2:256, on one classical view.¹⁰

9:111 — "Indeed, Allāh has purchased from the believers their lives and their properties [in exchange] for that they will have Paradise.  They fight in the cause of Allāh, so they kill and are killed.  [It is] a true promise [binding] upon Him in the Torah and the Gospel and the Qur'ān."⁴⁶

This is the plainest paradise-for-death transaction in the Quran, and it makes a checkable claim:  that the Torah and the Gospel contain the same promise.  That is where a Christian should press, because it is a factual claim about two documents anyone can read.  No text in the Gospels promises paradise for killing and being killed in God's cause.  Pressing instead on the verse's moral tone concedes the argument to taste.

9:123 — "O you who have believed, fight against those adjacent to you of the disbelievers and let them find in you harshness."⁴⁷

A geography of expansion:  nearest first.

47:4 — "So when you meet those who disbelieve [in battle], strike [their] necks until, when you have inflicted slaughter upon them, then secure [their] bonds, and either [confer] favor afterwards or ransom [them] until the war lays down its burdens."⁴⁸

The mercy clause at the end — release or ransom — became the subject of a live classical dispute about whether it survived.  Al-Tabari records Ibn Jurayj and al-Suddi holding that it was abrogated by 9:5 and 8:57, and Qatada holding that 8:57 alone did it.¹¹  Al-Qurtubi records the argument running the other way as well:  al-Dahhak, al-Suddi and Ata' said 9:5 was the verse abrogated, by 47:4;  Mujahid and Qatada said 9:5 abrogates 47:4;  and Ibn Zayd said both verses are operative.  Al-Qurtubi endorses the last:  "that is the correct view, because favour, killing and ransom never ceased to be the Messenger of God's ruling concerning them from the first war he fought, which was the day of Badr."¹⁰

That single paragraph is the best short demonstration of what the abrogation literature actually looks like.  Two respected scholars, reading the same two verses, produce opposite lists of what cancels what — and a third says neither cancels anything.

The kindness verses — Quran 60:8-9

60:8 — "Allāh does not forbid you from those who do not fight you because of religion and do not expel you from your homes - from being righteous toward them and acting justly toward them.  Indeed, Allāh loves those who act justly."⁴

60:9 — "Allāh only forbids you from those who fight you because of religion and expel you from your homes and aid in your expulsion - [forbids] that you make allies of them."⁴⁹

Ibn Zayd and Qatada both held that 9:5 abrogated 60:8.  Al-Tabari refuses again, and his wording matters because of how wide he makes it:  the verse means "God does not forbid you, regarding those who did not fight you over religion, from every category of creed and religion, to treat them well and be just to them … and there is no sense in the statement of the one who says it is abrogated."  He grounds the ruling on the report about Asma bint Abi Bakr and her pagan mother Qutayla.¹¹

The hadith: the order to fight

Sahih al-Bukhari 25, narrated by Ibn Umar:

"I have been ordered (by Allah) to fight against the people until they testify that none has the right to be worshipped but Allah and that Muhammad is Allah's Messenger, and offer the prayers perfectly and give the obligatory charity, so if they perform that, then they save their lives and property from me except for Islamic laws and then their reckoning (accounts) will be done by Allah."³¹

Sahih Muslim 22 carries the same report through the same transmitter.  Neither al-Bukhari (d. 256/870) nor Muslim (d. 261/875) prints a grade beside it, because inclusion in their collections is itself the claim being made;  there is no external verdict to cite, and an article that implies one is misreporting how these books work.

Three qualifications belong with it, and all three are inside the Muslim sources.

1. Sahih Muslim 21a has a shorter form from Abu Hurayra — "till they do not say: There is no god but Allah" — with no mention of prayer or the alms tax. The two forms were used to support different legal conclusions.
2. Sahih Muslim 20 puts the saying in the caliph Umar's mouth as an objection to Abu Bakr's war against tribes that had stopped paying zakat. Abu Bakr's answer is that prayer and the alms tax cannot be separated.³² The earliest recorded use of this hadith was as an argument for restraint.
3. al-nas, "the people," is read by most classical jurists as the Arab polytheists rather than all humanity — precisely because Quran 9:29 provides for Jews and Christians to live under the poll tax instead. The unrestricted reading is not the majority classical one.

Alongside it sit the martyrdom reports.  Sahih al-Bukhari 2797 has Muhammad saying "I would love to be martyred in Allah's Cause and then get resurrected and then get martyred, and then get resurrected again and then get martyred."⁵⁰  Sahih al-Bukhari 36 says the man who fights in God's cause with no other motive "will be recompensed by Allah either with a reward, or booty (if he survives) or will be admitted to Paradise."⁵¹

The hadith: rules of engagement, and their limit

The restraints are real and they are in the soundest collections.  Sahih al-Bukhari 3015 records that a woman was found killed after a raid, "so Allah's Messenger forbade the killing of women and children."²⁹  Sahih Muslim 1731a is the fullest rules-of-engagement text:  "do not embezzle the spoils;  do not break your pledge;  and do not mutilate (the dead) bodies;  do not kill the children," and then the three options offered to an enemy in order — Islam, the poll tax, fighting.³⁰

And here is the counterweight, which must not be suppressed.  Sahih Muslim 1745a stands under a chapter heading that Muslim's own editors wrote:  "Permissibility of killing women and children in night raids, so long as it is not done deliberately."  Asked about the women and children of the polytheists being killed in a night raid, Muhammad said:  "They are from them."⁵²

So the classical rule is a prohibition on targeting non-combatants, not a prohibition on foreseeably killing them.  In modern terms it is a doctrine of double effect, not a doctrine of immunity.  Christians should notice that this is roughly the position of Western just-war theory and of the laws of armed conflict as well, which is a reason to describe it accurately rather than to exclaim over it.

The "greater jihad" hadith

The saying is:  "We have returned from the lesser jihad to the greater jihad," with a fuller version adding that the greater jihad is "the servant's struggle against his desire."  It is quoted constantly in interfaith settings, and its status has to be stated exactly, because both sides get it wrong.

It is recorded by al-Bayhaqi (d. 458/1066) in al-Zuhd al-Kabir and by al-Khatib al-Baghdadi (d. 463/1071) in Tarikh Baghdad.  Both chains run through Yahya ibn al-Ala' al-Bajali, whom Ibn Hajar's Taqrib accuses of forgery, and through Layth ibn Abi Sulaym.  Ten separate graders across eight centuries have judged it, and not one of them called it sound:⁵³

What follows, and what does not.  A Muslim who presents "the greater jihad" as a teaching of Muhammad is citing a report that no hadith critic in a thousand years has graded sound.  But a Christian who calls it "a Sufi fabrication" overstates twice.  First, la asl lahu in hadith terminology means chainless, and this report does have chains.  Second, the underlying idea is independently attested in material the critics accepted:  Jami' at-Tirmidhi 1621 has "The Mujahid is one who strives against his own soul," and al-Tirmidhi (d. 279/892) grades the hadith in that chapter hasan sahih in his own words.⁵⁴  The greater-jihad saying is weak.  The greater-jihad concept is not an invention.

Abrogation: the load-bearing question

Everything turns on this.  If the later Medinan war verses cancel the earlier Meccan verses of patience, then the Quran's settled position is its most militant one and the restraint verses are historical curiosities.  If they do not, then the two sets have to be read together and the restraints stand.  So the doctrine of naskh — abrogation, the replacement of one revealed ruling by a later one — is where the argument is actually decided.

What the doctrine is, and what the Quran says about it

Four verses are cited for it.  The main two:

2:106 — "We do not abrogate a verse or cause it to be forgotten except that We bring forth [one] better than it or similar to it."⁵⁵

16:101 — "And when We substitute a verse in place of a verse - and Allāh is most knowing of what He sends down - they say, 'You, [O Muḥammad], are but an inventor [of lies].'"⁵⁶

Ibn Kathir's comment on 2:106 collects the classical range of meanings in one place, and they are not all the same thing:  "Whatever an Ayah We abrogate" (Ibn Abbas);  "Whatever an Ayah We erase" (Mujahid);  "We keep the words, but change the meaning";  and the erasure of wordings no longer in the text, of which the stock example is a supposed stoning verse.  He then quotes al-Tabari's restriction, which is the anchor of the modern Muslim position:  "The Nasakh only occurs with commandments, prohibitions, permissions, and so forth.  As for stories, they do not undergo Nasakh."²⁷

How many verses were abrogated? The real counts

The best short academic survey of this literature is David S. Powers' chapter in Andrew Rippin's Approaches to the History of the Interpretation of the Qur'an (Oxford, 1988).  Powers gives the figures directly:

"The number of verses that are considered to have been abrogated increased dramatically between the eighth and eleventh centuries (al-Zuhrī mentions 42 abrogated verses, al-Naḥḥās, 138, and Ibn Salāma, 238), at which point an upper limit seems to have been reached (Ibn al-'Atā'iqī identifies 231 abrogated verses, and al-Fārisī, 248)."⁹

And then the collapse:

"al-Suyūṭī (d. 911/1505) recognized only twenty instances of true abrogation, and Shāh Walī Allāh (d. 1762) reduced that number to five."⁹

A range from 5 to 248 is not a range of opinion about a fact.  It is a sign that different scholars were counting different things.  Powers says exactly that:  the increase came from "the expansion of the semantic range of the term naskh to include phenomena that did not originally fall within its scope."  He quotes Ibn al-Ata'iqi's modern editor distinguishing the jurists' naskh — replacing one legal ruling with another — from the commentators' much looser usage, which swept in "specification, exception, abandoning a legal rule because circumstances have changed, and mutual cancellation."⁹

One honesty note about two of those numbers.  Powers took the figures 20 and 5 not from al-Suyuti's al-Itqan and Shah Wali Allah's al-Fawz al-Kabir directly, but from Ernest Hahn's 1974 translation of Sayyid Ahmad Khan.  Powers says so in his own footnote.  This article has not gone behind Powers to the Arabic either, and anyone who intends to lean hard on those two figures should.⁹

Does the tradition say Quran 9:5 abrogates the peaceful verses?

Yes.  Powers, again verbatim:

"The distinction of abrogating the greatest number of verses in the Qur'ān belongs to Q. 9/5, known in Arabic as āyat al-sayf ('the sword verse'), which abrogated no less than 124 other verses.  This verse … abrogates every other verse in the Qur'ān which commands or implies anything less than a total offensive against the non-believers."⁹

His footnote for the figure of 124 is "Ibn Salāma, p. 184;  Ibn al-'Atā'iqī, p. 53."  So the number that circulates on Christian apologetics sites is real and traceable.  It is Hibat Allah ibn Salama's, from a manual printed in Cairo in 1315/1899.

Two cautions the sites never carry.  Powers' own printed list does not add up to 124.  His Appendix B reproduces Ibn Salama's list of verses abrogated by the sword verse;  counted off the page, it holds 103 entries covering about 108 verses, and one entry — "Sūra 93: 22" — cannot be right, since surah 93 has eleven verses.  The list also spans 50 surahs, where Powers' own prose on the facing page says the abrogated verses "occur in fifty-three sūras."  Neither gap is explained in the chapter.  This article prints 124 with Powers' citation and reports the discrepancy rather than quietly reconciling them.⁹  And the variant figure "114," which also circulates, was not found in any source consulted here and should not be used.

Al-Qurtubi supplies the strongest single classical sentence for the maximal reading.  On 9:5 he reports:  "al-Husayn ibn al-Fadl said:  this [verse] abrogated every verse in the Quran in which there is mention of turning away and of patience under the harm of enemies."  Al-Husayn ibn al-Fadl al-Bajali died in 282/895.¹⁰

Verses on Ibn Salama's list include 2:190, 2:191, 2:192, 2:256, 4:90, 4:91, 16:125, 47:4, 60:8 and 60:9 — that is, most of the restraint verses printed earlier in this article.⁹

And the tradition also says the opposite

Powers records the classical qualifications too, and they are not small:

"Remarkably, the sword verse, which abrogated no fewer than 124 other verses, is itself considered to have been abrogated by the conditional clause with which it concludes:  'But if they repent, and perform the prayer, and pay the alms, then let them go their way;  God is All-forgiving, All-compassionate.'"⁹

On the kind of cancellation involved, Powers adds:  since these verses are not phrased as commands, "it is only the ma'nā, or sense, of the verse that is abrogated, not the actual phrasing."⁹  The words stay in the Quran and stay recited;  what is said to lapse is their legal force.

The decisive fact for fairness is that al-Tabari argued against the sword-verse thesis verse by verse, in print, in the ninth century.  His refusals at 2:190, 2:256, 8:61 and 60:8 are all quoted above.  His method is consistent and it is a real argument:  an abrogator is only an abrogator where it negates the abrogated ruling entirely, so that the two cannot stand together;  a command that is general in wording but specific in reference is not abrogated at all, it is simply about something else.

So the claim "the classical tradition unanimously held that 9:5 abrogated the peaceful verses" is false, and a Christian who makes it can be refuted from al-Tabari in one move.

But the counter to that runs the other way just as hard.  None of it makes the sword-verse reading marginal.  It is in al-Qurtubi.  It is in Ibn Kathir.  It is the organising principle of Ibn Salama's manual, and manuals like it were the working reference for jurists for centuries.  A doctrine that the tradition's own greatest commentator has to argue against, verse by verse, four separate times, is a doctrine the tradition held.

The modern Muslim positions

Louay Fatoohi's Abrogation in the Qur'an and Islamic Law (Routledge, 2013) is the second book in English devoted to the subject, and it argues that abrogation should be rejected outright.  Fatoohi is a Muslim writer arguing a minority Muslim position, and he says so.  His own summary of where the majority stands is the most useful sentence in the book for present purposes:

"The overwhelming majority of Muslim scholars, past and modern, have accepted abrogation in both the Qur'an and the Sunna as an indisputable fact.  Only a very small minority has rejected Qur'anic abrogation."⁵⁷

And on the numbers:  "some scholars have identified over two hundred Qur'anic verses that they claim to have been abrogated by other Qur'anic verses or by sayings and actions of Prophet Muhammad, whereas others have reduced the number to a single digit!"⁵⁷  His own conclusion is that "abrogation is a phenomenon that lacks any support from the Qur'an," and his final chapter is titled "Conclusion:  The Myth of Abrogation."⁵⁷

The Western scholarly literature points in a compatible direction from a different angle.  John Burton argued that the Muslim accounts of how the Quran reached its present shape were introduced late by jurists who needed them to support their theories of abrogation — an argument Powers summarises in his opening footnote.⁹  Powers' own contribution is the demonstration that the doctrine was first applied to the Quran's bequest verses "a generation or more after Muḥammad's death."⁹

Where that leaves the question.  The strongest Muslim response to the sword-verse thesis does not require anything modern.  It requires al-Tabari.  The strongest Christian answer is that al-Tabari lost the argument in practice:  the abrogation manuals, not al-Tabari's commentary, are what the naskh tradition transmitted, and a legal system does not run on the dissenting opinion.

The history: what Muhammad actually did

The source-critical status of this section, stated once and applying throughout.  Everything in the table below comes from the maghazi and sira literature — al-Waqidi (d. 207/822), Ibn Ishaq (d. 150/767) as transmitted by Ibn Hisham (d. 218/833), and al-Tabari.  These are reports through named chains, written down 150 to 200 years after the events.  They are not hadith and they were not graded by the hadith critics.  This cuts both ways, and this article accepts both edges:  the material cannot be treated as decisive evidence against Muhammad in one paragraph and dismissed as legend in the next.  Where an episode also appears in Sahih al-Bukhari or Sahih Muslim, that is noted, and those are the strongest cases.

Page references are to Rizwi Faizer's translation of al-Waqidi's Kitab al-Maghazi (Routledge, 2011) and to Alfred Guillaume's translation of Ibn Ishaq (Oxford, 1955).

Al-Waqidi's own summary is the most useful single anchor in the whole literature, because it is a number the tradition volunteers about itself:

"The Prophet actively participated in twenty-seven raids.  He fought in nine of them:  Badr, Uhud, al-Muraysi', al-Khandaq, Qurayza, Khaybar, the Conquest of Mecca, Hunayn and al-Ta'if."¹⁴

Twenty-seven expeditions led in person in nine years, of which nine involved fighting, plus several dozen smaller raiding parties sent under other commanders.  That is not the career of a man who fought only when attacked, and no reading of the sources makes it one.  Tabuk is the clearest case:  an army marched about 400 miles north toward Byzantine territory in year 9 and came home without a battle, and nothing in the sources says Byzantium had attacked Medina.

Banu Qurayza

This is the hardest episode, and burying it would be dishonest.  Ibn Ishaq, in Guillaume's translation:

"Then they surrendered, and the apostle confined them in Medina in the quarter of d. al-Harith, a woman of B. al-Najjar.  Then the apostle went out to the market of Medina (which is still its market today) and dug trenches in it.  Then he sent for them and struck off their heads in those trenches as they were brought out to him in batches.  Among them was the enemy of Allah Huyayy b. Akhtab and Ka'b b. Asad their chief.  There were 600 or 700 in all, though some put the figure as high as 800 or 900."⁵⁸

The men of a Jewish tribe of Medina were executed and the women and children enslaved, after the tribe had negotiated with the besieging coalition during the battle of the Trench.  The sentence was pronounced by Sa'd ibn Mu'adh, chief of the Aws, whom the Qurayza themselves accepted as arbitrator.  The arbitration is not sira-only.  It is in Sahih al-Bukhari 4121 and Sahih Muslim 1769, which puts the core of the episode inside the material Sunni Muslims accept as sound.¹⁵  A Muslim who accepts al-Bukhari for law cannot reject this report merely because a critic quotes it.

The counter-considerations are also real and belong in the same paragraph.  The Qurayza had a treaty with Medina and, on the sources' own account, broke it under siege.  The arbitrator was of their own choosing and came from the tribe that had been their ally.  Ibn Ishaq records that "only one of their women was killed."  And the casualty figures come from a source written 120 to 150 years later with no independent check;  Ibn Ishaq himself offers a range from 600 to 900, which is a 50 per cent spread.  W. N. Arafat argued in the Journal of the Royal Asiatic Society in 1976 that the massacre account is unhistorical.  That article could not be obtained for this draft, so this article does not characterise its argument beyond naming it.

The conquests after Muhammad

The Arab conquests of the 630s and 640s took Syria, Egypt, Iraq and Iran within about twenty years.  What they were for is a separate question from how fast they went, and the earliest evidence is not what either polemic expects.

Robert Hoyland's In God's Path (Oxford, 2015) is the standard modern account.  On motive:

"The Arab conquerors do not seem to have expected or planned for this to happen.  God had ordained that the conquered people would be the Arabs' booty, not their equals.  Later Muslim historians maintained that the conquerors had offered their opponents the opportunity to convert before fighting them, but this is never mentioned in earlier sources.  As John of Fenek observed:  'Of each person they required only tribute, allowing him to remain in whatever faith he wished.'"¹⁶

John bar Penkaye was a Syriac Christian writing about 687, within living memory of the conquest, with no reason to flatter the conquerors.  His testimony is that they wanted tax, not converts.  That is one of the strongest pieces of evidence a Muslim apologist has, and it comes from a Christian source.

Two findings that cut against the crude Christian version

These two must be given at full strength, because they are the places where popular Christian apologetics is simply wrong on the facts.

Conversion was slow, not forced

Hoyland again, on the timing:

"In Egypt, where the Muslim presence was light for the first two centuries of Arab rule, conversion was very slow, and Islam only became the majority religion around the fourteenth century.  In Iraq and Khurasan, however, which bore the brunt of early Muslim settlement, opportunities for social intercourse were numerous and conversion correspondingly more frequent, with Muslims being a majority by the tenth century if not before."¹⁶

Read that again with the conquest dates beside it.  Egypt was conquered in 641.  It became majority Muslim around seven hundred years later.  Iraq was conquered in the 630s and reached a Muslim majority in the tenth century — roughly three centuries.  If military conquest converted populations, those gaps could not exist.  "Islam spread by the sword" is false as a demographic claim, and a Christian apologist who keeps using it is handing a Muslim opponent a free win.

The quantitative work behind this picture is Richard Bulliet's Conversion to Islam in the Medieval Period (Harvard, 1979).  That book could not be obtained for this draft, and its percentages are therefore not printed here.  What can be reported is his method and the size of his samples, because the Spanish Arabist Mayte Pénelas described both in detail, with page citations to Bulliet, in Al-Qantara in 2002.

Bulliet's idea was that biographical dictionaries record genealogies, that Muslims gave their children Muslim names, and that a genealogy whose earliest member has a Persian name therefore marks the generation that converted.  Working back from the death date recorded for each subject, he subtracted 70 years for an average lifespan, 34 years per generation, and added 25 as the likely age at conversion.⁵⁹  From nearly six thousand biographies of men from Nishapur and Isfahan he extracted 469 usable genealogies for Iran.⁵⁹  For Muslim Spain he had five dictionaries containing more than seven thousand biographies and got 154 usable sequences, and the resulting curve puts the midpoint of conversion at 350/961.⁵⁹

Pénelas' criticism is severe and should be given with the result.  One hundred and fifty-four individuals across more than four hundred years is a thin base.  Bulliet conceded as much himself:  Pénelas quotes him saying that for Spain "the result is obviously not a tidy bell-shaped curve of the type similar data produced for Iran," for which "the curve was made up from three times as many genealogies."⁵⁹  And the method's core assumption is admitted by Bulliet to be shaky:  it fails "if the initial convert changed his own name to an Arabic name, which was probably the normal practice."⁵⁹  Pierre Guichard has separately argued that Spanish chronicles of the late ninth and early tenth centuries do not fit Bulliet's picture of a Muslim majority arriving only after 950.⁵⁹

So the conversion-curve literature supports the shape of the argument — slow, generational, driven by social contact rather than by force — while its precise numbers are contested by the specialists.  Both halves belong in an honest account.

Jews fared better under Islam than under Christendom

Mark Cohen of Princeton is the historian who made the comparative case, in Under Crescent and Cross:  The Jews in the Middle Ages (Princeton, 1994).  That book was not obtainable either;  what follows is from Cohen's own restatement of the same argument in his prologue to A History of Jewish-Muslim Relations (Princeton, 2013), which was read in full.

Cohen begins by dismantling both myths.  The nineteenth-century picture of an interfaith utopia in Muslim Spain "was to a certain extent a myth;  it ignored, or left unmentioned, the legal inferiority of the Jews and periodic outbursts of violence."  And the twentieth-century reaction against it — the claim that Islam persecuted Jews relentlessly, "often as badly as medieval Christendom" — he names a "counter-myth of Islamic persecution," or a "neo-lachrymose conception of Jewish-Arab history."¹⁷

Then he states the comparison, and it is a comparison, not a compliment:

"Yet, when compared to the gloomier history of Jews in the medieval Ashkenazic world of Northern Europe and late medieval Spain, and the far more frequent and severe persecution in those regions, it contained a very large kernel of truth."¹⁷

His explanation has five parts, and every one of them is structural rather than sentimental:  religious exclusivity was diluted in Islam by the presence of several tolerated minorities rather than one;  Jews were not placed under an arbitrary special law manipulable by individual rulers;  they were not economically confined to despised trades like moneylending;  they were not socially excluded from the hierarchy;  and no Islamic equivalent developed of the medieval European "nationalism" that finally expelled them.¹⁷

Cohen is equally clear that this was not tolerance in the modern sense.  "If Islam seems to have been more tolerant than Christendom, this is true only in a qualified sense.  In the Middle Ages, tolerance, in the modern, liberal meaning of full equality, was not considered to be a virtue to be emulated."¹⁷  Jews and Christians under Islam were dhimmis, protected people:  they paid an annual poll tax and were subject in theory to the restrictions of the Pact of Umar, which barred new houses of worship and the repair of old ones, required humility in the presence of Muslims, required distinguishing dress, and excluded them from authority in government.¹⁷

And his verdict on how far those rules actually bit:

"De facto, however, these discriminatory regulations, most of them originating outside Islam, were largely honored in the breach, often with the tacit approval of Muslim rulers."¹⁷

He names the exceptions rather than hiding them:  the Almohad persecution in North Africa and Spain in the mid-twelfth century, which "destroyed entire Jewish communities" and forced thousands to accept Islam;  the destruction of houses of worship and forced conversions ordered by the Fatimid caliph al-Hakim in Egypt and Palestine in the early eleventh century;  and the killing of the vizier Joseph ibn Naghrela in Granada in 1066 and the massacre of the Jewish quarter that followed.¹⁷  His point is that these were exceptional, and that the Jews of Islam treated them as lapses of a system they otherwise trusted.

For a Christian apologist, this is the most uncomfortable evidence in the article, and it should be stated without softening.  A leading Jewish historian, writing about the treatment of his own people, concludes that they did better under Islamic rule than under Christian rule, and gives structural reasons for it.  Any Christian argument about Islam and violence that cannot survive that finding is not worth making.

Apostasy

The tension here is not manufactured by critics.  It is inside the sources, and the classical tradition felt it.

The hadith.  Sahih al-Bukhari 3017, narrated by Ikrima:

"'Ali burnt some people and this news reached Ibn 'Abbas, who said, 'Had I been in his place I would not have burnt them, as the Prophet said, "Don't punish (anybody) with Allah's Punishment."  No doubt, I would have killed them, for the Prophet said, "If somebody (a Muslim) discards his religion, kill him."'"¹⁸

The Arabic of the saying is man baddala dinahu fa-qtuluhu.  Sahih al-Bukhari 6922, in the chapter on apostates, carries the same report through a different chain.⁶⁰  Two things about the wording are worth noting, in both directions.  The words are quoted by Ibn Abbas, not narrated as a direct address by Muhammad to a gathering — which is a slightly weaker form of attestation than a direct narration, though still inside al-Bukhari.  And the occasion is a dispute about the method of execution, not about whether to execute:  Ibn Abbas objects to burning, not to killing.

The tension with Quran 2:256 is real, and al-Tabari's solution is the one that actually held.  He kept 2:256 un-abrogated and then defined its scope so that apostates fall outside it by name.¹¹  That is coherent.  It is also not a resolution a modern reader will find comforting, and pretending otherwise would be dishonest.

Modern Muslim scholars who reject the death penalty.  Several are frequently named:  Mahmoud Shaltut (d. 1963), Shaykh al-Azhar, who argued that unbelief alone does not justify killing and that the penalty attached to hiraba, armed sedition;  Taha Jabir al-Alwani, whose Apostasy in Islam (IIIT, 2011) argues the same from the Quran;  Yusuf al-Qaradawi, who distinguished "minor" from "major" apostasy;  and Ali Gomaa, then Grand Mufti of Egypt, who said in 2007 that Muslims are free to change their religion.  None of these works was read for this draft, so this article does not characterise their arguments beyond naming them, and a reader should be warned that several such positions retain the penalty for public apostasy judged seditious, which is a narrower concession than it first appears.

And the survey evidence, which shows how contested this is among ordinary Muslims.  Among Muslims who want sharia to be the law of the land, support for executing people who leave Islam ranges from 4 per cent in Kazakhstan to 86 per cent in Egypt.²¹  Whatever the texts say, the community does not speak with one voice about them.

The comparison: what the Bible commands

A Christian apologetics site has no standing to treat religious violence as an exotic Muslim problem.  The Bible contains commands more severe than anything in the Quran, and Christians have burned dissenters, sacked cities and blessed conquests.  This section handles both without evasion, because an argument that requires the reader not to check the Old Testament is not an argument.

The texts, printed rather than paraphrased

Quoted in the English Standard Version throughout, and read from the ESV module in the project source library rather than from memory or a web summary — the same rule this article applies to the Quran.⁶¹

• Deuteronomy 7:1-2 — seven named nations: "then you must devote them to complete destruction. You shall make no covenant with them and show no mercy to them."
• Deuteronomy 20:10-15 — the general law of war: "offer terms of peace to it"; if the city accepts, its people "shall do forced labor for you"; if it refuses, "you shall put all its males to the sword," and "the women and the little ones, the livestock, and everything else in the city, all its spoil, you shall take as plunder for yourselves."
• Deuteronomy 20:16-18 — the exception for the land itself: "But in the cities of these peoples that the LORD your God is giving you for an inheritance, you shall save alive nothing that breathes."
• Joshua 6:21; 8:24-26; 10:28-40; 11:11-14 — the execution of that command, city by city.
• 1 Samuel 15:2-3 — Amalek: "Do not spare them, but kill both man and woman, child and infant, ox and sheep, camel and donkey."

Those texts are worse than Quran 9:5 on the axis that matters most to a modern reader.  Quran 9:5 offers the polytheists an exit — repent and go free — and Quran 9:29 offers the People of the Book an exit that does not even require conversion.  Deuteronomy 20:16 offers no exit at all.  A Christian who does not concede that has not read his own book.

The structural differences, as testable claims

Four differences are real, and they are differences of scope rather than severity.

1. Bounded in space. The command names seven nations in one land. The same chapter gives a milder rule for every city "very far from you," and Deuteronomy 2:4-5, 2:9 and 2:19 forbid Israel to take land from Edom, Moab or Ammon. There is no general command to fight the nations.
2. Bounded in time. It attaches to one generation and one campaign. No later biblical text extends it.
3. Judicial rather than missionary. Genesis 15:16 makes the delay wait on "the iniquity of the Amorites" being "not yet complete," and Deuteronomy 9:4-6 denies explicitly that Israel's own righteousness is the ground. Nobody is offered conversion as an alternative to death and nobody is killed for refusing to worship the God of Israel. That is a real difference from a rule keyed to belief — and it is worse on one axis, not better, because it means no exit is offered.
4. Not a standing institution. Nothing in the Hebrew Bible or the New Testament creates a permanent legal category corresponding to jihad as a communal obligation (fard kifaya, a duty on the community that binds until enough people discharge it) binding until the end of the age, and nothing creates a permanent tributary status like dhimma. Classical Islamic law creates both. This is the strongest form of the Christian argument, and it is a claim about law, not about body counts.

And the New Testament's handling:  Matthew 5:38-48;  Matthew 26:52, "For all who take the sword will perish by the sword";  John 18:36, "My kingdom is not of this world.  If my kingdom were of this world, my servants would have been fighting, that I might not be delivered over to the Jews.  But my kingdom is not from the world";  Romans 12:17-21;  2 Corinthians 10:3-4.⁶¹  The New Testament never authorises a believer to use force to advance the faith.

The strongest Muslim reply, at full strength

"You object to a command to kill idolaters and to a permanent legal category of subjugated non-Muslims.  Your own scripture commands the extermination of entire populations including infants and livestock, offers those populations no option of surrender, and attributes the command directly to God — not to a prophet's judgement, not to a jurist's inference.  Islamic law at its harshest lets a Jew or a Christian live, keep his religion, run his own courts and pass his property to his heirs.  Deuteronomy 20:16 does not.  And you cannot appeal to the New Testament, because Jesus says he did not come to abolish the Law (Matthew 5:17);  what you are actually doing is abrogation, which is the very doctrine you say discredits us."

That is a serious argument and it should be answered rather than deflected.

The answers, and what they cost

On scope.  The conquest command has a terminus and the jihad obligation, in classical fiqh, does not.  This holds.  What it costs:  it is an argument about durability, not about severity, and it does nothing whatever for the infants of Jericho.

On the abrogation tu quoque.  Christians do claim a change of covenant, and that is a real structural parallel to naskh.  The difference a Christian can defend is the direction of travel.  The New Testament's supersession runs from a national and coercive order to a non-coercive one:  the later text is the gentler text.  The classical naskh chronology runs the other way, with Meccan patience cancelled by Medinan war:  the later text is the harsher text.  That is the best argument in this section.  What it costs:  it depends on accepting the classical naskh chronology, which the abrogation section above shows was contested — so a Christian who uses this argument has to grant al-Tabari's dissent when a Muslim invokes it.

On Matthew 5:17.  "Fulfil" is doing real work in that verse and this article will not slide past it.  A serious treatment belongs in a separate article on Old Testament warfare, which the Berean Archive does not yet have.

And the cost that has to be paid out loud.  A reader who finds "kill both man and woman, child and infant" morally intolerable will not be reassured by being told it was a one-off.  This is the hardest text in the Christian Bible.  Christians disagree about how to read it:  Richard Hess and Lawson Younger argue that the herem formula is ancient Near Eastern conquest hyperbole rather than literal head-count;  Paul Copan's Is God a Moral Monster? (2011) defends a related reading;  Nicholas Wolterstorff argues for a non-literal reading of Joshua.  None of those works was read for this draft, and other Christians reject all of them.  The Berean Archive has not solved this problem and does not claim to have.

The Crusades and the Inquisition

Two evasions are common here and both are refused.

"The Crusades were defensive."  Partly true and mostly not.  They were wars of aggression as well as of reconquest, preached by popes with indulgences attached.  The sack of Jerusalem in July 1099 was a massacre of the city's Muslim and Jewish inhabitants;  the traditional figure of 70,000 comes from Ibn al-Athir writing over a century later, and Benjamin Kedar's review of the numbers in Crusades 3 (2004) is the place to check them — it could not be obtained for this draft, so no figure is printed here.  The Fourth Crusade sacked Constantinople, a Christian city, in 1204.

"The Inquisition killed fewer people than everyone thinks."  Also partly true, and beside the point.  The Spanish Inquisition ran from 1478 to 1834 and executed people for heresy under state authority with church cooperation.  The revisionist trial-count work of Gustav Henningsen and Jaime Contreras, used by Henry Kamen, brought the totals down sharply;  neither was obtained for this draft, so no execution total is printed.  A lower number does not make a religious court that burns people acceptable.

To the list belong also the wars of religion, the burning of Michael Servetus by Calvin's Geneva in 1553, the persecution of Anabaptists by Catholics and magisterial Protestants alike, and the Christian blessing of colonial conquest.

Having conceded all of that, here is the point that survives it.  The useful comparison is not "which religion's adherents killed more people."  That is a question about states, demography and technology, not about texts, and it cannot be settled by counting.  The useful comparison is:  what does each body of scripture and law authorise, and what happens when a believer takes it most seriously?  A Christian who reads the Sermon on the Mount most seriously becomes a pacifist, and the Christian violence listed above happened in spite of the texts rather than because of them.  That last sentence is the weakest link in this article and it should not be waved through.  Christian violence did not proceed without texts.  Augustine read Luke 14:23, "compel them to come in," as warrant for coercing the Donatists, and that reading stood behind a thousand years of European religious compulsion.  Crusade preaching drew on Deuteronomy and on Maccabees.  Deuteronomy 20 was quoted at the conquest of the Americas.  What a Christian can defend is narrower:  that those uses required an argument the New Testament does not supply and that other Christians contested from within, whereas the jihad obligation is a settled category of classical law rather than a contested reading.  That is a real difference, and it is smaller than the sentence above makes it sound.  What a Muslim who reads the classical jurisprudence of jihad most seriously becomes is the question — and it has to be answered from the jurisprudence, not from the news.

The modern data, handled carefully

The suicide-attack statistic

The claim that circulates is:  "450 of 452 suicide terror attacks in 2015 were perpetrated by Muslim extremists."  It is real, it is traceable, and it is usable only with its caveats attached.

The sentence comes from Avi Issacharoff's report in The Times of Israel on 8 January 2016, summarising an interview with Yoram Schweitzer of the Institute for National Security Studies at Tel Aviv University.⁶²  The underlying study is INSS Insight No. 789 by Daria Schitrit, Einav Yogev and Yoram Schweitzer, published 26 January 2016, with an eleven-page country annex.⁶³  Both were obtained and read.

What the primary documents say:  "some 452 attacks carried out by some 735 male and female terrorists in 22 countries.  These attacks claimed the lives of approximately 4,370 people."⁶³  On motivation:  "With rare exceptions, the suicide attacks that took place in 2015 were committed by terrorist organizations associated with radical Islam."⁶³  And the two exceptions are identified by name in the country annex, both in Turkey:  "Two other suicide bombings were carried out by non-jihad organizations:  one by the Kurdish PKK underground, and one by the Marxist-Leninist DHKP/C party."¹⁹

Verdict:  the arithmetic checks.  452 attacks, two by non-jihadist organisations, 450 by jihadist ones.  The phrase "450 of 452" does not appear in INSS's own publications — it is Schweitzer's formulation in the interview — and should be attributed that way.

Six caveats, all of which must travel with the number.

1. It counts suicide attacks, not terrorism. Suicide bombing is one tactic, and it is overwhelmingly used by one movement. That is a fact about a tactic, not a measure of religiously motivated violence in general.
2. A multi-bomber operation counts as one attack. INSS says so: "a combined attack conducted simultaneously against a number of targets simultaneously is counted as a single event (e.g., the attack in Paris on November 13 was counted as one attack)."⁶³ Seven bombers, one attack.
3. The unit of attribution is the organisation, not the individual. "By Muslim extremists" means "by organisations identified with radical Islam." It is a claim about groups, not about what each attacker believed.
4. The base is 22 countries, nearly all active war zones — Iraq, Syria, Afghanistan, Pakistan, Nigeria, Cameroon, Chad, Niger, Yemen, Libya, Egypt, Somalia. The figure measures where the civil wars are as much as what motivates them.
5. One year is not a trend. 2015 had 452; 2014 had 592.⁶³
6. INSS is an Israeli national-security institute. Its counting rules are conservative and published, but it should be named for what it is rather than called "a study."

Usable, with all six attached, and never as a proxy for religiously motivated terrorism in general.

Pape's occupation thesis, and why it does not settle the question either

The strongest counter-explanation is Robert Pape's:  that suicide terrorism is a strategic response to foreign military occupation rather than a religious phenomenon.  Pape's 2003 paper in the American Political Science Review is paywalled and no open copy was found, so it is quoted here only through the direct quotations his critics reproduce, and that limitation is stated rather than hidden.

Pape collected "information on every suicide attack between 1980 and 2001 (188 cases in all)," and argued that all but ten "sought to persuade a democracy to remove its troops from the home soil of the terrorist organization."  His headline evidence against the religious explanation is that the Marxist-Leninist Tamil Tigers of Sri Lanka carried out more suicide attacks than any other organisation — 75 of the 188.⁶⁴  His conclusion, quoted verbatim by his critics:

"Perhaps most important, the close association between foreign military occupations and the growth of suicide terrorist movements in the occupied regions should give pause to those who favor solutions that involve conquering countries in order to transform their political systems."⁶⁴

And here is the methodological finding, which cuts against both sides equally.  Scott Ashworth, Joshua Clinton, Adam Meirowitz and Kristopher Ramsay showed in the same journal in 2008 that Pape sampled on the dependent variable:  his dataset contains only cases where suicide terrorism occurred, so it cannot tell you how likely suicide terrorism is given occupation, or given religious extremism.  Computing the bounds the data actually permit, they find that the risk attributable to occupation could be "any value between −0.957 and 0.944," and that for religious extremism the bounds run from −0.973 to 0.981 — so wide that "Pape's data do not restrict the feasible values of the risk of suicide terrorism attributable to religious extremism in any meaningful sense."⁶⁴

So Pape's data cannot show that occupation drives suicide terrorism, and the same data cannot show that religion drives it.  Anyone quoting Pape to prove "it's occupation, not religion" is quoting a study whose design cannot support that conclusion.  Anyone quoting the 2015 INSS numbers to prove "it's religion" is doing the same thing in reverse.

One fact from Pape's dataset should be carried in any case, because it is true and often denied:  the Tamil Tigers were a Hindu-background Marxist-Leninist organisation and account for 40 per cent of suicide attacks between 1980 and 2001.  "Suicide bombing is a Muslim phenomenon" is false as a statement about the twentieth century, and became overwhelmingly true only in the 2010s, after the Tigers were destroyed in 2009.

What the organisations say to their own side is a separate line of evidence, and it exists in translation.  Raymond Ibrahim, working in the Near East section of the Library of Congress, translated al-Qaeda's Arabic theological treatises and printed them beside its messages to the West in The Al Qaeda Reader (2007).  His finding was that the two genres give different reasons.  In the messages to the West, "bin Laden and Zawahiri insist that they are waging a 'Defensive Jihad' against an oppressive West.  When discussing the tenets of Islam, however, they argue to Muslims that Muslims should battle the West because it is the infidel, or the 'Great Satan.'"⁶⁵  The clearest document is al-Qaeda's reply to a 2002 open letter from Saudi intellectuals titled "How We Can Coexist."  It quotes the Buraydah hadith printed above — Islam, the jizya, or fighting — and draws its conclusion:

"Does Islam, or does it not, force people by the power of the sword to submit to its authority corporeally if not spiritually?  Yes.  There are only three choices in Islam:  either willing submission;  or payment of the jizya, through physical though not spiritual, submission to the authority of Islam;  or the sword—for it is not right to let him [an infidel] live.  The matter is summed up for every person alive:  Either submit, or live under the suzerainty of Islam, or die."⁶⁵

Ibrahim then observes that al-Qaeda's own letter to Americans, "Why We Are Fighting You," written in the same exchange, "does not mention those many Islamic doctrines delineated in al-Qaeda's declaration to the Saudis that intrinsically require Muslims to attack non-Muslims."⁶⁵  Three cautions belong here.  The attribution of the Saudi reply to bin Laden personally is Ibrahim's inference from four indicators, which he states;  the translations are his alone;  and Ibrahim has since written polemical histories that this article does not use.  What the documents establish is narrower than he claims and still matters for the question above:  an organisation's grievance narrative for Western audiences is not evidence for the occupation thesis when the same organisation tells its own audience that the three choices apply to "every person alive."

What Muslims actually say: the Pew figures

Pew Research Center surveyed more than 38,000 Muslims face to face in over 80 languages across 39 countries and published the results in April 2013.²¹  Saudi Arabia, Iran, China and India were not surveyed, which matters.

Suicide bombing "in defense of Islam" often or sometimes justified, per cent of all Muslims:²¹

Pew's own summary:  "In most of the 21 countries where the question was asked few Muslims endorse suicide bombing and other forms of violence against civilian targets as a means of defending Islam against its enemies.  But in a few countries, substantial minorities believe suicide bombing can be often justified or sometimes justified."²¹

Both readings of this table are available and both are partly right.  Single-digit figures across most of the sample are genuinely low.  Forty per cent in one place and thirty-nine in another are genuinely alarming.  Quoting either end alone is a misreport, which is why the whole table is printed here.

The same survey asked whether sharia should be the official law of the land.  The answers ran from 99 per cent in Afghanistan and 91 per cent in Iraq to 12 per cent in Turkey, 10 per cent in Kazakhstan and 8 per cent in Azerbaijan.²¹  The spread is the finding.  Any single global sentence about "what Muslims believe" is false before it is finished.

One arithmetic warning, because this number is misreported constantly in both directions.  The apostasy figures quoted earlier — 4 per cent in Kazakhstan to 86 per cent in Egypt — are percentages of those who already favour sharia as state law, not of all Muslims.²¹  To get a share of a country's Muslims you have to multiply:  Egypt is 74 per cent × 86 per cent ≈ 64 per cent;  Turkey is 12 per cent × 17 per cent ≈ 2 per cent;  Kazakhstan is 10 per cent × 4 per cent ≈ 0.4 per cent.  Pew does not print those products, so they are this article's arithmetic and are labelled as such.

The general terrorism data, which look different

Suicide bombing and terrorism are not the same dataset, and they do not give the same picture.

The University of Maryland's START consortium counted "nearly 8,500 terrorist attacks around the world" in 2019, killing more than 20,300 people — down about half from the 2014 peak of nearly 17,000 attacks.²⁰  Nine organisations carried out more than 100 attacks each:²⁰

Three of the nine were Maoist, with 564 attacks between them.  START also records that racially and ethnically motivated attacks in Australasia, North America and Western Europe killed at least 86 people in 2019, up from 52 in 2018.²⁰

Suicide bombing in 2015 was about 99.6 per cent jihadist.  Terrorist attacks in 2019 were not.  Any claim about "religiously motivated terrorism" has to name which dataset it is using and what that dataset counts, because the two give different pictures of the same world.

The strongest Muslim responses

Five, in order of force, each answered.

1.  "Naskh applies to law, not to narrative or to moral principle."  This is not a modern invention.  Al-Tabari says it, and Ibn Kathir quotes him saying it on 2:106:  "The Nasakh only occurs with commandments, prohibitions, permissions, and so forth.  As for stories, they do not undergo Nasakh."²⁷  Answer:  true, and it disposes of the crudest version of the sword-verse claim — the Quran's narrative and ethical material was never a candidate for abrogation.  But 2:190, 8:61, 47:4 and 60:8 are legal rulings about warfare, which is exactly the category naskh was built for.  The restriction does not protect them.

2.  "What the commentators called naskh was mostly specification, not abrogation."  This is Powers' own explanation of why the counts exploded, and it is al-Tabari's stated method:  an abrogator is not an abrogator unless it negates the abrogated ruling entirely, and a command general in wording but specific in reference is not abrogated at all.⁹  Answer:  this is the best argument on the Muslim side and it is largely right about the history of the doctrine.  It does not, however, tell a jurist which verses fall on which side, and the classical manuals answered that question the other way.

3.  "The maximalist count is a minority even inside the classical tradition, and it was falling for centuries."  238 in Ibn Salama, 20 in al-Suyuti, 5 in Shah Wali Allah.⁹  Answer:  the trend is real.  But Fatoohi, who rejects abrogation, is the witness against reading too much into it:  "The overwhelming majority of Muslim scholars, past and modern, have accepted abrogation in both the Qur'an and the Sunna as an indisputable fact."⁵⁷  A shrinking list is not a rejected doctrine.

4.  "The sword verse is itself abrogated, by its own final clause."  Powers reports this from the classical literature, and the abrogations attributed to it are of sense, not of wording.⁹  Answer:  correct, and it should be quoted whenever someone claims 9:5 stands alone and unqualified.  What it establishes is that the polytheists had an exit — conversion.  It does not establish that they had any other exit.

5.  "Al-Tabari refused the abrogation claim, verse by verse."  At 2:190, 2:256, 8:61 and 60:8.¹¹  Answer:  this is the strongest response in the set, because it requires the reader to accept nothing modern and no Western scholarship.  Its limit is that al-Tabari's own positive readings are not liberal ones.  He defends 2:256 by narrowing it to those from whom the poll tax may be taken, excluding Arab pagans and apostates by name;  he defends 8:61 by confining 9:5 to idolaters;  and he glosses the command in 2:190 as fighting until submission or the jizya "in humiliation."  Al-Tabari is a witness against the sword-verse thesis.  He is not a witness for religious liberty.

Common objections and responses

"Jihad means inner struggle, not war."

It means both, and the Quran settles which one is obligatory.  Quran 2:216 uses kutiba alaykum al-qital — "fighting has been written upon you" — the same construction used for the Ramadan fast.⁵  The classical law books have chapters titled kitab al-jihad, and they are about war:  who must fight, how spoils are divided, how truces are made.  The inner-struggle sense is genuine, and Jami' at-Tirmidhi 1621 attests it in a chapter al-Tirmidhi graded hasan sahih.⁵⁴  But the famous "greater jihad" saying that is used to prove the point has been graded weak or worse by every hadith critic who examined it.⁵³

"Quran 5:32 says killing one person is like killing all mankind."

The verse says that, and it is worth quoting for what it is.  But it opens "We decreed upon the Children of Israel," and it is a citation of Jewish law — the Mishnah, Sanhedrin 4:5 — rather than a statement of Islamic law.  The next verse, 5:33, prescribes execution, crucifixion, amputation or exile for those who "wage war against Allah and His Messenger."  Quoting 32 without 33 is the mirror image of quoting 9:5 without 9:6, and it fails for the same reason.

"Quran 8:12 orders Muslims to strike the necks of unbelievers."

It does not.  The verse's own opening clause addresses the angels:  "when your Lord inspired to the angels…"³⁹  This is one of the most frequently repeated errors in Christian polemic and it is refutable by reading one sentence.  Drop it.

"The Banu Qurayza massacre is a later invention."

The numbers may well be inflated;  Ibn Ishaq's own range of 600 to 900 shows the tradition was uncertain.⁵⁸  But the arbitration that produced the executions is in Sahih al-Bukhari 4121 and Sahih Muslim 1769, which is the material Sunni Muslims treat as soundest.¹⁵  Dispute the count if the evidence supports it.  The episode itself is inside the canon.

"Islam spread by the sword."

Not demographically.  Iraq and Khurasan reached a Muslim majority only in the tenth century and Egypt only around the fourteenth.¹⁶  The conquests spread Muslim rule quickly and Muslim belief slowly, and the earliest witness to what the conquerors demanded is a Christian one who says they wanted tribute.¹⁶

"Christians did the same thing, so the comparison is a wash."

It is not a wash, and the reason is not body counts.  Christians who sacked Jerusalem in 1099 or burned Servetus in 1553 were acting against their own foundational texts, and the texts were available to be quoted at them — as Erasmus, the Anabaptists and Sebastian Castellio in fact did.  The question this article puts is narrower and harder:  what does each body of law authorise for a believer who follows it most carefully?  Answering that requires reading the fiqh, not the newspaper.

"Most Muslims are peaceful, so the texts must be peaceful."

This does not follow, and the reverse does not follow either.  Most Christians are peaceful and Deuteronomy 20 still says what it says.  Adherents are not a reliable readout of a text, in either direction — which is precisely why this article is about the sources.

"Taqiyya means Muslims are permitted to lie to you."

Not as the classical commentators state it.  Taqiyya is the Quran's own word at 3:28, which forbids taking unbelievers as allies "unless you fear from them a precaution" — illa an tattaqu minhum tuqat.  Al-Tabari's gloss on that clause is:  "unless you are under their authority and fear them for yourselves, so that you show them allegiance with your tongues while harbouring enmity toward them — and do not go along with them in their unbelief, and do not help them against a Muslim by deed."  The early reports he collects say the same.  Ibn Abbas:  "taqiyya is speaking with the tongue while his heart is at rest in faith."  Ikrima:  "so long as he does not shed a Muslim's blood or make his property lawful."  Abu al-Aliya:  "taqiyya is with the tongue, not with deeds."  Al-Tabari's ruling is that the verse means "unless you fear from them a fear," and that it is taqiyya from unbelievers, not from anyone else.⁶⁶  Ibn Kathir says the same:  believers "who in some areas or times fear for their safety from the disbelievers" may "show friendship to the disbelievers outwardly, but never inwardly," and he quotes Abu al-Darda, "We smile in the face of some people although our hearts curse them," and al-Hasan, "The Tuqyah is allowed until the Day of Resurrection."⁶⁷

So the doctrine is real, it is Sunni as well as Shia, and it is conditioned on being under unbelievers' power and in fear of harm.  The strongest Christian-side statement, Raymond Ibrahim's 2010 essay in Middle East Quarterly, cites exactly these two commentators, translates al-Tabari accurately, and is right that the doctrine is not a Shia peculiarity.⁶⁸  Where it overreaches is the next step.  It calls bin Laden's grievance messages to the West "clear instances of taqiyya" and folds in the hadith "War is deceit."⁶⁹  That hadith is about stratagem against an enemy in the field, and al-Bukhari's chapter is titled for it;  it is not the 3:28 doctrine, and neither text licenses lying to a non-Muslim who holds no power over the speaker.  The usable claim is that classical law permits outward conformity under threat and deception against an enemy in war.  The conspiracy version — that any reassurance from any Muslim may be a lie — has no source in either text, and it is the one Christians should drop.

What evidence would change the conclusion?

What would strengthen the Muslim claim.  A demonstration from the early Arabic manuscript record that the abrogation manuals misrepresent what jurists actually practised.  Evidence that the sword-verse thesis was a minority position in working fiqh rather than in commentary, which would require systematic work in the legal literature rather than the tafsir.  A body of early legal texts showing that dhimma was framed as a temporary conquest arrangement rather than a permanent status.  Or a modern consensus among the major traditional institutions — al-Azhar, the Islamic Fiqh Academy — that offensive jihad and the apostasy penalty are abrogated in the other direction.  Each of these is checkable, and none is ruled out in advance.

What would weaken it.  Publication of the Arabic of Ibn Salama, al-Nahhas and Ibn al-Ata'iqi in a form that shows their lists were the working reference for jurists rather than a genre exercise.  Evidence that al-Tabari's dissent had no followers.  Or documentary evidence from the conquest period — treaty texts, tax registers — showing conversion demanded at the point of the sword, which is exactly what the earliest sources so far do not show.

What would change this article's method rather than its verdict.  Obtaining Bulliet's Conversion to Islam, Cohen's Under Crescent and Cross in full, Burton's The Sources of Islamic Law, Pape's own 2003 paper, W. N. Arafat's 1976 article on the Banu Qurayza, and the modern Muslim literature on apostasy — Shaltut, al-Alwani, Gomaa.  All six were sought for this draft and none was reached.  Where they bear on a claim, the claim above is stated more cautiously than it might otherwise be.

What must not be claimed

By the Christian side

1. Do not say the Quran contains no verses restraining warfare. 2:190, 2:192, 4:90, 8:61, 9:4, 9:6, 9:7 and 60:8 are all there.
2. Do not quote 9:5 without 9:4, 9:6 and 9:7, or 4:89 without 4:90-91. Both elisions are checkable in seconds and both destroy credibility on contact.
3. Do not present 8:12 as a command to Muslims. It addresses angels.
4. Do not say the classical tradition unanimously held that 9:5 abrogated the peaceful verses. Al-Tabari argued against it, in print, in the ninth century, four separate times.
5. Do not call the "greater jihad" hadith a modern Sufi fabrication. It has chains, it is at least as old as al-Bayhaqi, and its content is paralleled in material al-Tirmidhi graded hasan sahih. Say "no hadith critic has graded it sound," which is true and sufficient.
6. Do not use "Islam spread by the sword" as a demographic claim.
7. Do not use the "450 of 452" statistic without its six caveats, and never extend it from suicide attacks to terrorism generally.
8. Do not treat the Old Testament conquest as a solved problem. It is the hardest text in the Christian Bible and Christians disagree about it.
9. Do not generalise from texts to 1.9 billion people. The Pew tables run from 1 per cent to 40 per cent on the same question.
10. Do not print numbers from books nobody has opened — Bulliet's percentages, the Inquisition's execution total, Kedar's Jerusalem figure.
11. Do not say taqiyya licenses Muslims to lie to non-Muslims at will. Al-Tabari and Ibn Kathir condition it on being under unbelievers' power and fearing harm, and "war is deceit" is a different doctrine, about enemies in the field.

By the Muslim side

1. Do not say jihad means only inner struggle. Qital — fighting — is what 2:216 makes obligatory, and the fiqh chapters titled kitab al-jihad are about war.
2. Do not present "no compulsion in religion" as an unrestricted classical principle. Al-Tabari, defending it against the abrogationists, restricted it to those from whom the poll tax may be accepted and excluded Arab pagans and apostates by name.
3. Do not say abrogation is a fringe doctrine. Fatoohi, who rejects it, writes that the overwhelming majority of Muslim scholars past and modern have accepted it as an indisputable fact.
4. Do not present the "greater jihad" hadith as a saying of the Prophet. Ten graders across eight centuries, including al-Suyuti and Ibn Rajab, call it weak or worse.
5. Do not say the poll tax carried no humiliation. Al-Tabari glosses wa-hum saghirun as "while they are abased and subdued" and reports Ikrima's gloss that the collector sits while the payer stands.
6. Do not say the killing of the Banu Qurayza men is a late invention. The arbitration is in Bukhari 4121 and Muslim 1769.
7. Do not cite Pape to prove that religion plays no role. Ashworth and his colleagues showed his design cannot support any conclusion about religion — in either direction.
8. Do not say Muhammad never fought except when attacked. Al-Waqidi's own summary has him leading twenty-seven expeditions and fighting in nine, and Tabuk was a march toward Byzantine territory by an army no one had attacked.
9. Do not use the Pact of Umar's late date to imply that dhimmi restrictions were never enforced. Cohen, who says they were largely honoured in the breach, also names the Almohads, al-Hakim and Granada in 1066.
10. Do not answer the Old Testament comparison with "so you admit it." A Christian who concedes that Deuteronomy 20 is hard has not thereby conceded that jihad is defensive.

Related articles in this series:  where the Quran came from, which hadith count as canon and how that was decided, and whether Muhammad's life looks like a prophet's or a founder's.

Sources

Every Quran quotation was taken as raw JSON from the api.quran.com/api/v4 endpoint in August 2026 and checked against quran.com, not through any summarising fetcher.  Every hadith was parsed out of sunnah.com's own markup.  Where a source could not be opened, the footnote says so.  The translation this article calls Sahih International spells its own name "Saheeh International" on quran.com;  the house spelling is used here for consistency across the series.