Is the Bible corrupted?

Summary

Ask a Muslim why the Bible disagrees with the Quran about Jesus and the usual answer is one word:  tahrif, corruption.  Jews and Christians altered the books God gave them.  This article tests that claim against the Quran's own wording, against what Muslim scholars actually said for the first four centuries of Islam, and against the physical manuscripts.

• The Quran never says the biblical text was rewritten. Four verses use a form of the Arabic verb harrafa — Quran 2:75, 4:46, 5:13 and 5:41 — and the phrase is yuharrifuna al-kalima 'an mawadi'ihi, "they shift the words out of their places."¹ Gabriel Said Reynolds, professor of Islamic studies at Notre Dame: "the Qur'an intends scriptural falsification that involves reading or explaining scripture out of context, not erasing words and rewriting them."¹ The noun tahrif itself does not occur in the Quran at all.
• All four of those verses are aimed at Jews, never at Christians.¹ Of Christians the Quran says only that they "forgot a portion" of what they were told (Quran 5:14).
• A much larger group of verses treats the Torah and the Gospel as present, intact and usable in Muhammad's own lifetime. Quran 5:43 asks the Jews of Medina why they come to Muhammad for judgement "while they have the Torah, in which is the judgement of Allah."² Quran 5:47 commands Christians to "judge by what Allah has revealed therein."³ Quran 10:94 tells Muhammad himself that if he doubts, he should "ask those who have been reading the Scripture before you."⁴ Reynolds again, in his 2018 Yale commentary: "neither here nor elsewhere … does the Qur'an argue that the Bible of the Jews and Christians is falsified (5:47 and 10:94 actually suggest that the Bible is reliable)."⁵
• The charge that the text was corrupted is a later development inside Islam, and the best evidence for that is Islamic. Sahih al-Bukhari's chapter heading in the Book of Divine Unity quotes Ibn Abbas (d. 68/687) glossing the word: "'They distort' — [it means] they displace; and no one displaces the wording of a book from among the books of God, mighty and majestic — rather they distort it, interpreting it otherwise than its true interpretation."⁶ Camilla Adang's study of nine medieval Muslim writers on the Hebrew Bible found that the majority held only the sense had been changed; "Only al-Maqdisi and Ibn Hazm believed that the text itself had suffered distortion."⁷
• The tradition is genuinely split, and this article says so. The same Bukhari collection has Ibn Abbas saying the opposite two books away: "the people of the scripture changed their scripture and distorted it, and wrote the scripture with their own hands."⁸ Anyone who tells you early Islam was unanimous in either direction is mining one side.
• The decisive physical fact: the Bible Christians read today is the Bible that existed centuries before Muhammad was born. The Great Isaiah Scroll was copied around 125 BC.⁹ Codex Sinaiticus, a complete Greek Bible, was written in the middle of the fourth century.¹⁰ Compare the first-century-BC scroll 1QIsa-b against the medieval Hebrew codex used for modern translations — a thousand years apart, forty-six chapters of Isaiah — and every difference "concern[s] minutiae," with one arguable exception at Isaiah 53:11.⁹
• That forces the Muslim claim to name a date, and both available dates cost something. If the corruption happened after Muhammad, the manuscripts refute it, and Ibn Taymiyya (d. 728/1328) said so himself: "Whoever says that, after the Prophet, all copies have been corrupted, he has said what is manifestly false."¹¹ If it happened before Muhammad, then the Quran calls a corrupt book "guidance and light" (Quran 5:46) and orders Christians to judge by it (Quran 5:47).
• The Bible does have variants, and the honest numbers are large. About 500,000 non-spelling variants across roughly 5,300 Greek New Testament manuscripts.¹²¹³ But at least 75% are spelling slips and nonsense readings, fewer than 1% are both meaningful and viable,¹⁴ and about forty verses are genuinely doubtful, every one flagged in modern Bibles. No manuscript of any Gospel, in any language, at any date, omits the crucifixion or puts someone else on the cross. There is no variant there to argue about.
• The Gospel of Barnabas is not a candidate. It survives in two manuscripts, neither older than the sixteenth century, and its Jesus describes a Jubilee "which now cometh every hundred years" — a custom that existed only between AD 1300 and 1350, in Latin Christendom.¹⁵ It also makes Muhammad the Messiah, which the Quran denies.

Verdict:  the charge that Jews and Christians rewrote their scriptures is not supported by the Quran, by the earliest Muslim exegesis, or by the manuscripts.  What the Quran charges is misuse — hiding, forgetting, twisting in the reading.  That is a theological accusation, not a textual one, and it has to be argued on theological ground.  The textual version of the charge is a medieval Islamic development that reached full form with Ibn Hazm (d. 456/1064), and it collides with objects that can be examined in Jerusalem, London and the Vatican.

• Where this article's own case is weak, stated up front. The Quran does not say in so many words that the biblical text is corrupt — which cuts both ways. It means Christians cannot claim the Quran endorses their Bible outright, and it means quoting Quran 5:47 at a Muslim is not the checkmate popular apologetics treats it as. The standard Muslim reply is that the true Injil — the Gospel God gave to Jesus — was lost, and that the four Gospels are human reports about Jesus rather than revelation to him. No manuscript can refute that position, because it makes no claim about any manuscript. It can only be pressed by asking what evidence there is that such a book ever existed. The related claim that prophecies of Muhammad were deleted from the Bible is, as usually stated, unfalsifiable — which is a reason to set it aside, not a proof that it is false.

The claim being tested

Tahrif is the Arabic noun built from the verb harrafa, "to turn aside" or "to shift."  In Islamic usage it means the corruption of an earlier scripture.  Muslim scholars split it into two kinds, and the whole argument turns on the difference:

• Tahrif al-lafz — corruption of the wording. Words physically erased, added, or rewritten on the page. Some sources call this tahrif al-nass, corruption of the text.
• Tahrif al-ma'na — corruption of the meaning. The words on the page are the ones God gave, but the readers hide parts of them, misread them, or refuse to see what is there.

Camilla Adang, whose Muslim Writers on Judaism and the Hebrew Bible is the standard survey, puts both on the table in one sentence:  "From a very early period onwards, the accusation of tahrif, whether understood as a distortion of the biblical text (tahrif al-nass) or of its sense (tahrif al-ma'ani) came up in disputations between Muslims and the People of the Book."⁷

Popular argument blurs four separate propositions.  This article separates them, because they are refuted or supported by completely different kinds of evidence:

1. The Quran teaches that the biblical text was rewritten. A question about Arabic words in a seventh-century book.
2. The biblical text was in fact rewritten. A question about manuscripts.
3. The rewriting happened before Muhammad, so the Bible he encountered was already corrupt. A question about dates.
4. The book Christians call the Gospel is not the Injil at all — the real one was lost, and the four Gospels are human writing about Jesus rather than God's revelation to him. Not a textual claim at all, and no manuscript touches it.

Proposition 4 is the serious one.  Propositions 1 through 3 are the ones usually argued, and they are the ones this article shows do not hold.  A reader who wants only the strongest Muslim position should skip to the strongest Muslim responses.

Two related questions are handled elsewhere.  Whether Jesus died on the cross is treated in the Quran and the crucifixion.  Whether the Bible predicts Muhammad is treated in Muhammad in the Bible.  The manuscript figures used below come from the Berean audits in New Testament manuscripts and variants and which Bible text gives the right date for creation, which deliberately adopt the datings least favourable to the Christian side.

The strongest case for the Muslim claim

Stated at full strength, without answers.  A Muslim reader should recognise this section as fair.

The Quran does accuse.  Six verses say something went wrong with the earlier scriptures, and they are not gentle.  Quran 2:75 says "a party of them used to hear the words of Allah and then distort it after they had understood it while they were knowing."¹⁶  Quran 2:79 is harsher:  "So woe to those who write the 'scripture' with their own hands, then say, 'This is from Allah,' in order to exchange it for a small price."¹⁷  Quran 3:78 describes "a party who alter the Scripture with their tongues so you may think it is from the Scripture, but it is not from the Scripture."¹⁸  A Muslim who reads those three verses together and concludes that human hands produced a false book has not strained the text.

The Quran also says its own message is guarded in a way the earlier ones were not.  Quran 15:9 has God saying "indeed, it is We who sent down the message and indeed, We will be its guardian."  No comparable guarantee is given for the Torah or the Gospel.  If God had promised to preserve them, the argument runs, the promise would be on record.

The Bible is demonstrably a human artefact in ways the Quran is not.  The four Gospels are four different books by four named or traditionally named authors, written decades after the events, disagreeing in detail, and preserved in manuscripts that differ from one another in about half a million places.  Nothing in that description matches what the Quran says the Injil was:  a single revelation "sent down" to Jesus, as the Quran was sent down to Muhammad (Quran 5:46, 57:27).  Martin Whittingham, director of the Centre for Muslim-Christian Studies at Oxford, states the mismatch without polemic:  "Since the canonical New Testament gospels do not seem to fit the Muslim expectation of what the Injil should be like, Muslim scholars have proposed the existence of a pure, original Injil which is partly or sometimes barely preserved in the New Testament gospels."¹⁹

The internal Christian evidence supports at least some tampering.  Christian scholars themselves agree that Mark 16:9–20 and John 7:53–8:11 were added by later copyists, that 1 John 5:7 in the King James Version is a Latin intrusion, and that scribes sometimes changed readings for doctrinal reasons.  A Muslim who says "your own textual critics admit your text was altered" is quoting real scholarship, not inventing it.

And Christians did lose books.  Paul mentions a letter to the Laodiceans (Colossians 4:16) and an earlier letter to Corinth (1 Corinthians 5:9) that no one has.  If apostolic writings could vanish, a Gospel given to Jesus could vanish.

Finally, the strongest form of the claim does not depend on any of this.  Ibn Taymiyya drew the line that most thoughtful Muslims still draw, and it makes the manuscript evidence beside the point:

"What is in Injil of stories about the crucifixion of Jesus, his death, his coming to the disciples after he was 'raised,' are not what Jesus said.  Rather they are [reports] from those who saw those things after him.  What God revealed was what was heard [directly] from Jesus who is the transmitter [of God's word] from Him."¹¹

On that view the Gospels can be perfectly transmitted and still not be scripture.  They are eyewitness memoirs.  Muslims are not obliged to accept an eyewitness memoir over a later revelation, and no fourth-century codex changes that.

What the Quran actually says

Two sets of verses, both quoted in full.  Unless marked otherwise the English is Saheeh International, the Muslim-produced translation served by default at quran.com.  Its square brackets are the translators' additions and are kept here rather than silently deleted;  the Arabic does not contain them.²⁰

The verses that accuse

Three things about that list decide most of the argument.

First, the noun tahrif never appears in the Quran.  What appears is the verb, four times, and each time in the same fixed phrase.¹

Second, the phrase is spatial, not textual.  Mawadi' is the plural of mawdi', "place."  It is not the word for "meaning," which would be mawadu'.  Reynolds works through the vocabulary:  "one might begin with the literal sense of mawadi' (sing. mawdi') 'places' (and not 'meanings' …) and the primary meaning of the root h-r-f 'to move;  to turn'."¹  Classical Arabic does use the related noun harf for "letter," which is what suggested to medieval scholars that the verse was about altering letters — but that is a secondary sense, and the only time harf occurs in the Quran itself (Quran 22:11) it carries the primary sense of "edge."¹  His conclusion:  "there is no compelling reason to associate Qur'anic tahrif with an alteration of letters.  Instead, the phrase yuharrifuna l-kalima 'an mawadi'ihi seems to involve turning or shifting words out of their places or contexts."¹

The four English translations in the comparison set do not agree on how to render it, which is itself the point.  At Quran 4:46, Saheeh International has "distort words from their [proper] places [i.e., usages]";  Pickthall (1930) has "change words from their context";  Yusuf Ali (1934) has "displace words from their (right) places";  Abdel Haleem (Oxford, 2004) has "distort the meaning of [revealed] words."²¹  Three of the four render it as displacement or context, not rewriting.  No case about a physical text should rest on one English rendering.

Third, look at the examples the Quran itself gives.  Quran 4:46 does not describe scribes at a desk.  It describes people speaking:  they say "We hear and disobey," they twist ra'ina — "look at us" — into an insult with their tongues.  Abdel Haleem's rendering makes the scene plain:  "they say, 'We hear and disobey,' and 'Listen,' [adding the insult] 'May you not hear,' and 'Raina [Look at us],' twisting it abusively with their tongues."²¹  That is mockery in a conversation.  Nothing in it is about copying manuscripts.

Reynolds surveys the eight different verbs the Quran uses across all its complaints about scripture — concealing, forgetting, exchanging, writing falsely, twisting the tongue, shifting words — and concludes:

"Evidently the Qur'an is principally concerned with the misuse of scripture.  In none of these examples does the Qur'an insist that passages of the Bible have been rewritten or that books of the Bible have been destroyed and replaced by false scripture.  Instead, the Qur'an argues that revelation has been ignored, misread, forgotten, or hidden."¹

And a fourth point, which surprises most readers.  All four harrafa verses are aimed at Jews.  None is aimed at Christians.¹  The nearest thing to a parallel charge against Christians is Quran 5:14, that they "forgot a portion of that of which they were reminded" — and that verse comes immediately after Quran 5:13, which charges Jews with both forgetting and shifting words.  Reynolds:  "the Qur'an seems to excuse the Nasara from this latter charge."¹  Nasara is the Quran's word for Christians.  So the verse most often quoted at Christians about corrupting the Bible is not, on its face, about Christians at all.

The verses that treat the Torah and Gospel as present and usable

These are more numerous, and they are the ones popular argument leaves out.

Three of these do the real work.

Quran 5:43 asks the Jews of Medina, in the 620s, why they bring their disputes to Muhammad "while they have the Torah, in which is the judgement of Allah."  Abdel Haleem:  "but why do they come to you for judgement when they have the Torah with God's judgement."²  Abdullah Saeed, Sultan of Oman Professor of Arab and Islamic Studies at Melbourne and a Muslim scholar, reads it the obvious way:  the verse is "an obvious reference to Tawrat, which was indeed what existed with the Jewish community of Medina at the time of the Prophet."¹¹  Not a Torah in heaven.  The scrolls in the synagogue down the road.

Quran 5:47 is a command, in the imperative, addressed to Christians alive at the time:  "let the People of the Gospel judge by what Allah has revealed therein."  Abdel Haleem:  "So let the followers of the Gospel judge according to what God has sent down in it."³  Reynolds:  "if the Qur'an speaks against certain Christians, it speaks in support of the Gospel, and moreover assumes that the valid Christian revelation is still at hand in its day."¹

Quran 10:94 is the strangest of the three, because of who is being addressed.  It tells Muhammad that if he is in doubt about what has been revealed to him, he should "ask those who have been reading the Scripture before you."  Pickthall:  "then question those who read the Scripture (that was) before thee."⁴  Nobody settles a doubt by consulting a forgery.

A warning to Christians about two verses they overuse.  Quran 6:115 and 18:27 both say la mubaddila li-kalimatihi, "none can alter His words," and apologists reach for them constantly as proof that God preserved the Bible.  That is a bad argument, and a competent Muslim will say so.  Saqib Hussain, reviewing Whittingham's history of Muslim views of the Bible for the Review of Qur'anic Research, makes the objection cleanly:  "In context, the verses cited have little to do with scriptural preservation, and refer instead to God's threats and promises of punishment, reward, guidance, and misguidance being fulfilled."²²  Quran 5:43, 5:47 and 10:94 are the strong ones.  Use those.

The standard Muslim harmonisation

The two sets of verses above look like a contradiction.  Muslim scholarship has an answer, and it is not a dodge.  It runs in three moves:

1. The affirming verses refer to the original Torah and Injil as God revealed them to Moses and to Jesus — real books, genuinely "guidance and light."
2. The accusing verses describe what human beings then did to them.
3. Therefore the Bible now in circulation is not the Tawrat or the Injil the Quran praises. It is a human production that contains scattered remains of them.

On that reading Quran 5:47 is not endorsing the New Testament.  It is endorsing the Injil, a book that no longer exists in usable form.

Saeed sets out the three live Muslim positions, in order of increasing severity:

"(a) the scriptures of Jews and Christians of which the Qur'ān approves as uncorrupted are only those that were actually revealed to Moses (Tawrāt) and Jesus (Injīl), not those that existed with the Jews and Christians at the time of the Prophet Muhammad or exist today;  (b) significant parts of the scriptures that exist today are distorted and corrupted and it is difficult to know which these are;  (c) there are no uncorrupted scriptures of Jews and Christians remaining today — those that the Qur'ān refers to as Tawrāt or Injīl have been obliterated.  This last appears to be the most popular and widely-held view."¹¹

Note who is saying that.  Saeed is a Muslim professor writing in The Muslim World, and he goes on to argue against position (c).  His conclusion, at the end of the same article:  "The wholesale dismissive attitude held by many Muslims in the modern period towards the scriptures of Judaism and Christianity do not seem have the support of either the Qur'an or the major figures of tafsir."¹¹  The disagreement traced in this article is not Christians against Muslims.  It runs through Islamic scholarship itself.

The harmonisation has one obvious cost, and it is the hinge of everything that follows.  It requires the corruption to have happened before Muhammad.  If the Torah and Gospel were already lost or overlaid by the 620s, then the Quran told Christians to judge by a book that had been ruined, called it "guidance and light," and sent its own Prophet to consult its readers when in doubt.  Ibn Taymiyya saw the problem clearly enough to build a whole theory to escape it, and that theory is examined below.

How the doctrine developed

The charge that the biblical text was rewritten is not where a Christian apologist usually expects to find it.  It is thin in the earliest Islamic sources and grows steadily denser over four centuries.  This section traces that, mostly through Muslim sources and through the two scholars who have surveyed them most carefully.

Hadith and sira: the theme is nearly absent

Martin Whittingham's A History of Muslim Views of the Bible: The First Four Centuries (De Gruyter, 2021) searched the whole hadith corpus.  His finding, quoted in Hussain's review:

"Biblical corruption of either text or interpretation barely features in hadith and sīra literature.  Despite the great prominence of this theme in more recent centuries, tradition literature actually yields more regarding use of the Bible."²²

Across the entire corpus, Sunni and Shi'a alike, Whittingham identifies one Shi'a tradition and three Sunni ones that accuse Jews — and in one isolated report Christians — of corrupting their scriptures.²²  Four reports, out of a body of literature that fills shelves.

That is a striking absence.  If the first Muslim generations had believed the Bible was a forgery, the hadith collections are where it would show.

Al-Bukhari's own chapter heading

The single most useful text in this whole debate sits in Sahih al-Bukhari (compiled by al-Bukhari, d. 256/870), and any reader can check it in a minute.  In the Book of Divine Unity — Book 97 — chapter 55 carries a heading built out of Quranic glosses.  Inside that heading, attributed to Ibn Abbas, the Prophet's cousin and the tradition's most-quoted early exegete, is this:

{يُحَرِّفُونَ} يُزِيلُونَ، وَلَيْسَ أَحَدٌ يُزِيلُ لَفْظَ كِتَابٍ مِنْ كُتُبِ اللَّهِ عَزَّ وَجَلَّ، وَلَكِنَّهُمْ يُحَرِّفُونَهُ يَتَأَوَّلُونَهُ عَلَى غَيْرِ تَأْوِيلِهِ

"'They distort' — [it means] they displace;  and no one displaces the wording of a book from among the books of God, mighty and majestic — rather they distort it, interpreting it otherwise than its [true] interpretation."⁶

The gloss is doing exactly the work this article's argument needs, and it is doing it inside the most authoritative Sunni collection there is.  It says the verb in Quran 4:46 does not mean what modern popular argument says it means, and it says so on the authority of Ibn Abbas.

Two practical notes for anyone checking it.  The heading appears only in the Arabic column at sunnah.com/bukhari/97;  the English column does not render chapter headings.  And it is absent from the single-hadith page for Sahih al-Bukhari 7553, which is the hadith the heading precedes.  Go to the whole-book page.

And one concession about its status, which should be made before a Muslim makes it.  This is not a hadith.  It is a taʿlīq — a report al-Bukhari quotes in a chapter title with the chain of transmission suspended, introduced simply by "and Ibn Abbas said."  Hadith scholars treat the taʿlīqāt in the chapter headings as the weakest material in the book, precisely because the chain is not given.  So the gloss cannot be waved about as "Sahih al-Bukhari says the Bible is not textually corrupt."  What it does show is what al-Bukhari, in the ninth century, thought it was reasonable to put in the heading of a chapter in his Kitab al-Tawhid — and that is still worth something, because he was not writing for Christians.

Now the counterweight, which belongs in the same breath.  Two books earlier in the same collection, Ibn Abbas says something that sounds like the opposite.  Sahih al-Bukhari 7363, in the chapter "Do not ask the people of the Scripture about anything":

"Why do you ask the people of the scripture about anything while your Book (Qur'an) which has been revealed to Allah's Messenger is newer and the latest?  You read it pure, undistorted and unchanged, and Allah has told you that the people of the scripture (Jews and Christians) changed their scripture and distorted it, and wrote the scripture with their own hands and said, 'It is from Allah,' to sell it for a little gain."⁸

Sahih al-Bukhari 2685 is the same report in different wording:  "the people of the scriptures have changed with their own hands what was revealed to them."²³  Neither hadith carries a grade line, because inclusion in Bukhari is itself the claim being made about them.

So al-Bukhari preserves Ibn Abbas on both sides of the question, in the same book.  A Christian who quotes only the chapter heading is doing what he complains of when a Muslim quotes only Quran 2:79.  The honest description of the earliest evidence is that it is divided.

Two further Bukhari texts show what Muslims were actually told to do with the Bible in front of them.  Sahih al-Bukhari 4485 has Muhammad respond to Jews reciting the Torah in Hebrew and explaining it in Arabic:  "Do not believe the people of the Scripture or disbelieve them, but say:— 'We believe in Allah and what is revealed to us.'"²⁴  That is a posture of suspended judgement, not a verdict of forgery.  It is hard to reconcile with a settled early belief that the book was a fake.

The early exegetes: mostly corruption of meaning

The pattern holds when the argument moves from hadith to commentary.  Adang read nine medieval Muslim writers on Judaism and the Hebrew Bible and reported where each stood.

Ibn Rabban al-Tabari (d. after 250/864) was a Christian who converted to Islam and wrote a defence of his new religion out of the Bible.  In his own preface he says he translated the Torah, the Gospel and the prophets into Arabic "letter for letter," and that he "did not seek to beautify or embellish its wording for fear of distortion (al-tahrif)."⁷  A man who thought the book was a forgery would not worry about distorting it in translation.

Ibn Qutayba (d. 276/889) used tahrif twice, both times about interpretation.  Adang:  "In both these cases, however, it is clear that what is meant by tahrif is giving a wrong interpretation to an otherwise genuine text.  Ibn Qutayba does not question the authenticity or validity of the Jewish scriptures, and nowhere does he accuse the Jews of having distorted them."⁷

Al-Tabari (d. 310/923), the greatest of the early commentators, defined the term explicitly.  Adang:  "Al-Tabari explicitly states what he understands by distorting the word of God:  changing its meaning and interpretation, deliberately bending its original meaning to something else."⁷  Hussain, reading him independently, agrees that al-Tabari "seems to overwhelmingly support the notion that tahrif refers to deliberate misinterpretation of the text."²²  This is not unanimous:  Whittingham finds al-Tabari's position hard to pin down, and Gordon Nickel's study of tampering narratives in the earliest commentaries agrees that he does not stake out a clear line.²²

Al-Baqillani (d. 403/1013) used the word for something milder still.  Adang:  "The term tahrif as used by him stands for inadvertent errors made in the process of translation, rather than deliberate alterations effected in the text of the Torah."⁷

Whittingham's summary of the earliest commentators, quoted by Hussain:  "the earliest commentaries seem to assume a broadly authentic text."²²

Ibn Hazm is the turn

The change comes in eleventh-century Spain, with Ibn Hazm of Cordoba (d. 456/1064).  Adang's verdict on her nine authors:

"It was found that the majority of our authors subscribe to a mild interpretation of the Koranic allegation of large-scale tampering with the Torah by the Jews (taḥrīf);  according to this interpretation, only the sense of the biblical text had been changed while the text itself remained intact.  Only al-Maqdisī and Ibn Ḥazm believed that the text itself had suffered distortion.  The person held responsible by Ibn Ḥazm for the corruption of the Torah was Ezra the scribe, who was generally put in a very positive light by Ibn Ḥazm's predecessors."⁷

Ibn Hazm did the work.  Adang:  "It is only Ibn Hazm, however, who made a thorough study of the Bible with express polemical purposes."⁷  He read it closely, catalogued its difficulties, and concluded that Ezra had rewritten the Torah:  in Ibn Hazm's eyes "Ezra's emendations were no mere corrections, but rather radical changes which altered the entire character of the Torah, turning it into a profane piece of writing, a mere forgery."⁷  On the Gospel he was equally flat:  "The gospel sent down from God disappeared, except for small sections which God preserved as a proof against them, and a shame for them."¹⁹

And he won.  Whittingham, on the chapter he devotes to Ibn Hazm and his contemporaries:

"the tide turned decisively against the acceptance of the Bible.  Decisively does not mean exclusively, of course, but the views expressed in this chapter flowed into the mainstream of Muslim Biblical criticism.  Where before they had surfaced either in passing, or as one among a series of points in interfaith exchanges, explicit criticism of textual reliability gradually assumes a more dominant position."²²

His summary of the whole four centuries:

"To draw a broad but foundational conclusion, in comparison with the modern assumptions about Muslim views, there is less impulse to condemn large parts of the Bible, more impetus to use it, and the accusation of corruption of the text lingers in the background, present but not dominant, until the 5th/11th century."²²

Four hundred years after Muhammad.  That is when the doctrine most Muslims now hold as obvious became dominant.  This does not make it false — doctrines can be developed correctly — but it does mean it cannot be presented as the plain teaching of the Quran that every early Muslim understood.

Ibn Taymiyya's compromise

Ibn Taymiyya, writing in Damascus three centuries after Ibn Hazm, refused both extremes.  His position is the one a Christian apologist should actually engage, and almost never does:

"It is said that in the world there is no single copy [or version of the scripture] that corresponds to what God revealed in Tawrāt and Injīl.  All that exist are changed (mubaddal). … It is [also] said:  No one has changed any text of the scriptures.  Rather they have falsified their meanings by [false] interpretations.  Many Muslims have held both of these views.  The correct [view] is the third view, which is that in the world there are true (ṣaḥīḥ) copies [versions], and these remained until the time of the Prophet, and many copies [versions] which are corrupted.  Whoever says that nothing in [these] copies was corrupted he has denied what cannot be denied.  Whoever says that, after the Prophet, all copies have been corrupted, he has said what is manifestly false.  The Qur'ān commands them to judge with what Allah revealed in Tawrāt and Injīl. … There is nothing in the Qur'ān to indicate that they altered all copies."¹¹

Read that carefully.  Ibn Taymiyya says that sound copies of the Torah and Gospel existed in Muhammad's day, that the Quran's command to judge by them proves it, and that the claim of universal corruption after Muhammad is "manifestly false."  The version of tahrif most commonly heard today — that the whole Bible is a corrupted book — is one that Ibn Taymiyya, of all people, explicitly rejected.

The manuscripts that predate Muhammad

Muhammad's public preaching began around AD 610.  The standardised Quran is traditionally dated to Uthman's reign, around AD 650.  Everything in the table below is older than that, most of it by centuries, and every item can be inspected today.

That table is the empirical core of this article.  Its argument is worth stating at exactly its real strength, no more.

What it proves

The text Christians read today is substantially the text that existed centuries before Muhammad was born.  Whatever the Bible says now, it said then.  This is not a matter of trusting Christians;  the objects are in Jerusalem, Dublin, London, Geneva and the Vatican, and the Isaiah scroll is on public display in the Shrine of the Book.

The clearest single demonstration is the Isaiah material, because the comparison is between a scroll copied before Christ and the medieval Hebrew codex from which modern Bibles are translated — a gap of roughly a thousand years.  Emanuel Tov, editor-in-chief of the international Dead Sea Scrolls publication project, made the comparison for 1QIsa-b, which preserves about forty-six chapters:

"When comparing 1QIsa-b, dating from the first century BCE, with codex L written one thousand years later, one easily recognizes the close relation between the two texts that are sometimes almost identical."⁹

He then lists every category of difference across those forty-six chapters, describing them as "features in the scroll that represent differences from codex L, all of which concern minutiae":  107 differences of spelling, 16 added conjunctions, 13 missing conjunctions, 4 differences in the definite article, 10 differences in letters, 5 missing letters, 14 in number, 6 in pronouns, 24 in grammatical form, 9 in prepositions, 11 different words, 5 words absent, 6 words added, 4 differences in word order.⁹  Tov quotes the 1959 verdict of Bleddyn Roberts:  "The almost complete absence of textual variants is the clearest indication of the close affinity between the two text-forms, for in one instance only can a case be made for a significant variant reading.  It is in 53.11."⁹

One arguable significant variant, in forty-six chapters, across a thousand years of copying.  And Tov's verdict on the whole Isaiah tradition, Hebrew and Greek together:  "the known texts of Isaiah do not differ from each other recensionally."⁹

Isaiah is the right test case, because Isaiah 53 is the chapter Christians press hardest on Jewish and Muslim readers.  If anyone had rewritten it to serve Christian claims, this is where it would show — and the one significant variant in the whole book sits at Isaiah 53:11, where the pre-Christian scrolls and the Greek translation carry an extra word, or, "light," which the medieval Hebrew lacks.⁹  The pre-Christian witnesses have the longer reading.  Whatever happened there, it is not a Christian addition.

The New Testament side is younger but still decisive.  Codex Sinaiticus is "a manuscript of the Christian Bible written in the middle of the fourth century," containing "the earliest complete copy of the Christian New Testament."¹⁰  That is roughly 250 years before Muhammad began preaching.  Its Gospels contain the crucifixion, the death, the burial, and the resurrection appearances.

What it does not prove

It says nothing about whether the text was already corrupted before the surviving manuscripts.  Saeed's position (a) — that the true Tawrat and Injil were lost long ago — is untouched by a fourth-century codex or a second-century-BC scroll, because that position places the loss earlier still.  Anyone who thinks the manuscript table ends the argument has not understood the argument.

It also cannot show that the manuscripts are complete.  A book that was suppressed early enough and thoroughly enough would leave no trace.  Absence of evidence is weak evidence here, not proof.

What it does do, and this is the useful part

It forces the Muslim claim to name a date.  There are only three options, and each costs something.

If the corruption happened after Muhammad, the manuscripts refute it flatly, with no appeal to faith required.  This is also the option Ibn Taymiyya rejected in the strongest terms available to him:  "Whoever says that, after the Prophet, all copies have been corrupted, he has said what is manifestly false."¹¹

If the corruption happened between Jesus and Muhammad, then the Quran of the 620s is affirming a book that was already ruined.  It calls the Gospel "guidance and light" (Quran 5:46), commands Christians to judge by it (Quran 5:47), tells the Jews of Medina that they have "the judgement of Allah" in the Torah (Quran 5:43), and sends Muhammad to its readers when he doubts (Quran 10:94).  Saeed, a Muslim, states the problem for his own tradition without softening it:

"We know from the history of these two religious traditions that by the time the Prophet was preaching in the early seventh century CE, the scriptures of both Jews and Christians were documented.  The Tawrāt they had at the time is what they have now.  The same is true for the Gospels.  Since the Qur'ān refers to those same scriptures, its references to them should equally apply in the modern era.  This is perhaps the main challenge to Ibn Taymiyya's position."¹¹

If the corruption happened before Jesus — the position Ibn Hazm took about Ezra — then the Torah that Jesus quoted was already a forgery, and Quran 5:46's claim that the Gospel confirms the Torah collapses, since it would have Jesus confirming a corrupt book.

None of these three is comfortable.  The second is where most modern argument lands, and it is the one the Quran's own affirming verses make hardest to hold.

"But the Bible has variants too"

This is the strongest empirical move on the Muslim side, and it deserves the real numbers rather than the comfortable ones.  The figures below are carried over unchanged from the Berean audit at New Testament manuscripts and variants, which cut down several inflated apologetic claims in the process.

• About 5,300 Greek New Testament manuscripts survive. Not 5,800, and not 25,000. Jacob Peterson, a textual critic writing in a book produced by evangelical scholars to correct their own side: "It is best to say that there are about fifty-three hundred Greek New Testament manuscripts in existence, although fifty-one hundred might be the safer estimate."¹³
• About 500,000 non-spelling variants exist across the whole Greek tradition. That is higher than Bart Ehrman's widely quoted "400,000 or more." It is Peter Gurry's estimate, built from actual collation data rather than guesswork.¹² Gurry's own gloss on his number: it "reflects the frequency with which scribes copied more than their infidelity in doing so."¹²
• At least 75% of those variants are spelling slips and nonsense readings. Fewer than 1% are both meaningful and viable — that is, they make a difference to the sense and have a real chance of being original.¹⁴
• About forty verses are genuinely doubtful, and every modern Bible flags them in a footnote. The two large blocks are Mark 16:9–20 and John 7:53–8:11. Both were already known to be doubtful in 1516, when Erasmus printed the first published Greek New Testament.³⁰
• Ehrman himself, the best-known skeptic of the New Testament text, wrote in his own book's appendix that his position "does not actually stand at odds with Prof. Metzger's position that the essential Christian beliefs are not affected by textual variants in the manuscript tradition of the New Testament."³¹

Half a million variants is a real number and Christians should not minimise it.  But it is a measurement, and the reason it can be measured is that 5,300 manuscripts survive to be compared.  A text preserved in three copies would have almost no countable variants.  That would be a fact about its poverty, not its purity.

The point that actually answers the objection

The Muslim objection is not really about variant counts.  It is about a specific claim:  that the Gospels were altered to say Jesus was crucified and rose, when he was not and did not.  So test that claim directly.

No manuscript of any of the four canonical Gospels, in any language, at any date, omits the crucifixion or reports a substitute.  There is no variant to argue about at the point that matters.  This is not an argument from the reliability of the text in general;  it is an observation about one narrative, and it can be checked against the apparatus of any critical edition — the Nestle-Aland Novum Testamentum Graece prints the variants for Matthew 27, Mark 15, Luke 23 and John 19 on the page, and none of them is of that kind.³²

Be precise about this, because the loose version is false.  Two of the roughly forty doubtful verses are inside the Passion narrative:  Luke 22:43–44, the angel in Gethsemane and the sweat like blood, and Luke 23:34a, "Father, forgive them, for they know not what they do."  Both are missing from P75 and Codex Vaticanus, and modern editions print them in double brackets.³⁰  Neither touches the question at issue.  A copyist adding or dropping a sentence about an angel is not a copyist inventing an execution.  What no manuscript anywhere does is leave Jesus off the cross.

The only ancient texts that deny the crucifixion of Jesus are second- and third-century Gnostic works.  The best known is the system of Basilides, reported by Irenaeus of Lyon around AD 180:

"Wherefore he did not himself suffer death, but Simon, a certain man of Cyrene, being compelled, bore the cross in his stead;  so that this latter being transfigured by him, that he might be thought to be Jesus, was crucified, through ignorance and error, while Jesus himself received the form of Simon, and, standing by, laughed at them."³³

That is remarkably close to the substitution stories in later Islamic commentary.  But it is no help to a Muslim, for a reason a Muslim scholar states better than a Christian could.  The Gnostic denial rests on Jesus having had no real body — Irenaeus reports Basilides saying Jesus "was an incorporeal power."  The Quran flatly denies that:  Jesus and his mother "both used to eat food" (Quran 5:75).³⁴  Mahmoud Ayoub, a Muslim scholar of Christian-Muslim relations, put it plainly:  "Islam, however, does not admit of docetism in any form.  Its very human images of the life to come, as well as its insistence on the humanity of God's messengers, argues against such a view."³⁵

No church ever accepted any of these Gnostic works.  They are not suppressed Gospels;  they are the writings of movements the churches argued against in public, which is how they are known at all.

A note on fairness.  It would be easy here to change the subject to the Quran's own textual history — the Sanaa palimpsest, the variant readings, the report in Sahih al-Bukhari 4987 that Uthman "ordered that all the other Qur'anic materials, whether written in fragmentary manuscripts or whole copies, be burnt."³⁶  That is a real subject with real evidence, and it is treated in has the Quran been perfectly preserved?.  It is not treated here, because answering "your book has variants" with "so does yours" concedes the question instead of answering it.

The Gospel of Barnabas

A book called the Gospel of Barnabas circulates widely in Muslim da'wah — outreach — literature as the suppressed true Gospel.  In it Jesus denies being the Messiah and predicts Muhammad by name.  It is worth handling carefully and kindly, because many Muslims who cite it have simply never been told what it is, and being told rudely will make them stop listening.

What survives.  Two manuscripts.  An Italian codex in the Austrian National Library in Vienna (Cod. 2662), and a Spanish version now lost, known through eighteenth-century descriptions.  There is no Greek, no Syriac, no Coptic, no pre-modern Arabic, no patristic quotation, and no ancient canon list that names this book.  The standard edition is Lonsdale and Laura Ragg's 1907 Oxford text, which prints the Italian with a facing English translation and a long critical introduction.¹⁵

The decisive argument is a single sentence inside the book.  In chapter 82 Jesus tells the disciples that in the time of the Messiah "the jubilee every year — that now cometh every hundred years."¹⁵  Ragg's analysis:

"This cannot be the Hebrew Jubilee, which recurred every fiftieth year;  and the Qoran is innocent of Jubilees.  There remains, then, the Jubilee of Western Christendom:  and there was only one period, so far as we know, when this celebration could have been spoken of as 'recurring every 100 years'.  The first recorded Jubilee [was] that of Boniface VIII in [1300];  the second, [authorised by] Clement VI in 1350.  The first Jubilee had been so brilliant a financial success that it was decided to shorten the interval!  This would give us precisely the first half of the fourteenth century as the period in which the passage in question must have been written."¹⁵

The Hebrew jubilee is every fiftieth year (Leviticus 25:11).  The hundred-year jubilee existed in Latin Christendom for exactly fifty years, from 1300 to 1350, and then was shortened.  A first-century Jewish text cannot contain it.  Nothing else about the book needs to be argued.

Everything else points the same way.  Ragg lists parallels to Dante's Divine Comedy and notes that "Barnabas has nine Heavens, with Paradise — like Dante's Empyrean — for the tenth."  He records that "the scribe of the Vienna MS. was certainly a Venetian of the sixteenth or early seventeenth century," and that three Italian philologists he consulted — Ernesto Monaci of Rome, Giovanni Alfredo Cesareo of Palermo, and Vincenzo Crescini of Padua — put the word-forms and style in the fifteenth or sixteenth century.¹⁵  His conclusion on authorship:

"In short, some one about 1575 — possibly the Fra Marino of the Spanish Preface — either copied out or invented this 'Gospel of Barnabas'."¹⁵

And on the writer's background:

"The Italian Barnabas is, to all intents and purposes, an original … it is the work of one who, whether priest or layman, monk or secular, has a remarkable knowledge of the Latin Bible — as remarkable, perhaps, as Dante's … It is the work of one whose knowledge of the Christian Scriptures is considerably in advance of his familiarity with the Scriptures of Islam:  presumably, therefore, of a renegade from Christianity."¹⁵

On the claim of a lost Arabic original, Ragg is blunt:  "as a matter of fact the evidence for an Arabic original is vague and fugitive.  Experts say that the Italian of our MS. by no means suggests or warrants it."¹⁵  His note quotes Carlo Alfonso Nallino of Palermo, professor of Arabic:  "Non vi trovo affatto tracce di arabismi" — "I find in it no traces of Arabisms at all."¹⁵  Ragg also observed that the book depends on the canonical Gospels for its material:  "no candid reader of Barnabas, furnished with even elementary critical acumen, can fail to recognize that our document is dependent for the bulk of its matter on the four Canonical Gospels of Christendom."¹⁵

And it does not even support Islam.  The Gospel of Barnabas makes Muhammad the Messiah and has Jesus deny that title.  The Quran calls Jesus al-Masih, the Messiah, at Quran 3:45, 4:157, 4:171 and 5:75.³⁷  Granting the book everything it asks for would contradict the Quran on the one point it exists to make.  Later scholarship has moved the composition later still and located it among Spanish Muslim exiles around 1600 — the line of argument developed by Jan Slomp, Mikel de Epalza, Luis Bernabé Pons and Gerard Wiegers — but none of that work was consulted for this article, and Ragg's fourteenth-century Jubilee argument does not need it.³⁸

The missing prophecies

The last fallback is the most common one in conversation.  When a Bible passage is produced and does not say what was claimed, the answer is that the bisharat — the "glad tidings," the prophecies of Muhammad — were removed.  The Quran describes believers as those who follow "the unlettered prophet, whom they find written [i.e., described] in what they have of the Torah and the Gospel" (Quran 7:157),³⁹ so if the description is not there, someone must have taken it out.

The move, as usually made, cannot be tested.  Its own content is that the evidence which would test it no longer exists.  Produce a manuscript and the answer is that this copy is already corrupted.  Point out that no manuscript anywhere shows the variant and the answer is that the corruption was universal and early.  A claim that predicts the same observation no matter what the world is like is not doing any work.  That is not a proof it is false.  It is a reason to stop treating it as evidence.

Two things make it harder to hold here than elsewhere.

First, the earliest and best Muslim writers on this exact subject did not make it.  They argued from the Bible they had.  Ibn Qutayba used tahrif twice in his Dala'il al-nubuwwa, both times about Paran and Habakkuk — passages he read as pointing to Muhammad — and Adang's judgement is that in both cases he meant "giving a wrong interpretation to an otherwise genuine text."⁷  Ibn Rabban al-Tabari built the earliest systematic list of biblical testimonies to Muhammad out of the text as he found it, and worried about distorting it in translation.⁷  Men who believed the prophecies had been cut out would not have spent their careers finding them.

Second, the test is available and has been run.  What would falsify "the prophecy was removed" is a pre-Islamic witness that contains it.  The best-known case is John 14–16, where the Greek word is parakletos, "advocate" or "comforter," and the Muslim claim is that it replaced periklytos, "praised one," which would translate the name Ahmad.  Every pre-Islamic manuscript that preserves those verses reads parakletos:  P66 and P75, from the late second to fourth century, and Codex Sinaiticus and Codex Vaticanus from the fourth.⁴⁰  A full audit is in Muhammad in the Bible, which also states, at length, where the Christian side of that argument overstates.  The clean positive test the Muslim side can run at any time is to produce one pre-Islamic witness — Greek, Latin, Syriac, Coptic, Armenian, Georgian, Ethiopic, or a quotation in any church father — reading periklytos at John 14:16, 15:26 or 16:7.  None has ever been produced.

The strongest Muslim responses

These are the serious positions.  Each is stated in its own advocates' words and then answered.

"The true Injil was lost, and the four Gospels are not it"

This is the real Muslim position, and it is coherent.  The Quran describes the Injil as a book God gave to Jesus, in the singular, the way the Quran was given to Muhammad (Quran 5:46).  The four Gospels are not that.  They are four accounts about Jesus, by human authors, written decades later.  Whittingham lays out the mismatch fairly:

"So between Muslims and Christians, there are two different traditional concepts of Gospel or Injīl in operation.  One is of a message sent down to Jesus, all at once.  The other is of records which Christians view as divinely inspired but also humanly authored, of a message about Jesus.  Since the canonical New Testament gospels do not seem to fit the Muslim expectation of what the Injīl should be like, Muslim scholars have proposed the existence of a pure, original Injīl which is partly or sometimes barely preserved in the New Testament gospels."¹⁹

This cannot be refuted by manuscripts, and Christians should stop pretending otherwise.  No codex bears on the existence of a book that, by hypothesis, was lost before any codex was made.

What can be asked is what evidence there is that the book existed.  And the sharpest form of that question comes from a Muslim.  Muhammad Abu Laylah, in The Qur'an and the Gospels, raises four objections to the theory — objections that arise, as Whittingham stresses, "not from apparent tensions with historical or manuscript evidence, but from factors arising from Islamic principles":

"First, if a sound version of the Injīl endured until the time of Muhammad, and presumably beyond, why did no early Muslim mention it in books?  Secondly, why was it not preserved by Muslims?  Thirdly, Muhammad would surely have safeguarded a proposed original Injīl.  Fourthly, Muhammad would not have allowed the four New Testament gospels to eclipse the pure Injīl."¹⁹

Add the external form of the same point.  The proposed Injil has no manuscript, no fragment, no quotation in any church father, no lectionary, no translation into any language, no mention in any canon list, and no enemy who complains about it.  Every other early Christian book that was rejected — the Gospel of Thomas, the Gospel of Peter, the Gospel of the Hebrews — is known to us precisely because somebody argued against it in writing.  A book that vanished so completely that even its opponents forgot to mention it is an unusual kind of object.

And Whittingham's own closing question shows the practical difficulty:  "it is difficult to know how contemporary Christians could themselves have identified what the true Injīl stated.  Was it a command to follow Muhammad, precepts taught by Jesus, or, more broadly, the New Testament if rightly interpreted?"¹⁹  Quran 5:47 commands Christians to judge by the Gospel.  If nobody, then or now, can say what that book contained, the command cannot be obeyed.

"The Gospels are reports about Jesus, not revelation to Jesus"

Ibn Taymiyya's move, and the one that best survives the manuscript evidence:

"What is in Injīl of stories about the crucifixion of Jesus, his death, his coming to the disciples after he was 'raised,' are not what Jesus said.  Rather they are [reports] from those who saw those things after him.  What God revealed was what was heard [directly] from Jesus who is the transmitter [of God's word] from Him."¹¹

On this view the Gospels can be transmitted perfectly and still not be scripture.  They are testimony.  A Muslim is not required to accept human testimony over a later revelation.

The answer is not to argue about manuscripts, which do not touch this.  It is to notice two things.

First, the move proves more than its user wants.  Almost the whole Quran is, on Islam's own account, a report:  the Quran was heard by Muhammad, spoken to companions, memorised, written down, collected under Abu Bakr and standardised under Uthman, and transmitted by named chains of men whose reliability the hadith scholars spent centuries grading.  If the category "report about" disqualifies the Gospels from being revelation, the same category must be handled carefully in Islam's own case.  What distinguishes them is not the presence or absence of human transmission.  It is a claim about what God originally gave, which is the thing in dispute.

Second, this is where a Muslim and a Christian genuinely part company, and it is honest to say so.  Christianity does not claim the Gospels are dictation.  It claims that God acted in history and that witnesses recorded it.  The Muslim expectation of what revelation must look like — a single book, sent down, dictated — is being applied to a religion that never made that claim.  That is a real disagreement about the nature of revelation, and neither side wins it by producing a manuscript.

"You cannot quote Quran 5:47 at us"

Correct, and Christians should know why.  Whittingham's survey lists five different Muslim readings of that verse, and none of them concedes what the apologist wants:

1. It addressed people who lived before Muhammad; after him they should follow the Quran. Ibn Kathir's option.
2. Sound copies as well as corrupt ones were circulating in the seventh century. Ibn Taymiyya's position in his Tafsir.
3. "What God sent down in it" means the command to follow Muhammad, and nothing else. Ibn Hazm's move, and Ibn Taymiyya's in al-Jawab al-sahih: "God handed down in the Gospel the command to follow Muhammad, just as He commanded it in the Torah."¹⁹
4. The Injil is Jesus' teaching, so the parts of the Gospels that record his own words are the Injil. Maududi's suggestion.
5. The Injil is whatever all four Gospels agree on — rejected by Abu Laylah himself, because that would include the crucifixion.

Reading 1 has a problem inside Islam.  Ibn Taymiyya rejected it, on the ground that "God would not issue a command to those who had died before that command was issued."¹⁹  Reading 3 empties the verse:  it makes God tell Christians to judge by their Gospel, when what he means is a single command they cannot find in it.  Readings 4 and 5 return the argument to the text of the Gospels, which is where the manuscripts apply again.

"Muslims can use the Bible without believing it"

Shabir Ally, the best-known contemporary Muslim debater in English, sets out the working position in an article posted in January 2014:

"The easiest way to prove a point to a person who believes in the Bible is to make reference to the Bible.  This does not mean you believe in the Bible or consider it true in even that portion. … You will have to remind him that the Bible is a proof against him, not against you.  You were just asking him why his statement about Jesus is opposite to his Bible's statement about Jesus.  He cannot now ask you why your statement about God is different from the Bible because you never claimed to believe in the entire bible."⁴¹

Ally distinguishes carefully:  the hadith of Umar, in which Muhammad is displeased to find Umar reading a Torah scroll, "teaches that we cannot use the Bible for guidance.  That hadith does not prevent us from using the Bible for Dawah."⁴¹

This is logically clean, and no Christian should call it dishonest.  Arguing from a premise your opponent accepts is standard practice, and Christians do it constantly.

But it has a cost that is worth naming gently.  Ally's own hundred-page book Is Jesus God?  The Bible Says No! argues from the Bible page after page — and contains no discussion of tahrif at all.⁴²  The Bible is treated as sufficiently reliable to settle the deity of Christ.  If the same book is reliable enough to establish what Jesus said about the Father, the argument that it is too corrupt to establish what happened at Golgotha needs a reason that applies to one passage and not the other.  Ibn Taymiyya's distinction between revelation to Jesus and reports about him supplies exactly such a reason — which is why that is the position to engage, and this one is not.

Where the Christian argument is weak

Four concessions.  All of them are conceded by non-Muslim scholarship, not extracted under pressure.

1.  The Quran does not say the biblical text is corrupt — and it does not say it is preserved either.  Reynolds is the authority for the first half:  "neither here nor elsewhere … does the Qur'ān argue that the Bible of the Jews and Christians is falsified."⁵  But the same restraint cuts the other way.  The Quran nowhere guarantees the preservation of the Torah or the Gospel the way Quran 15:9 guarantees its own message.⁴³  A Christian who says "the Quran teaches that the Bible is preserved" is overreaching just as much as a Muslim who says it teaches the opposite.  What the affirming verses establish is narrower:  that the Quran treats those books as available and authoritative in the 620s.  That is enough for the argument in this article, and it is all that should be claimed.

2.  Quoting Quran 5:47 at a Muslim is not a checkmate.  Five distinct answers are on record, listed above.  Presenting the verse as though it settled the matter will convince nobody who has read Ibn Taymiyya, and will make the person quoting it look as though he has not.

3.  The lost-Injil position is coherent and cannot be refuted by evidence.  It can only be pressed for evidence, which it does not have.  That is a weaker result than a refutation, and this article does not claim more.

4.  The Islamic tradition is not uniformly on the Christian side, and mining it is dishonest.  Bukhari preserves Ibn Abbas in both directions.  Adang's majority held the mild view — but al-Maqdisi and Ibn Hazm did not, and Ibn Hazm's position became the mainstream.⁷  A Christian who cites Bukhari's chapter heading and hides Sahih al-Bukhari 7363 is doing to Islamic sources what he complains of when a Muslim cites Quran 2:79 and hides Quran 5:47.

5.  Isaiah is the best case for the Hebrew Bible, not the average one.  The 1QIsa-b comparison above is genuinely impressive, and it is also the showcase.  Two things pull the other way and belong here.  First, the biblical text at Qumran was plural, not uniform:  of the forty-six Qumran Torah manuscripts extensive enough to classify, Tov's revised statistics put only about half in the Masoretic-like group, with a comparable share "non-aligned."  Second, the three text-families of the Hebrew Bible — Masoretic, Samaritan and the Hebrew behind the Septuagint — disagree about the ages in Genesis 5 and 11 by more than a thousand years in total, and that disagreement is older than Christianity.  Both facts are documented at length in the Berean audit of the Hebrew text, which does not soften them.²⁵  Neither is tahrif in the Muslim sense — nobody rewrote a doctrine — but a Christian who cites Isaiah and stops there is showing the reader his best card and not the hand.

One more, about method.  This article rests heavily on al-Bukhari for the early Islamic evidence.  That is deliberate:  it is a collection Sunni Muslims accept as sound, so the argument does not require anyone to adopt a Western dating.  But the rule runs both ways.  A report from Bukhari that helps the Christian case is no better attested than one that hurts it, and this article has printed both.

Common objections and responses

"The Quran plainly says the Bible was corrupted"

It says a party of Jews shifted words out of their places, twisted their tongues, concealed, and forgot.  The noun tahrif never occurs.  Al-Bukhari's own chapter heading in Book 97 glosses the verb as misinterpretation and adds that "no one displaces the wording of a book from among the books of God."⁶  A Muslim is entitled to reject that gloss.  He is not entitled to say the Quran plainly teaches textual corruption when the tradition's most authoritative collection preserves the opposite reading.

"The Bible has 400,000 variants"

The real figure is higher — about 500,000 — and it exists because about 5,300 Greek manuscripts survive to be compared.¹²¹³  At least three quarters are spelling slips;  fewer than 1% are both meaningful and viable.¹⁴  About forty verses are genuinely doubtful, all flagged in modern Bibles, and none of them is in the Passion narrative.  Ehrman, the skeptic whose book made the number famous, wrote that his position does not conflict with Metzger's view that "the essential Christian beliefs are not affected by textual variants."³¹

"Christians changed the Bible to hide Muhammad"

Then the change must predate the oldest manuscripts, which for John is around AD 175–225 and for the Hebrew prophets is the second century BC.  A change that early would have to be universal, simultaneous, and invisible in every language into which the text had already been translated — Greek, Latin, Syriac, Coptic — and would have to have been carried out by communities that were persecuting each other and would have used any such tampering as ammunition.  No trace of that argument survives in any ancient source, Christian, Jewish or pagan.

"Modern textual critics admit the New Testament was altered"

They do, and Christians should say so first.  Scribes harmonised the Gospels to each other, smoothed rough Greek, and in a handful of cases changed a reading in a doctrinally convenient direction.  What the same scholars also say is that the alterations are visible precisely because so many manuscripts survive, that they cluster in additions rather than deletions, and that no Christian teaching rests on a disputed reading.

"The Gospel of Barnabas is the suppressed true Gospel"

It refers to a Jubilee coming "every hundred years," a custom that existed only from 1300 to 1350 in Latin Christendom.¹⁵  Its oldest manuscript was written by a Venetian scribe of the sixteenth or early seventeenth century.¹⁵  It makes Muhammad the Messiah, contradicting the Quran, which calls Jesus al-Masih.³⁷

"You are just as biased as we are"

True, and irrelevant to the manuscripts.  Nothing in this article asks anyone to trust a Christian.  The Isaiah scroll was published by Israeli and American scholars, mostly Jewish and secular;  the papyrus datings used here are the later, less convenient ones set by Pasquale Orsini, Willy Clarysse and Brent Nongbri;  the variant figures come from a book written by evangelical textual critics to correct evangelical apologists;  and the key statements about early Muslim views come from Camilla Adang, Martin Whittingham, Gabriel Said Reynolds and the Muslim scholars Abdullah Saeed and Saqib Hussain.

What evidence would change the conclusion?

What would strengthen the Muslim claim:

• A manuscript, fragment, translation, lectionary, or patristic quotation of a Gospel materially different from the four — especially one that omits the crucifixion for reasons other than Gnostic denial of Jesus' body.
• Any pre-Islamic witness in any language reading periklytos at John 14:16, 15:26 or 16:7.
• An early Christian or Jewish source complaining that scriptures had been rewritten, from the period the rewriting is supposed to have happened. Ancient religious communities argued in writing constantly; a change of this magnitude leaving no complaint is what needs explaining.
• Evidence from before the eleventh century that Muslim scholars generally understood the Quran to teach textual, not interpretive, corruption. Adang's and Whittingham's surveys are the current state of that question and could be overturned by fuller work in the manuscript sources.
• Convincing evidence of a distinct Injil — a citation, a title in a canon list, a quotation in a Syriac or Arabic Christian writer, anything with a date.

What would weaken it further:

• Publication of more pre-Islamic biblical manuscripts agreeing with the received text, particularly in Syriac and Arabic, from the fifth and sixth centuries.
• A closer reading of the earliest tafsir confirming that al-Tabari and his contemporaries meant interpretive corruption. This is currently disputed, with Hussain reading al-Tabari one way and Whittingham and Gordon Nickel finding him harder to pin down.²²

What would not change anything either way:  variant counts.  Both sides use them as scoreboards, and they measure how much evidence survives, not how faithfully it was copied.

Sources