New Testament Manuscripts & Variants

About 5,300 Greek manuscripts preserve the New Testament better than any ancient text. Roughly 40 verses remain doubtful, no doctrine depends on them, and the unobserved first century of copying remains the main gap.

Summary

Skeptics hear that the New Testament was copied by hand for centuries and conclude nobody can know what it originally said.  Apologists answer with big numbers:  5,800 manuscripts, 25,000 counting translations, a fragment of Mark from the first century.  Both sides are wrong in ways that can be checked.  This article audits the numbers from the primary sources and reports what survives.

  • The New Testament is the best-documented text from the ancient world, by a wide margin. About 5,300 Greek manuscripts survive, plus a few thousand more in Latin, Syriac, Coptic and other languages.11 The best-attested classical works survive in a few hundred copies; most survive in a handful. Bart Ehrman, the best-known skeptic of the New Testament text, agreed on the record: asked what to call the gap between the originals and first copies of Suetonius, Tacitus, or Pliny after calling the New Testament's 150-year Galatians gap "enormous," he answered "Very enormous… Ginormous doesn't cover it. For the New Testament we have much earlier attestation than for any other book."2
  • But the favorite apologetic numbers are inflated, and Christians should stop using them. The catalog holds 5,700 entries, not 5,900; after removing destroyed, missing, and double-counted items, the honest count is about 5,100–5,300.31 "25,000 total manuscripts" is, in the words of a textual scholar writing for an apologetics-friendly publisher, "an exaggeration too big to keep using."1 The Iliad comparison figure is 1,757 manuscripts, not 643.4 There are no first-century New Testament papyri, and the famous "first-century Mark" turned out to be second-to-third century — and stolen.56
  • The manuscripts are mostly late and mostly partial. 83% date from after AD 1000. Only about 60 contain the complete New Testament. Roughly 54 manuscripts, most of them fragments, predate AD 300.17 The earliest scrap, P52, holds five verses of John and is dated anywhere from AD 125 to 200.8
  • About 500,000 textual variants exist — because about 5,300 manuscripts exist. The best published estimate, built from actual collation data, is roughly half a million non-spelling variants across the whole Greek tradition.9 The estimate's author notes the number "reflects the frequency with which scribes copied more than their infidelity in doing so."9 At least 75% of variants are spelling slips and nonsense readings. Less than 1% are both meaningful and viable.10 Westcott and Hort put the "substantial variation" at about a thousandth of the text in 1881, and nothing found since has moved that needle much.11
  • The genuinely doubtful text amounts to about 40 verses, and every modern Bible flags them. The two big blocks are Mark 16:9–20 and John 7:53–8:11. Erasmus, working in 1516 with two twelfth-century manuscripts, already knew about the doubt attached to both.12 Five centuries and five thousand manuscripts later, the list of doubtful passages has barely grown.
  • No Christian doctrine stands or falls on a disputed reading. That is the stated position of Bruce Metzger, D.A. Carson, and Daniel Wallace — and of Bart Ehrman, who 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."131415
  • The honest weakness: the first hundred years of copying cannot be directly checked. No manuscript survives from the first century. Ehrman's serious argument is not the variant count; it is that the earliest copying period "was also the least controlled," and Helmut Koester argued the Gospel text "was very unstable during the first and second centuries."1516 The best reply is evidence, not assertion: where the early period can be tested — the papyri against the great fourth-century codices — the text proves stable, and a scholar in the same volume as Koester showed the claim of extensive early ideological rewriting "runs counter to all of the textual evidence."1718 But a gap is a gap, and it should be stated, not papered over.

Verdict:  the core argument survives the audit once its numbers are cut down to honest size.  Copying was not a game of telephone;  it was a branching tree whose earliest surviving branches can be compared against each other and against branches a thousand years younger, and they agree to a degree no other ancient text approaches.  What the manuscripts prove is that the text was transmitted faithfully.  Whether what the text says is true is a separate question, and manuscripts alone cannot answer it.1

What survives

The Institut für Neutestamentliche Textforschung (INTF) in Münster keeps the official catalog of Greek New Testament manuscripts, the Kurzgefasste Liste.  Its own 2023 tally:3

Category Description Count Share
Papyri Written on papyrus, mostly fragments, mostly early 135 2.4%
Majuscules Parchment, all-capital letters, roughly 3rd–10th century 283 5.0%
Minuscules Parchment/paper, cursive script, 9th century onward 2,860 50.2%
Lectionaries Church reading books with the text rearranged for services 2,422 42.5%
Total 5,700

That 5,700 counts catalog entries, not surviving objects.  The INTF itself flags 56 entries as destroyed and 105 as location unknown, which alone brings the figure down to 5,541.3  Jacob Peterson, in a book written by textual critics to correct their own side's apologists, works through the double-counted fragments, wrongly admitted items, and losses, and concludes:  "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."1

Three more honest qualifications:

  • Most manuscripts are partial. Only about 60 — about 1% — contain the whole New Testament. Most contain one section, often the Gospels, and many are fragments of a few pages or a few verses.1
  • Most manuscripts are medieval. 83% date from after AD 1000.1 The count peaks in the twelfth century. The early manuscripts are few and fragmentary; the complete ones are late.7
  • The number keeps growing. 167 new catalog numbers were added between 2019 and 2023 alone,3 and the Center for the Study of New Testament Manuscripts has photographed collections in over 50 countries and found roughly 60 previously unknown manuscripts in the process.19

Beyond Greek, the New Testament was translated early:  Latin, Syriac, and Coptic by the third century, Armenian and Gothic by the fifth.12  Popular apologetics adds these up to "over 25,000 manuscripts."  Jeremiah Coogan's correction, in the same self-audit volume:  "There are probably not ten thousand Latin manuscripts, and claiming twenty-five thousand New Testament manuscripts beyond the Greek is an exaggeration too big to keep using.  It is better to say there are a few thousand versional manuscripts and leave it at that."1  No comprehensive census of Latin Bibles exists, so the traditional "10,000 Vulgates" is a guess.1

There is also a popular claim that the whole New Testament, minus 11 verses, could be reconstructed from quotations in the church fathers alone.  Trace it to its source and it dissolves:  a man recalling, fifty years later, what a friend said Sir David Dalrymple told him around 1780, with the punch line "except seven or eleven verses, (I forget which)."  Dalrymple's actual notebooks survive in the National Library of Scotland.  He never finished, and his best collation was missing about 4,336 verses — more than half the New Testament.20  Patristic quotations are genuinely massive and genuinely useful for dating readings, but the "11 verses" story is a legend, and the reconstruction argument is circular anyway:  identifying a loose quotation requires already knowing the text.1

The oldest copies

Nothing survives from the first century.  The oldest piece is P52, a credit-card-sized papyrus scrap holding five verses of John 18.  Its discoverer dated it to AD 100–150, and the date "AD 125" hardened into apologetic folklore.  Brent Nongbri's 2005 study of the paleography concluded:  "any serious consideration of the window of possible dates for P52 must include dates in the later second and early third centuries."8  The John Rylands Library, which owns it, now allows "a date nearer to 200 AD."8

The careful census of the early material is Pasquale Orsini and Willy Clarysse's 2012 study, and its abstract deserves quoting because it corrects the standard apologetic in both directions:  "There are no first-century New Testament papyri and only very few papyri can be attributed to the (second half of the) second century."5  They allow about seven items from the second century:  P52, P90, P104, P30, the fragments P4+64+67, and the parchments 0171 and 0212.5

The first substantial manuscripts arrive around AD 200–300:

  • P46 — a codex of Paul's letters, usually dated around AD 200–225.5
  • P66 — most of John. Traditionally around AD 200; Nongbri has argued for as late as the fourth century, and the paleographer who defended the traditional method now accepts "a date between third and fourth centuries."21
  • P75 — much of Luke and John, traditionally AD 175–225, now extended by the same debates to possibly the early fourth century.21
  • P45 — the four Gospels and Acts in one codex, early third century, about 30 of its original 220 leaves surviving.5

The first complete New Testaments are Codex Vaticanus (c. AD 300–350) and Codex Sinaiticus (c. AD 325–360, the earliest fully complete one), both containing the whole Bible in Greek.22  Daniel Wallace's summary claim — "the New Testament is completely found within three hundred years of composition, with more than 43% of the verses attested within 125 years of its completion" — is fair on the traditional dates, but note that the 43% figure leans on the early datings of P45/P46/P66/P75 that Nongbri has put in play.23

This article deliberately uses the later, skeptic-friendly date ranges above.  The argument does not need the early ones.

Warning

Case study in overstatement:  "First-Century Mark."  In a 2012 debate with Bart Ehrman, Daniel Wallace announced a fragment of Mark dated to the first century — which would have made it the earliest New Testament manuscript by a wide margin.  He wrote that it had been "dated by one of the world's leading paleographers.  He said he was 'certain' that it was from the first century."24  Six years of hype later, the fragment was published as P.Oxy. 5345 (= P137):  five partial verses of Mark 1, dated to the late second or early third century.25  The Egypt Exploration Society explained that an old record card had suggested a first-to-second-century date, revised on rigorous comparison.6  Wallace publicly apologized:  "I must first apologize to Bart Ehrman, and to everyone else, for giving misleading information about this discovery."26  The affair got worse:  the Oxford scholar who had handled the fragment, Dirk Obbink, was accused by the Society of selling it stolen papyri, was sued by Hobby Lobby for $7 million over 32 fragments, and was arrested in 2020.2728  The lesson for apologists:  unverifiable claims about unpublished evidence eventually cost more credibility than they buy.

Compared with other ancient books

The comparison with classical literature is legitimate, but it must be done with current numbers and matching counting rules.  The table uses the updated counts published in a 2012 apologetics-journal audit and figures verified against the holding institutions.429

Work Written Oldest surviving copy Gap Copies
Homer, Iliad ~700s BC Fragments 3rd c. BC; first complete copy ~10th c. AD ~400 / ~1,700 yrs 1,757
Herodotus, Histories ~430 BC Fragments 1st–2nd c. AD; medieval copies 10th c. ~500 / ~1,400 yrs ~109
Plato, dialogues d. 347 BC AD 895 (Clarke Plato) ~1,240 yrs ~210
Caesar, Gallic War ~50 BC 9th c. AD ~900 yrs 251
Josephus, Antiquities AD 93 11th c. (Greek) ~950 yrs handful in Greek
Tacitus, Annals ~AD 116 ~AD 850 (books 1–6); ~1040 (books 11–16) ~730–925 yrs 2 medieval + ~31 15th-c.
Suetonius, Twelve Caesars ~AD 121 ~AD 800 ~680 yrs dozens, medieval
New Testament AD 50–95 Fragments ~AD 150–200; most books by ~300; complete ~350 ~60–260 yrs ~5,100–5,300 Greek + a few thousand versional

Three honest caveats, from the same textual critics:

  1. The counting rules differ. Apologists count every Greek New Testament scrap ("inclusive count") but count only the manuscripts classicists actually use ("functional count") for the comparison works — "comparing apples and oranges," as James Prothro puts it. Counted functionally, the New Testament figure "would be nowhere near five thousand," though it would still win.1
  2. The classical numbers were badly out of date. Generations of apologists repeated "643 manuscripts of the Iliad" from a 1963 count. The updated count is 1,757. Herodotus is ~109, not 8. Plato is ~210, not 7.4 Using stale numbers to win a comparison that current numbers still win is self-inflicted damage.
  3. The comparison proves less than it is asked to prove. "The comparative argument, by its nature, can only demonstrate that the New Testament has a better textual basis for scholars to work with than classical works, not that it has a perfect one."1

With those caveats made, the comparison still stands, and the other side grants it.  In his 2009 debate with James White, Ehrman was asked whether the skepticism of Misquoting Jesus applied to the rest of antiquity would mean "Misquoting Suetonius":  "Absolutely!  Scholars do this…  There is no scholar who is an expert in Suetonius or Cicero or the Gospel of Thomas who would tell you that we absolutely know what these texts originally said."  And:  "What you can't do is, then say, 'Well, then you can't trust any book from antiquity.'"2  The New Testament's textual evidence is not merely adequate by ancient standards;  it defines the top of the scale.

Half a million variants

Ehrman's most-quoted line is that "there are more variations among our manuscripts than there are words in the New Testament," with scholarly estimates of "200,000… 300,000… 400,000 or more."15  The line is true.  It is also, without its denominator, meaningless.

The only published estimate actually built from collation data is Peter Gurry's 2016 study, which counted the variants in the INTF's Text und Textwert sample (about 3% of the New Testament collated across the full manuscript tradition) and extrapolated:  "a reasonable estimate for the number of textual variants in the Greek New Testament (not including spelling differences) is about 500,000."9  So the real number is higher than Ehrman's.  Why does the estimate's author consider it unalarming?  Because the count is a product of two factors:  how error-prone copying was, and how many copies survive.  The Greek New Testament has about 138,000 words and about 5,300 manuscripts — hundreds of millions of copied words.  As Gurry's abstract puts it, the number "reflects the frequency with which scribes copied more than their infidelity in doing so."9  A text with three surviving copies cannot have many variants.  That is a fact about its poverty, not its purity.

What are the variants?  Daniel Wallace's standard breakdown:10

Type Share Example
Spelling and nonsense slips ≥75% A movable nu — the Greek equivalent of "a" vs "an"
Untranslatable word-order changes and synonyms most of the rest "Jesus Christ" vs "Christ Jesus"; Greek word order is free
Meaningful but not viable small A late reading with no chance of being original
Meaningful and viable <1% Mark 1:41: did Jesus feel "compassion" or "anger"?

Independent measures agree on the scale.  Westcott and Hort, 1881:  "the proportion of words virtually accepted on all hands as raised above doubt is very great, not less, on a rough computation, than seven eighths of the whole," with genuinely substantial variation "hardly… more than a thousandth part of the entire text."11  The Alands counted 62.9% of New Testament verses — 4,999 of 7,947 — in which seven rival critical editions agree word for word.30  K. Martin Heide measured the stability of the text across the manuscripts at an average of 92.6% of words in the test passages, and found that the Shepherd of Hermas — a non-canonical text copied over the same period — scores worse than the worst New Testament manuscript.31

None of this makes the variants uninteresting.  A few dozen are genuinely important, and they are the subject of the next section.

The passages actually in doubt

Compare a King James Version with any modern translation and about 40 verses seem to have vanished — 16 whole verses plus two twelve-verse blocks, along with several partial verses.32  They were not removed from the Bible;  they were added to it by later copyists, and the older manuscripts recovered since 1611 lack them.  Every modern translation flags each one with a footnote.  The main cases:

Passage Content The manuscript facts
Mark 16:9–20 Resurrection appearances, snake handling, "whoever believes and is baptized will be saved" Absent from the two oldest Greek copies (Sinaiticus, Vaticanus), the oldest Old Latin and Syriac copies, and ~100 Armenian mss. Eusebius and Jerome say it was absent from almost all Greek copies known to them. Irenaeus quotes it in the 2nd century, so the addition is early.33
John 7:53–8:11 The woman caught in adultery Absent from P66, P75, Sinaiticus, Vaticanus; no Greek commentator mentions it before the 12th century. Metzger: "the evidence for the non-Johannine origin of the pericope of the adulteress is overwhelming," yet "the account has all the earmarks of historical veracity."33
1 John 5:7–8 (Comma Johanneum) The only explicit Trinity proof-text: "the Father, the Word, and the Holy Spirit… and these three are one" In no Greek manuscript's text before the 14th century; quoted by no Greek father even during the Trinitarian controversies; entered Greek via a 1215 Latin council document and entered print via Erasmus's third edition.33
Luke 22:43–44 The angel and the sweat like blood Absent from P75, Vaticanus, and others; cited by Justin and Irenaeus in the 2nd century. Modern editions double-bracket it.33
Luke 23:34a "Father, forgive them" Absent from P75, Vaticanus, Codex Bezae and others, yet "bears self-evident tokens of its dominical origin" (Metzger). Double-bracketed.33
John 5:3b–4 The angel stirring the pool A gloss absent from P66, P75, Sinaiticus, Vaticanus.33
Acts 8:37 The eunuch's confession of faith A "Western" addition; no Greek manuscript before the 6th century, though Irenaeus knows the tradition in the late 2nd.33
Matt 17:21 and 15 similar verses Mostly harmonizations from parallel passages Copyists filled in wording from the parallel Gospel; e.g. Matt 17:21 imported from Mark 9:29.32

Notice the pattern.  The doubtful passages are additions — harmonizations, marginal traditions, liturgical smoothing — not deletions or rewrites of doctrine.  If modern editors are wrong to omit them, Bibles contain slightly too little, not too much.12

The Comma Johanneum deserves one more paragraph, since it is the closest thing to a doctrinally motivated corruption in the whole list.  It is a Trinitarian sentence that no Greek manuscript contains before the late Middle Ages and no Greek father quotes even when arguing about the Trinity — which dates it decisively, since no one fighting the Arian controversy would have skipped it had it existed.33  Two ironies:  first, the doctrine it "supports" was formulated at Nicaea in 325, a millennium before the sentence appears in any Greek text, so the Trinity plainly never rested on it.  Second, the famous story that Erasmus rashly promised to include it if a single Greek manuscript could be produced — and was tricked with a made-to-order forgery — is itself a legend, traced no earlier than 1818 by the Erasmus specialist Henk de Jonge:  "The current view that Erasmus promised to insert the Comma Johanneum if it could be shown to him in a single Greek manuscript, has no foundation in Erasmus' works."34  Bad stories accumulate on both sides of this debate.

The live scholarly disputes today are smaller still, and worth naming so the scale is visible:  whether Mark 1:41 says Jesus was "moved with compassion" or "moved with anger" (the 2011 NIV prints "indignant");33  whether 1 Thessalonians 2:7 reads "gentle" or "infants" (one Greek letter);33  whether 2 Peter 3:10 says the earth "will be found" or "will not be found" — where the latest critical edition prints a reading found in no Greek manuscript at all, following the early translations.35  That last case is worth remembering whenever someone claims text critics smuggle theology into the text:  the editors printed the harder reading against every Greek witness, in public, with their reasoning published.

Does any doctrine change?

Set the two twelve-verse blocks aside and run down the rest:  nothing in the list above is the unique support of any Christian teaching.  The resurrection does not depend on Mark 16:9–20;  Matthew, Luke, John, Acts, and Paul all report appearances, and even within Mark the empty tomb and the announcement "he has been raised" are in the undisputed text (Mark 16:6).  The Trinity does not depend on the Comma.  The deity of Christ does not depend on any disputed passage;  the texts skeptics would need to delete — John 1:1, John 20:28, Philippians 2:6–11, Colossians 1:15–20 — are textually secure.  The one practice that genuinely rests on a doubtful text is snake handling (Mark 16:18), and the churches that practice it should take the manuscript evidence seriously.

This is not an apologetic talking point;  it is the stated consensus of both sides:

  • Bruce Metzger, asked how many doctrines are jeopardized by variants: "I don't know of any doctrine that is in jeopardy." (The version of this quote that says "seriously dented" circulates widely but is not in the book. Use the real one.)13
  • D.A. Carson: "nothing we believe to be doctrinally true, and nothing we are commanded to do, is in any way jeopardized by the variants. This is true for any textual tradition."14
  • Daniel Wallace: "viable textual variants that disturb cardinal doctrines found in the NT have not yet been produced."36
  • Bart Ehrman, in the paperback appendix of Misquoting Jesus, answering a question about exactly this: "The position I argue for in Misquoting Jesus 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."15

Honesty requires adding Ehrman's own framing of that concession.  On his blog he argues the question is malformed:  doctrines are built from many passages, so no single variant could threaten one — "a textual variant by its very nature cannot 'threaten' any doctrine."37  His real claim is that individual variants change what individual passages teach, and that is true:  Mark 1:41 changes Jesus's emotion in one scene;  the ending of Mark changes how one Gospel closes.  The right conclusion is precise:  variants affect exegesis of particular verses;  they do not affect what Christianity teaches.

How the text is reconstructed

The King James Version's Greek base — the Textus Receptus — descends from the edition Erasmus rushed out in 1516, which he himself described as "precipitated rather than edited."  For most of the text he used two twelfth-century manuscripts from Basel.  For Revelation he had one manuscript, missing its last page, so he back-translated the final six verses from Latin into Greek — creating readings "that have never been found in any known Greek manuscript of these verses—but that are still perpetuated today in printings of the so-called Textus Receptus."38

That sounds like an indictment.  The remarkable thing is what happened when five centuries of discovery — Sinaiticus, Vaticanus, the papyri, five thousand catalogued manuscripts — were thrown at Erasmus's text:  not much.  Peter Williams's test case:  the first fourteen verses of John, 188 words, are letter-for-letter identical in Erasmus's 1516 edition, the Nestle-Aland editions used by scholars, the Byzantine-priority edition, the SBL edition, and the Tyndale House edition — five editions built on rival philosophies and manuscript bases a millennium apart.12  Erasmus even knew where the soft spots were:  of the 35 Gospel verses modern editors doubt, his own notes flag the uncertainty of about 27, including both twelve-verse blocks.  Of the pericope of the adulteress he wrote in 1527:  "The story of the adulterous woman is not contained in the majority of Greek copies."12

Two independent checks show why the text is this stable:

  • The early text was not rewritten in the fourth century. When P75 was published, its text of Luke and John — copied at least a century before Codex Vaticanus — turned out to agree with Vaticanus at about 85–89% of all points of variation, the highest agreement of any manuscript pair measured, against about 16% between the least similar pairs.17 The Alands' verdict: the theory "of thoroughgoing revisions of the New Testament text made in the fourth century, was no longer defensible."30 The great codices were not a new edited text; they were careful copies of a line already old.
  • Readings do not disappear. The Alands call it tenacity: "Once a reading occurs it will persist with obstinacy… there always continued to be a stream of the tradition (sometimes broad, sometimes narrow) which remained unaffected."30 This is why the disputed passages are known: the manuscripts disagree openly, in copies made under different empires and jurisdictions. Nobody was ever able to impose uniformity — which means nobody was ever able to impose a change.12 The papyri discovered over the last 116 years put the claim to an empirical test: as Wallace notes, across all 127 published papyri, "not a single new reading has commended itself as authentic" — new discoveries keep confirming readings already present in the known tradition rather than revealing a lost original.24
  • Since 2012 the field has gone computational: the Editio Critica Maior collates the full tradition and feeds a genealogical algorithm (the CBGM), which produced 34 changes in the Catholic Epistles and 54 in Acts in the latest critical editions — changes on the scale of "Jesus" vs "Lord" in Jude 5, not on the scale of paragraphs.35 In the Catholic Letters, the average agreement between any two of the 123 collated manuscripts is 87.6%.39

The debate-closing summary belongs to the skeptic.  White asked Ehrman whether, if Ehrman himself edited a Greek New Testament using all his own preferred readings, it would differ from the standard scholarly text less than the Textus Receptus does.  Ehrman:  "Yes."2  The most famous textual skeptic alive, editing with a free hand, would produce a New Testament closer to the church's critical text than the King James's base text is.

Could it have changed before our earliest copies?

This is the strongest skeptical argument, and it should be stated at full strength.  Ehrman's scholarly thesis is that "scribes occasionally altered the words of their sacred texts to make them more patently orthodox and to prevent their misuse by Christians who espoused aberrant views," and that "the vast majority of all textual variants originated during… the second and third centuries," when copying "was also the least controlled."4015  Helmut Koester went further:  "the text of the Synoptic Gospels was very unstable during the first and second centuries…  there is no guarantee that the archetypes of the manuscript tradition are identical with the original text of each Gospel."16  No manuscript survives to check the first hundred years directly.  That is simply true, and no honest treatment can wave it away.

The replies are evidence-based, not rhetorical:

  1. Where the claim can be tested, it fails. Frederik Wisse, writing in the same 1989 volume as Koester, checked the actual manuscript evidence for early ideological rewriting: "the claims of extensive ideological redaction of the Gospels and other early Christian literature runs counter to all of the textual evidence." The whole tradition attests exactly two major additions (the Mark and John blocks), and "none of these interpolations change in an appreciable way the theology of the writing in question." There is "no basis to assume" the undocumented earliest decades worked differently from every documented period.18 Koester's specific mechanism — that canonical Mark is a revision and the original survives only in Matthew and Luke's copies of it — is a hypothesis with no manuscript witness, of the kind skeptics would reject instantly if a Christian proposed it.
  2. The change window is short and shrinking. P52 and its second-century companions sit within roughly a century of composition — for John, plausibly within decades. Every discovery since Erasmus has narrowed the gap, and none has revealed a different text. Someone proposing radical change just before the earliest copies is, in Williams's phrase, "filling in gaps in the evidence using their imaginations rather than what we already know."12
  3. The logistics never existed. By AD 64 Tacitus reports "an immense multitude" of Christians in Rome; by 112 Pliny complains from northern Turkey that the temples "had been almost deserted."4142 The books spread with the movement — copied in Egypt, Italy, Greece, Syria — and from the second century circulated both as single books and as collections. A successful alteration would have had to reach every copy in every format in every province, before any of the tens of thousands of surviving patristic quotations froze the older wording. "The scenario of widespread deliberate change starts to become fantastical."12
  4. The copyists' behavior is measurable, and it was conservative. The scribes preserved wording that no forger or improver would bother with: "When you come, bring the cloak that I left with Carpus at Troas, also the books, and above all the parchments" (2 Tim 4:13, NRSVUE). Greetings to named individuals, travel logistics, requests for a coat — expensive to copy, useless to any later church, faithfully transmitted for fourteen centuries. These are the fingerprints of copyists who treated every word as fixed, not of editors shaping a story.

What this argument establishes is a strong induction:  every period of transmission that can be observed shows a stable, tenaciously copied text;  the burden is on the claim that the one unobservable period was different.  What it cannot do is turn induction into proof.  The gap is real.  It is the honest cost of dealing in ancient documents, and the New Testament pays less of it than any other text from antiquity.

Common objections

"Copying was a game of telephone — errors compounding in a chain." The analogy fails structurally.  Telephone is a single chain, oral, unrepeatable, and played for laughs.  Copying is a branching tree of durable documents:  a fourth-generation copy can be checked against a first-generation copy that still exists, and modern editors compare branches that split in the second century.  Errors in one branch do not enter the others — which is exactly how editors detect them.9

"There are more variants than words in the New Testament." True, and the count rises every time a new manuscript is found — because the variant count is mostly a count of manuscripts.  Divide the ~500,000 variants by 5,300 manuscripts and the tradition averages fewer than 100 non-spelling variants per manuscript copy of documents totaling 138,000 words — better than 99.9% per-copy fidelity, most of the residue being spelling.9  The right question is not "how many variants exist" but "how much text is in doubt," and the answer has been stable since 1881:  roughly an eighth debatable at all, a sixtieth worth printing in an apparatus, a thousandth substantial.11

"We only have copies of copies — no originals." No originals exist for any ancient text — no Plato, no Cicero, no Tacitus, no Qur'anic autograph.  Ehrman, on the record:  "What you can't do is, then say, 'Well, then you can't trust any book from antiquity.'"2  If copies-not-originals invalidates the New Testament, it invalidates ancient history.

"The church corrupted the text to support its doctrines." Then it did an inexplicably bad job.  The one clearly doctrine-motivated addition (the Comma Johanneum) fooled no Greek father, appears in essentially no Greek manuscript for a millennium, and was expelled from the text the moment scholars compared copies.33  The passages skeptics most need to be interpolations — the resurrection accounts, the high Christology — show no textual disturbance at all.  Ehrman's documented "orthodox corruptions" are one-word nudges that survive in some branches and not others, which is precisely how they were caught.40  A church with the power to rewrite its scriptures would not have left four Gospels that disagree with each other on the details.

"Constantine (or Nicaea) standardized the Bible and burned the rest." The manuscript evidence refutes this directly:  the copies keep disagreeing, before and after Constantine, in all the same small ways.  A successful imperial standardization would look like the opposite of what exists.  The council of Nicaea did not discuss the canon or the text at all — and the era's actual book-burnings (Diocletian's persecution) targeted Christian scriptures for destruction, not for editing.12

"Modern Bibles deleted 40 verses." Backwards.  Later copyists added them;  older manuscripts recovered since the KJV lack them;  modern editions print the older text and footnote the additions.  The reader of a modern Bible has more information than a KJV reader, not less:  every disputed line is marked as disputed.  And this is not a modern conspiracy — Erasmus flagged most of the same passages in the 1520s.12

"The early papyri are dated too early by apologists." Often they have been, and this article uses the corrected ranges:  no first-century papyri, about seven second-century items, P52 possibly as late as 200.5  Notice what the redatings cannot do:  even shifted a half-century later, the New Testament's gap and manuscript count remain an order of magnitude better than any classical author's.  The argument survives on skeptic-approved dates, which is why this article uses them.

"A well-copied text can still be false." Correct, and worth conceding in full.  Textual criticism establishes what the authors wrote, not whether the authors told the truth.  Peterson again:  "manuscripts alone cannot prove the truth of Christianity.  What manuscripts can do is provide evidence of a reliable text."1  The historical questions — authorship, dating, eyewitness contact, corroboration — are separate arguments treated in separate articles.  This article establishes only this:  whatever the first-century authors wrote, that is what the church still reads.

Sources

  1. 1.

    Myths and Mistakes in New Testament Textual Criticism.  IVP Academic.  Edited by Elijah Hixson and Peter J. Gurry.  2019.  An edited book of essays by evangelical textual critics, written to correct their own side's inflated apologetic claims.  It is the closest thing this field has to a self-audit.  The article draws on its chapters on the manuscript count, the classical comparison, the early translations and the church fathers.

    1. 1a.

      Chapter Math Myths: How Many Manuscripts We Have and Why More Isn't Always Better, pages 48–69:  Chapter by Jacob W. Peterson.

    2. 1b.

      Page 69:  The chapter's stated takeaway:  “Most manuscripts of the New Testament are only manuscripts of part of the New Testament, and providing an exact count of them is a fool's errand.  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.”  Page 52:  The chapter's starting figure: summing the catalog.  “results in a count of 5,885 manuscripts…  The 5,885 total just cited, for instance, is surely incorrect, as will be shown.” ↩ ↩ ↩

    3. 1c.

      Pages 62–63:  “If there are approximately fifty-three hundred manuscripts in total, 83 percent of them come from the year 1000 or later.  This does not mean the number of manuscripts from the second to tenth centuries is insignificant;  clearly the New Testament is well attested in this period.” ↩ ↩

    4. 1d.

      Pages 62–63:  “There are just over sixty manuscripts of the whole New Testament.  This leaves the other 99 percent containing only portions of the text.” ↩

    5. 1e.

      Pages 68–69:  “In the negative, manuscripts alone cannot prove the truth of Christianity.  What manuscripts can do is provide evidence of a reliable text.  A reliable text attested by thousands of manuscripts is just that:  a reliable text.  But a reliable text is not a guarantee of reliable content.”  The chapter adds:  “providing arguments for the trustworthiness of a text's actual claims is not something with which textual criticism can help.  Those types of important arguments must come from other fields of inquiry.” ↩ ↩

    6. 1f.

      Chapter Myths About Early Translations, pages 285–304:  Chapter by Jeremiah Coogan.

    7. 1g.

      Page 303:  “an exaggeration too big to keep using”. ↩ ↩

    8. 1h.

      Page 288:  “there is no comprehensive list of Latin New Testament manuscripts corresponding to the list of Greek New Testament manuscripts kept by the INTF, so we have no reliable estimate of the total number.” ↩ ↩

    9. 1i.

      Chapter Responsibly Comparing the New Testament to Ancient Works, pages 70–89:  Chapter by James B. Prothro, in the section titled.  “Myths About Classical Literature:  Responsibly Comparing the New Testament to Ancient Works”.

    10. 1j.

      Page 89:  “Scholars and apologists often count all the manuscripts for the New Testament that exist (an inclusive count), whereas classicists generally only count the ones they need to use (a functional count)…  Apologists' numbers too often reflect this inclusive count for the New Testament but a functional count for manuscripts of classical works and end up comparing apples and oranges.”  Page 81:  On F.F. Bruce's classical figures, the chapter asks:  “what if we counted the same for the New Testament?…  I do not know exactly what the number would be, but I know it would be nowhere near five thousand.” ↩

    11. 1k.

      Page 86. ↩

    12. 1l.

      Chapter Myths About Patristics: What the Church Fathers Thought About Textual Variation, page 232:  Chapter by Andrew Blaski.  On the reconstruct-from-quotations claim, the chapter says:  “The claim, then, appears to be a conflation of two unsubstantiated, unpublished nineteenth-century studies, and yet it has appeared in literally dozens of current apologetic texts.”  Pages 250–251:  The chapter's stated takeaway:  “The argument that we can reconstruct all but eleven verses of the New Testament from 36,289 quotations by the church fathers is not only false but it is a conflation of two different arguments that are both riddled with problems.  It should not be used.” ↩

  2. 2.

    White, James, and Bart D. Ehrman.  Did the Bible Misquote Jesus?  Coral Ridge Presbyterian Church, Fort Lauderdale.  Jan 21, 2009.  “Did the Bible Misquote Jesus?  Debate transcript”.  A 48-page transcript of a debate between James White and Bart D. Ehrman.  The article uses Ehrman's answers in it on the gap between the originals and the copies, and on whether his own edition of the Greek text would differ from the standard one.

    1. 2a.

      White asks:  “You describe this time period as 'enormous,' that's a quote.  Could you tell us what term you would use to describe the time period between, say, the original writings of Suetonius or Tacitus or Pliny, and their first extant manuscript copies.”  Ehrman answers:  “Very enormous.”  White asks:  “So ginormous would be a good one?”  Ehrman answers:  “Ginormous…  Ginormous doesn't cover it.  The New Testament we have much earlier attestation than for any other book or manuscript would be.” ↩

    2. 2b.

      White asks:  “So, it would be correct to write a book called Misquoting Suetonius?”  Ehrman answers:  “Absolutely!  Scholars do this…  There is no scholar who is an expert in Suetonius or Cicero or the Gospel of Thomas who would tell you that we absolutely know what these texts originally said.”  Earlier in the same exchange:  “What you can't do is, then say, 'Well, then you can't trust any book from antiquity.'  Okay?  Yes, right.  That's right.” ↩ ↩

    3. 2c.

      White asks:  “Peter Williams of Cambridge suggested that if you were to edit an edition of the Greek New Testament using all of your own decisions regarding textual variants, that it would differ less than from the Nestle-Aland/UBS platform than the Textus Receptus does.  Would you agree?”  Ehrman answers:  “Yes.” ↩

  3. 3.

    Leggett, Katie, and Greg Paulson.  “How Many Greek New Testament Manuscripts Are There REALLY? The Latest Numbers.”  INTF Blog.  Institut für Neutestamentliche Textforschung, Münster.  Sep 29, 2023.  A blog post by the institute that maintains the official catalog of Greek New Testament manuscripts, reporting the catalog's own tallies.  Several secondary sites repeat wrong per-category figures from this post.  The article uses its 2023 total of 5,700 catalog entries as the starting figure for its count.

    1. 3a.

      Tables As of Sept. 29, 2023:  The table gives 135 papyri, 283 majuscules, 2,860 minuscules, 2,422 lectionaries, and a total of 5,700.  The post cautions that the total.  “is only a snapshot in time”.  It also says.  “one GA number does not always correspond to a single manuscript”. ↩ ↩

    2. 3b.

      “There are currently 56 manuscripts in the Liste tagged as 'Manuscript Destroyed'…  105 manuscripts in this category [location unknown]…  subtract this number from the total number of manuscripts, which results in 5,541 manuscripts.” ↩

    3. 3c.

      “Just since 2019, an additional 167 Gregory-Aland numbers have been added (2 papyri, 3 majuscules, 81 minuscules, and 81 lectionaries.” ↩

  4. 4.

    Jones, Clay.  “The Bibliographical Test Updated.”  Christian Research Journal 35(3).  2012.  A journal article by an apologist who wrote to the holding institutions and current editors to update the classical manuscript counts that apologists had been repeating from the 1960s.  The article uses its updated counts in its comparison with classical literature. ↩

    1. 4a.

      “they have reported the number of manuscripts for Homer's Iliad to be 643, but the real number of Iliad manuscripts is actually 1,757”.  The Iliad figures come from Martin West's catalog of 1,569 papyri plus T.W. Allen's 188 parchment manuscripts.  Herodotus: 49 papyri plus about 60 medieval copies.  Plato: 210 in the Plato Microfilm Project.  Caesar: 251.  Thucydides: text based on 8 manuscripts plus papyri back to the 3rd century BC. ↩ ↩

  5. 5.

    Orsini, Pasquale, and Willy Clarysse.  “Early New Testament Manuscripts and Their Dates: A Critique of Theological Palaeography.”  Ephemerides Theologicae Lovanienses 88(4):443–474.  2012.  A journal article by a palaeographer, who studies ancient handwriting, and a papyrologist, who studies ancient papyrus.  Neither has an apologetic stake.  It critiques the dates that scholars have given early New Testament manuscripts, and it finds that the dates proposed by Comfort-Barrett and Jaroš are often too early.  The article uses it as the careful census of the early material.

    1. 5a.

      Section Abstract:  “There are no first-century New Testament papyri and only very few papyri can be attributed to the (second half of the) second century.  It is only in the third and fourth centuries that New Testament manuscripts become more common, but here too the dates proposed by Comfort-Barrett…  and Jaroš…  are often too early.” ↩ ↩ ↩

    2. 5b.

      Their second-century items: P52 (125–175), P90 (150–200), P104 (100–200), P30 (175–225), P4+64+67 (175–200), 0171, 0212. ↩

    3. 5c.

      They assign P46.  “to the early third century”.  “excluding dates in the first or the first half of the second century”.  Young Kyu Kim's claim of c. AD 80.  “has been widely rejected”.  Bruce Griffin's 1996 analysis put the most probable date at 175–225. ↩

    4. 5d.

      Kenyon's original early-third-century dating stands; about 30 of an original ~220 leaves survive, covering all four Gospels and Acts. ↩

  6. 6.

    Egypt Exploration Society.  P.Oxy LXXXIII 5345.  2018.  A news item from the society on the papyrus P.Oxy LXXXIII 5345; the fragment was excavated by Grenfell and Hunt around 1903, and a 1980s record card suggested "a late first- or early second-century date," but "after rigorous comparison with other objectively dated texts, the hand of this papyrus is now assigned to the late second to early third century AD."  The Society also states it never offered the papyrus for sale.  The article cites it for the Society's explanation of the revised date. ↩ ↩

  7. 7.

    Gurry, Peter.  “Distribution of New Testament Manuscripts by Century.”  Evangelical Textual Criticism (blog).  May 27, 2015.  A blog post with a chart of how many New Testament manuscripts were copied in each century.  The article uses it for the picture of a few early manuscripts and a peak in the medieval centuries.

    1. 7a.

      4 manuscripts from the 2nd century, 50 from the 3rd, 48 from the 4th; peak of 1,109 in the 12th.  Before AD 300: 54; before 400: 102.  Gurry follows the Alands in assigning split-century dates to the later century. ↩ ↩

  8. 8.

    Nongbri, Brent.  “The Use and Abuse of P52: Papyrological Pitfalls in the Dating of the Fourth Gospel.”  Harvard Theological Review 98(1):23–48.  2005.  A journal article by Brent Nongbri of Yale University arguing that scholars use the papyrus P52 in inappropriate ways.  The study argues that handwriting evidence, the study of which is called paleography, cannot fix the papyrus's date as early as many scholars have claimed, so the window of possible dates must reach into the later second and early third centuries.  The article relies on it for that conclusion.

    1. 8a.

      Page 46:  “What I have done is to show that any serious consideration of the window of possible dates for P52 must include dates in the later second and early third centuries.” ↩ ↩

    2. 8b.

      The John Rylands Research Institute and Library's own catalog commentary now allows.  “a date nearer to 200 AD”.  while noting no conclusive evidence. ↩

  9. 9.

    Gurry, Peter J.  “The Number of Variants in the Greek New Testament: A Proposed Estimate.”  New Testament Studies 62(1):97–121.  2016.  A journal article that counted variants in the Text und Textwert collations, about 3% of the New Testament collated across essentially the whole manuscript tradition, and extrapolated over the 138,020 words of the NA27 text.  It estimates about 500,000 non-spelling variants.  The article uses it as the best published estimate, built from actual collation data.

    1. 9a.

      “We suggest that a reasonable estimate for the number of textual variants in the Greek New Testament (not including spelling differences) is about 500,000.  This estimate—and we emphasize that it is still an estimate—is based on a sample size of about three percent of the entire Greek New Testament.” ↩ ↩ ↩

    2. 9b.

      Abstract:  “The result is a higher number than almost all previous estimates.  Proper use shows that the number reflects the frequency with which scribes copied more than their infidelity in doing so.”  The paper also identifies Ehrman's Misquoting Jesus as the first place the 400,000 figure appears attributed to "some scholars." ↩ ↩ ↩

  10. 10.

    Taylor, Justin.  “An Interview with Daniel B. Wallace on the New Testament Manuscripts.”  The Gospel Coalition.  Mar 22, 2012.  An online interview with Daniel B. Wallace, whose standard breakdown of the kinds of variants the article uses.

    1. 10a.

      “Spelling and nonsense readings are the vast majority, accounting for at least 75% of all variants…  The smallest category by far is the last category:  meaningful and viable variants.  These comprise less than 1% of all textual variants.  Yet, even here, no cardinal belief is at stake.” ↩ ↩

  11. 11.

    Westcott, Brooke Foss, and Fenton John Anthony Hort.  The New Testament in the Original Greek: Introduction, Appendix.  Harper & Brothers.  1882.  The introduction that two Cambridge divinity professors, Brooke Foss Westcott and Fenton John Anthony Hort, wrote to explain the Greek New Testament they edited.  It sets out how scholars compare manuscripts to recover the original wording, and applies that method to the whole text.  The article uses its fractions as its measure of how much of the text is in doubt.

    1. 11a.

      Pages 2–3:  “The proportion of words virtually accepted on all hands as raised above doubt is very great, not less, on a rough computation, than seven eighths of the whole…  setting aside differences of orthography, the words in our opinion still subject to doubt only make up about one sixtieth of the whole New Testament…  the amount of what can in any sense be called substantial variation…  can hardly form more than a thousandth part of the entire text.” ↩ ↩ ↩

  12. 12.

    Williams, Peter J.  Can We Trust the Gospels?  Crossway.  2018.  A short book for general readers that sets out the evidence for the trustworthiness of the four Gospels.  Its author is Principal of Tyndale House, Cambridge, and chairs the International Greek New Testament Project.  Chapter 6, "Has the Text Changed?", pages 111–122, covers how the Gospel text was transmitted, and the article draws on it for the doubtful verses, the early translations, and the stability of the text since Erasmus.

    1. 12a.

      Chapter 6:  Erasmus's Gospel manuscripts are now numbered 1 and 2, both twelfth century.  Of the 35 Gospel verses in his edition that scholars have since called into question, his manuscript 1 and his 1527 Annotationes show he already knew of the doubt attached to about 27, namely both twelve-verse blocks plus three of the eleven individually questioned verses.  On John 7:53–8:11, his 1527 Annotationes, page 234, says:  “The story of the adulterous woman is not contained in the majority of Greek copies.” ↩ ↩ ↩

    2. 12b.

      Chapter 6:  “From no later than the third century the Gospels were also translated into other languages:  Coptic, Latin, and Syriac;  from the fifth century, into Armenian and Gothic;  and by the turn of the first millennium, into Anglo-Saxon, Arabic, Georgian, and Old Church Slavonic, among others.” ↩

    3. 12c.

      Chapter 6:  “If many modern scholars are wrong in thinking that these verses should be omitted, it only means that modern copies of the Gospels might contain too little, not too much.” ↩

    4. 12d.

      Chapter 6:  “as we look at these opening fourteen verses of John's Gospel, a sequence of 188 words or 812 letters, we find no differences in these editions”.  The editions are Erasmus 1516, the German Bible Society editions of 1979–2012, Robinson-Pierpont 2005 (Byzantine priority), SBLGNT 2010, and the Tyndale House edition of 2017.  A quantitative comparison by Drayton Benner found the Tyndale House edition closer to the Nestle-Aland text than to any other edition. ↩

    5. 12e.

      Chapter 6:  “they show that Gospel manuscripts vary, and therefore there has been no successful attempt by rulers or scribes to make them all agree or to cover up debate.  Gospel manuscripts came from many different countries and were made under various jurisdictions.” ↩ ↩

    6. 12f.

      Chapter 6:  “It is those who suppose that major change occurred before our earliest manuscripts who are proposing a radical discontinuity between all the centuries we know about and the time immediately before our earliest copies.  We might say that they are filling in gaps in the evidence using their imaginations rather than what we already know.” ↩

    7. 12g.

      Chapter 6:  “The further the Gospels spread, the harder it would have been for anyone logistically to travel and change everyone's copies…  anyone wanting to change a Gospel would have had to change it in both media (collected and individual), as well as in a wide range of locations.  The scenario of widespread deliberate change starts to become fantastical.” ↩

  13. 13.

    Strobel, Lee.  The Case for Christ.  Zondervan.  1998.  A book whose chapter 3 interviews Bruce Metzger.  The article uses the interview for Metzger's answer on whether any doctrine is in jeopardy because of textual variants.

    1. 13a.

      Chapter 3, page 65:  “'How many doctrines of the church are in jeopardy because of variants?'  'I don't know of any doctrine that is in jeopardy,' he responded confidently.  'None?'  'None,' he repeated.”  The "seriously dented" wording often attributed to this interview does not appear in the book. ↩ ↩

  14. 14.

    Carson, D. A.  The King James Version Debate: A Plea for Realism.  Baker.  1979.  A book on the King James Version debate.  The article uses Carson's statement that no doctrine is jeopardized by the variants.

    1. 14a.

      Page 56:  “What is at stake is a purity of text of such a substantial nature that nothing we believe to be doctrinally true, and nothing we are commanded to do, is in any way jeopardized by the variants.  This is true for any textual tradition.  The interpretation of individual passages may well be called in question;  but never is a doctrine affected.” ↩ ↩

  15. 15.

    Ehrman, Bart D.  Misquoting Jesus: The Story Behind Who Changed the Bible and Why.  HarperSanFrancisco.  2005.  A bestselling book by the biblical scholar Bart Ehrman on the mistakes and intentional changes that ancient scribes made to the New Testament.  The article quotes it on three points:  that no Christian doctrine depends on a disputed reading, that the earliest copying was the least controlled, and the figure of 400,000 variants.

    1. 15a.

      Pages 89–90:  “some say there are 200,000 variants known, some say 300,000, some say 400,000 or more!  We do not know for sure because, despite impressive developments in computer technology, no one has yet been able to count them all…  There are more variations among our manuscripts than there are words in the New Testament.”  Page 10:  The introduction adds:  “Most of these differences are completely immaterial and insignificant.” ↩

    2. 15b.

      “the earliest period of textual transmission was also the least controlled.  This is when nonprofessional scribes, for the most part, were copying our texts—and making lots of mistakes in their copies.” ↩ ↩

    3. 15c.

      Page 252:  Paperback edition (2007), added Q&A, answering whether his conclusions conflict with Metzger's:  “The position I argue for in Misquoting Jesus 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.”  He continues that where one passage's reading changes,  “there are still other passages that could be used to argue for the same belief.  For the most part, I think that's true.” ↩ ↩

  16. 16.

    Koester, Helmut.  “The Text of the Synoptic Gospels in the Second Century.”  Gospel Traditions in the Second Century, 19–37.  University of Notre Dame Press.  Edited by William L. Petersen.  1989.  A chapter by Helmut Koester on the text of the Synoptic Gospels (Matthew, Mark and Luke) in the second century.  The article quotes his conclusion that the text was very unstable in the first and second centuries.

    1. 16a.

      Page 37:  “All of the evidence presented here points to the fact that the text of the Synoptic Gospels was very unstable during the first and second centuries.  With respect to Mark, one can be fairly certain that only its revised text has achieved canonical status, while the original text (attested only by Matthew and Luke) has not survived.  With respect to Matthew and Luke, there is no guarantee that the archetypes of the manuscript tradition are identical with the original text of each Gospel.” ↩ ↩

  17. 17.

    Fee, Gordon D.  The Significance of Papyrus Bodmer II and Papyrus Bodmer XIV–XV for Methodology in New Testament Textual Criticism.  University of Southern California.  1966.  A doctoral thesis, written in 1966 at the University of Southern California, that asks how scholars should measure how closely ancient manuscripts are related.  It tests this on two early papyrus copies of the Gospels, Bodmer II (P66) and Bodmer XIV–XV (P75).  The article uses it for the agreement figures behind its claim that the early text was not rewritten.

    1. 17a.

      Pages 37–38:  Across 61 variation-units (places where the manuscripts differ) in John 4, P75 and Vaticanus agree 85.2% (88.8% when more manuscripts are included in the tally), the highest agreement of any pair measured.  Vaticanus and Codex Bezae agree at 16.4% on the same passages.  The widely quoted "92% in John, 94% in Luke" traces to Philip Comfort's tallies. ↩ ↩

  18. 18.

    Wisse, Frederik.  “The Nature and Purpose of Redactional Changes in Early Christian Texts: The Canonical Gospels.”  Gospel Traditions in the Second Century, 39–53.  University of Notre Dame Press.  Edited by Petersen.  1989.  A chapter in Petersen's 1989 volume on the nature and purpose of redactional changes, meaning deliberate editing, in the canonical Gospels.  The article cites it for the finding that the manuscript evidence shows no extensive early ideological rewriting of the Gospels.

    1. 18a.

      Page 52:  “the claims of extensive ideological redaction of the Gospels and other early Christian literature runs counter to all of the textual evidence.”  Page 39:  The manuscript evidence attests.  “only two major redactional additions…  None of these interpolations change in an appreciable way the theology of the writing in question.”  Page 47:  “Since there is no basis to assume that the early, poorly attested history of the transmission of the text was governed by factors different from those operative in the canonical period, the existing textual evidence may be taken to be indicative of the nature and purpose of redactional activity from the beginning.” ↩ ↩

  19. 19.

    Center for the Study of New Testament Manuscripts.  Manuscripts 101: GA Numbers.  Oct 24, 2022.  “more than 5,700 extant”.  A web article from the center on its GA numbers, the catalog numbers used for New Testament manuscripts.  Its figure of more than 5,700 extant manuscripts repeats the entry count of the Liste, which slightly overstates the surviving total because entries include destroyed and missing items.  The article cites it for the point that the count keeps growing. ↩

  20. 20.

    Islamic Awareness.  Reconstruction Of The New Testament From The Church Fathers' Citations: Are The Claims Justified?  n.d.  A website hostile to Christian claims that checked the primary documents behind the claim.  It traces the anecdote to Robert Philip's 1841 biography of John Campbell, which recalls what Walter Buchanan said David Dalrymple told him around 1780, so the anecdote is third-hand.  Dalrymple's surviving notebooks in the National Library of Scotland show roughly 4,336 New Testament verses never located.  The article cites it because the correction rests on documents, not on the author's sympathies. ↩

  21. 21.

    Orsini, Pasquale.  “Palaeographic Method, Comparison and Dating: Considerations for an Updated Discussion.”  Evangelical Textual Criticism.  Feb 2018.  A guest post on a textual-criticism blog about how to date ancient manuscripts from their handwriting, by comparing them with securely dated documents.  The article cites it for the later dates that a scholar who defended the traditional method now accepts for P66 and P75.

    1. 21a.

      On P66, accepting comparisons with securely dated documents of AD 298 and c. 330–340:  “These comparisons are convincing for me, and for this reason I accept a date between third and fourth centuries.”  Nongbri's fuller argument is "The Limits of Palaeographic Dating of Literary Papyri," Museum Helveticum 71 (2014) 1–35. ↩

    2. 21b.

      On P75, comparing manuscripts of 317–325:  “I agree to extend the chronological terms, including the early fourth century.”  Nongbri's argument is "Reconsidering the Place of Papyrus Bodmer XIV–XV (P75)," Journal of Biblical Literature 135.2 (2016) 405–437. ↩

  22. 22.

    “Codex Sinaiticus and Codex Vaticanus.”  Wikipedia.  n.d.  An online encyclopedia article that gives the dates and histories of two codices, Sinaiticus and Vaticanus, which the article uses for the first complete New Testaments.  Sinaiticus is dated c. 325–360 and was found by Tischendorf at St Catherine's Monastery (leaves in 1844, main portion in 1859); Vaticanus is dated c. 300–350, was in the Vatican's catalog by 1481, and its New Testament lacks 1–2 Timothy, Titus, Philemon, Hebrews after 9:14, and Revelation, which a 15th-century hand supplied.  Brent Nongbri (Journal of Theological Studies, 2022) would extend Sinaiticus's window into the early fifth century.  Mirror:  En.wikipedia.org ↩

  23. 23.

    Wallace, Daniel B.  “The Bart Ehrman Blog and the Reliability of the New Testament Text.”  danielbwallace.com.  May 1, 2012.  A blog post by Daniel Wallace that discusses Bart Ehrman's blog and how reliably the New Testament text is attested.  The article accepts Wallace's summary for the traditional dates, but notes that its 43% figure leans on early dates that Brent Nongbri has put in play.

    1. 23a.

      “the New Testament is completely found within three hundred years of composition, with more than 43% of the verses attested within 125 years of its completion.” ↩

  24. 24.

    Wallace, Daniel B.  “Earliest Manuscript of the New Testament Discovered?”  Dallas Theological Seminary.  Feb 9, 2012.  The original announcement of a first-century fragment of Mark.  It was left online after the claim collapsed, to Wallace's credit.  The article quotes its claim that a leading paleographer dated the fragment.

    1. 24a.

      “It was dated by one of the world's leading paleographers.  He said he was 'certain' that it was from the first century.”  The same article claimed "as many as eighteen New Testament manuscripts from the second century and one from the first";  compare the roughly seven second-century items of Orsini and Clarysse. ↩

    2. 24b.

      “As with all the previously published New Testament papyri (127 of them, published in the last 116 years), not a single new reading has commended itself as authentic…  The original New Testament text is found somewhere in the manuscripts that have been known for quite some time.” ↩

  25. 25.

    Hixson, Elijah.  “First-Century Mark Published at Last?”  Evangelical Textual Criticism.  May 23, 2018.  A blog post that identified a long-rumored fragment of Mark with the newly published papyrus P.Oxy. 5345 (P137), which contains Mark 1:7–9 and 16–18, was edited by Dirk Obbink and Daniela Colomo, and is dated late second/early third century.  The article cites it for that identification and date. ↩

  26. 26.

    Wallace, Daniel B.  “First-Century Mark Fragment Update.”  danielbwallace.com.  May 23, 2018.  Wallace's own update on the fragment, in which he apologizes for his earlier announcement.  The article cites it for that apology.

    1. 26a.

      “I must first apologize to Bart Ehrman, and to everyone else, for giving misleading information about this discovery.”  Wallace explains he had signed a non-disclosure agreement and was urged to make the announcement by a representative of the fragment's would-be purchasers. ↩

  27. 27.

    Egypt Exploration Society.  Professor Obbink and Missing EES Papyri.  Oct 2019.  A news item from the Egypt Exploration Society on papyri missing from its collections.  Thirteen EES texts identified at the Museum of the Bible had been sold by Obbink to Hobby Lobby, and are described as "taken without authorisation."  The article cites it for the claim that Obbink, who had handled the fragment, sold stolen papyri. ↩

  28. 28.

    Silliman, Daniel.  “Hobby Lobby Sues Oxford Scholar.”  Christianity Today.  Jun 2021.  A news report on the lawsuit that Hobby Lobby brought against Oxford scholar Dirk Obbink.  The suit alleges Obbink stole about 120 papyri from the Egypt Exploration Society (EES) and sold 32 to Hobby Lobby for $7,095,100 between 2010 and 2013, including four "first-century" Gospel fragments for $1.8 million.  He was arrested in Oxford in March 2020.  The article cites it for the Hobby Lobby lawsuit over 32 fragments and for Obbink's arrest in 2020. ↩

  29. 29.

    Pearse, Roger.  “Tacitus and his manuscripts.”  Tertullian.org.  n.d.  “differ significantly”.  A web page on the oldest surviving copies of Tacitus and other classical authors, with a date for each.  Tacitus's Annals 1–6 survive in a copy written about AD 850 at Fulda, and Annals 11–16 in one written about 1038–55 at Monte Cassino; Suetonius's oldest copy is from the late 8th or early 9th century (Tours); Josephus's Antiquities survives in Greek no earlier than the 11th century (Ambrosianus 370); the earliest complete Iliad is the 10th-century Venetus A, with Ptolemaic papyrus fragments from the 3rd to 1st centuries BC; and the oldest substantial Plato is the Clarke Plato of AD 895.  The article cites these dates for the late copies of the classical texts it sets beside the New Testament's early copies. ↩

  30. 30.

    Aland, Kurt, and Barbara Aland.  The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism.  Eerdmans.  2nd edition.  1989.  A handbook to the critical editions of the Greek New Testament and to the methods of modern textual criticism, by Kurt Aland and Barbara Aland.  Kurt Aland founded the Münster institute and co-edited the Nestle-Aland text, and this is the field's standard handbook of its generation.  The article uses it for three of the Alands' findings: the share of verses on which the editions agree, the rejection of a fourth-century rewriting of the text, and the persistence of readings once they appear.

    1. 30a.

      Pages 28–29:  Of 7,947 New Testament verses, 4,999 (62.9%) are identical in all seven major critical editions compared, and the Alands describe them as agreeing.  “with no differences other than in orthographical details.” ↩

    2. 30b.

      Page 87:  The Alands, on the papyrus P75 (Papyrus Bodmer XIV–XV), say its.  “text proved to be so close to that of Codex Vaticanus (B) that the theory of recensions, i.e., of thoroughgoing revisions of the New Testament text made in the fourth century, was no longer defensible.  One of the main pillars supporting the dominant theory of New Testament textual history was now demolished.” ↩

    3. 30c.

      Pages 291–295:  The Alands on how stubbornly a reading survives once it appears:  “The transmission of the New Testament textual tradition is characterized by an extremely impressive degree of tenacity.  Once a reading occurs it will persist with obstinacy…  When an alteration was made in the text of the New Testament — however strategically important the text…  there always continued to be a stream of the tradition (sometimes broad, sometimes narrow) which remained unaffected, and this for purely technical reasons.” ↩

  31. 31.

    Heide, K. Martin.  “Assessing the Stability of the Transmitted Texts of the New Testament and the Shepherd of Hermas.”  The Reliability of the New Testament: Bart Ehrman and Daniel Wallace in Dialogue.  Fortress.  Edited by Robert B. Stewart.  2011.  “an average textual stability of 92.6% (of the words in the test passages), the spread of results for individual manuscripts ranging from 87.1% to 99.7%”.  “the worst value of the New Testament text, as represented by P45”.  These are the words of Larry Hurtado's summary of Heide's figures, not Heide's own wording.  A chapter in a 2011 book of dialogue between Bart Ehrman and Daniel Wallace.  It measures how stable the text is across the manuscripts, averaging 92.6% of the words in the test passages, and it tests the Shepherd of Hermas, a text outside the New Testament, which averages about 86%.  The article uses these figures as an independent measure of stability. ↩

  32. 32.

    Wikipedia.  List of New Testament verses not included in modern English translations.  n.d.  A tabulation of the 16 whole verses, 8 partial verses, and the two 12-verse blocks that modern translations leave out, with the manuscript evidence for each.  The decisions behind each entry are documented passage by passage in Metzger's Textual Commentary.  The article uses it for the count of about 40 verses that modern translations flag. ↩ ↩

  33. 33.

    Metzger, Bruce M.  A Textual Commentary on the Greek New Testament.  United Bible Societies.  2nd edition.  1994.  The published reasons for every major decision in the standard critical text, given passage by passage.  Metzger wrote it on behalf of the editorial committee of the United Bible Societies' Greek New Testament.  The article uses its verdicts on the disputed passages: Mark 16:9–20, John 7:53–8:11, 1 John 5:7–8 and Luke 22:43–44.

    1. 33a.

      “The last twelve verses of the commonly received text of Mark are absent from the two oldest Greek manuscripts (א and B), from the Old Latin codex Bobiensis, the Sinaitic Syriac manuscript, about one hundred Armenian manuscripts, and the two oldest Georgian manuscripts…  Clement of Alexandria and Origen show no knowledge of the existence of these verses;  furthermore Eusebius and Jerome attest that the passage was absent from almost all Greek copies of Mark known to them…  The earliest patristic witnesses to part or all of the long ending are Irenaeus and the Diatessaron.”  Many manuscripts that contain it carry scribal notes that older copies lack it. ↩

    2. 33b.

      “The evidence for the non-Johannine origin of the pericope of the adulteress is overwhelming.  It is absent from such early and diverse manuscripts as P66 P75 א B…  No Greek Church Father prior to Euthymius Zigabenus (twelfth century) comments on the passage, and Euthymius declares that the accurate copies of the Gospel do not contain it.”  Yet, Metzger writes:  “the account has all the earmarks of historical veracity.”  Its earliest Greek witness is Codex Bezae. ↩

    3. 33c.

      “The passage is absent from every known Greek manuscript except four, and these contain the passage in what appears to be a translation from a late recension of the Latin Vulgate…  The passage is quoted by none of the Greek Fathers, who, had they known it, would most certainly have employed it in the Trinitarian controversies (Sabellian and Arian).  Its first appearance in Greek is in a Greek version of the (Latin) Acts of the Lateran Council in 1215.”  The four Greek manuscripts are 12th–16th century, two with the passage only in the margin;  the earliest with it in the text is the 14th–15th-century ms 629.  Its earliest sure citation as Scripture is the Latin Liber Apologeticus (c. 385), from the circle of Priscillian of Ávila. ↩ ↩ ↩

    4. 33d.

      “The absence of these verses in such ancient and widely diversified witnesses as P69(vid) P75 א(1) A B T W…  strongly suggests that they are no part of the original text of Luke.  Their presence in many manuscripts, some ancient, as well as their citation by Justin, Irenaeus, Hippolytus, Eusebius and many other Fathers, is proof of the antiquity of the account.” ↩

    5. 33e.

      “The absence of these words from such early and diverse witnesses as P75 B D* W Θ…  is most impressive and can scarcely be explained as a deliberate excision…  At the same time, the logion, though probably not a part of the original Gospel of Luke, bears self-evident tokens of its dominical origin.” ↩

    6. 33f.

      “Ver. 4 is a gloss, whose secondary character is clear from (1) its absence from the earliest and best witnesses…  (3) the presence of non-Johannine words or expressions…  and (4) the rather wide diversity of variant forms in which the verse was transmitted.” ↩

    7. 33g.

      “Ver. 37 is a Western addition, not found in P45 P74 א A B C…  There is no reason why scribes should have omitted the material, if it had originally stood in the text…  the tradition of the Ethiopian's confession of faith in Christ was current as early as the latter part of the second century, for Irenaeus quotes part of it.”  Erasmus inserted it from a manuscript's margin, judging.  “that it had been omitted by the carelessness of scribes.” ↩

    8. 33h.

      Metzger's committee gives the reading its weakest confidence grade:  “It is difficult to come to a firm decision concerning the original text”.  It is easy to see why "being angry" would be softened to "being filled with compassion" but not the reverse; the committee kept "compassion" on the external evidence, noting Mark elsewhere reports Jesus's anger without scribes correcting it.  Ehrman argues for "angry"; the 2011 NIV prints "Jesus was indignant." ↩

    9. 33i.

      “From a transcriptional point of view it is difficult to decide whether νήπιοι ['infants'] arose by dittography…  or whether ἤπιοι ['gentle'] arose by haplography”.  A one-letter difference that the committee decided on external evidence. ↩

  34. 34.

    de Jonge, Henk Jan.  “Erasmus and the Comma Johanneum.”  Ephemerides Theologicae Lovanienses 56:381–389.  1980.  An article by Henk Jan de Jonge, a specialist editor of Erasmus's Apologiae, on the Comma Johanneum, the disputed sentence in 1 John 5:7–8.  De Jonge traces the "rash promise" story no earlier than 1818, and notes that the version he refutes is the one in Metzger's own handbook.  The article uses it for the story that Erasmus rashly promised to include the sentence if a single Greek manuscript contained it. ↩

  35. 35.

    Wallace, Daniel B.  “Nestle-Aland 28: The New Standard in Critical Texts of the Greek New Testament.”  danielbwallace.com.  Dec 17, 2012.  A blog post by Daniel B. Wallace introducing Nestle-Aland 28, a new edition of the Greek New Testament.  The article uses it for 2 Peter 3:10, where the new edition prints a reading that no Greek manuscript has, and for the 34 changes made in the Catholic Epistles.

    1. 35a.

      Reporting Klaus Wachtel's SBL presentation:  “The textual differences from the previous edition only occur in the Catholic Epistles…  A total of 34 textual changes have been made.”  Jude 5 now reads "Jesus" where NA27 read "[the] Lord."  The Editio Critica Maior work on Acts later produced 54 changes there. ↩

    2. 35b.

      “Perhaps the most controversial reading in the text of NA28 is found in 2 Peter 3.10:  οὐχ εὑρεθήσεται ['will not be found'].  This is not found in any Greek witnesses”.  The printed reading follows early Coptic and Syriac translations against every Greek manuscript. ↩

  36. 36.

    Revisiting the Corruption of the New Testament: Manuscript, Patristic, and Apocryphal Evidence.  Kregel.  Edited by Daniel B. Wallace.  2011.  A collection of six chapters, each by a different scholar, on how far copying damaged the New Testament text, with Wallace also writing the first chapter.  The article uses that chapter, "Lost in Transmission," for Wallace's own verdict on whether copying changed doctrine.

    1. 36a.

      Chapter Lost in Transmission, pages 54–55:  “Suffice it to say that viable textual variants that disturb cardinal doctrines found in the NT have not yet been produced.”  The same pages quote and cite the Ehrman paperback appendix at its page 252. ↩

  37. 37.

    Ehrman, Bart D.  “Do Textual Variants Really Matter for Anything?”  The Bart Ehrman Blog.  n.d.  A blog post in which Ehrman argues that no single variant can threaten a doctrine.  "Because of the way that theology works, a textual variant by its very nature cannot 'threaten' any doctrine."  The article cites it for Ehrman's own framing of his concession. ↩

  38. 38.

    Metzger, Bruce M., and Bart D. Ehrman.  The Text of the New Testament: Its Transmission, Corruption, and Restoration.  Oxford University Press.  4th edition.  2005.  A textbook in its fourth edition on how the New Testament text was copied, damaged and restored, by Bruce M. Metzger of Princeton Theological Seminary and Bart D. Ehrman of the University of North Carolina at Chapel Hill.  The article uses its account of how Erasmus's rushed Greek text, the base of the King James Version, was made.

    1. 38a.

      Pages 142–146:  Erasmus on his own first edition: it was "precipitated rather than edited"; Scrivener later called it "in that respect the most faulty book I know."  “For most of the text he relied on two rather inferior manuscripts from a monastic library at Basle…  both dating from about the twelfth century…  For the Book of Revelation, he had but one manuscript…  this manuscript lacked the final leaf…  Erasmus…  depended on the Latin Vulgate and translated the missing verses into Greek.”  The interpolation at Acts 9:6 likewise entered from the Vulgate.  “and is found in no Greek manuscript at this passage.” ↩

  39. 39.

    Gurry, Peter J.  “What Every Pastor Should Know about Developments in Textual Criticism, Part 2: A New Method.”  Phoenix Seminary.  n.d.  “The average textual agreement between all pairs of witnesses reaches 87.6 percent”.  In the Catholic Letters, one pair of witnesses agrees at 99.1 percent of variant points.  A plain-language web article describing the Coherence-Based Genealogical Method (CBGM), by a co-author of its standard introduction (Wasserman and Gurry, "A New Approach to Textual Criticism," SBL Press, 2017).  The article uses its figure of 87.6 percent agreement among the Catholic Letters manuscripts. ↩

  40. 40.

    Ehrman, Bart D.  The Orthodox Corruption of Scripture: The Effect of Early Christological Controversies on the Text of the New Testament.  Oxford University Press.  1993.  A scholarly book by Bart Ehrman on how early disputes over the nature of Christ, meaning who Christ was, changed the wording of New Testament manuscripts.  The article cites it for Ehrman's thesis that scribes altered the text to make it more orthodox.

    1. 40a.

      Section Preface:  “My thesis can be stated simply:  scribes occasionally altered the words of their sacred texts to make them more patently orthodox and to prevent their misuse by Christians who espoused aberrant views.”  Page 29:  “The vast majority of all textual variants originated during the period of our concern, the second and third centuries.” ↩ ↩

  41. 41.

    Tacitus.  “Annals.”  Livius.org.  Translated by Church and Brodribb.  c. 116.  Book 15, chapter 44:  “an arrest was first made of all who pleaded guilty;  then, upon their information, an immense multitude was convicted.”  A Roman history by Tacitus that describes Nero's persecution of Christians in AD 64.  The article uses it to show how many Christians were in Rome by then. ↩

  42. 42.

    Pliny the Younger.  Letters.  c. 112.  Book 10, section 96:  “the temples, which had been almost deserted, have begun to be frequented…  the established religious rites, long neglected, are being resumed.”  A letter in which Pliny reports from Bithynia on his repression of the Christians.  The article uses it for Pliny's complaint, by 112, that the temples had been almost deserted. ↩