Bart Ehrman's Main Arguments

"This oldest form of the text is no doubt closely (very closely) related to what the author originally wrote, and so it is the basis for our interpretation of his teaching." -- Bart Ehrman, Misquoting Jesus, 2005¹

"To be clear, I am not saying the opposite either—that historians can use the historical disciplines in order to demonstrate that Jesus was not raised from the dead." -- Bart Ehrman, How Jesus Became God, 2014²

Summary

Bart Ehrman taught the New Testament and early Christianity at the University of North Carolina at Chapel Hill from 1988 until he retired in 2025, and he has written both technical studies and best-selling books that challenge traditional Christian beliefs about the New Testament.³  He is not a mythicist and not a radical skeptic about Jesus.  He holds that Jesus certainly existed, that he was crucified, that some of his followers, Paul among them, came to believe soon afterward that he had been raised, and that the oldest recoverable text of the New Testament is "very closely" related to what its authors wrote.⁴¹⁵⁵

This article tests nine of his principal historical arguments.  For each it states his position in his own words, gives the strongest reply found, and says what the evidence settles and what it leaves open.  The arguments were chosen because he makes them often and they bear directly on Christian claims;  the examples inside each are the ones he puts forward himself, or the hardest case available.  The verdicts describe these arguments.  They are not a grade for all of Ehrman's scholarship, and they do not measure how often the New Testament is right or wrong.

1. The text. Ehrman is right that the manuscripts differ widely, that some changes were made for doctrinal reasons, and that some original wording is still undecided. His claim has an important qualification: in a 2007 interview he agrees that "the essential Christian beliefs are not affected" by variants, because other passages teach them, while insisting that "the theology of entire books" can turn on a single reading.⁶ His further conclusion, that a God who inspired the words would also have preserved them, rests on a premise about what God would do that no manuscript can test. Verdict: right on the facts. His evidence challenges claims to know every original word with certainty; it does not prove that no surviving text has all the original readings. It does not touch inerrancy defined, as the Chicago Statement defines it, for the original documents, though it leaves that view with some undecided places and an early period no one can measure.
2. Forgery. Ehrman is right that ancient writers who mention forgery almost always condemn it, and he deals with the main exceptions. On 1 Peter, the case tested here, the language and proposed late date raise real difficulties. Karen Jobes challenges the language argument with a measured comparison of its syntax.⁷ Whether Peter composed it, authorised help or was impersonated remains unsettled. The letter claims to come from Peter; the evidence examined here supplies no independent account of its composition. Secretaries are well attested as recorders and occasional editors; Richards finds no ordinary secretary composing for an author, while distinguishing such a secretary from a trusted colleague.⁸ Verdict: right about ancient attitudes; the step from there to "forged" has to be argued letter by letter, and it is strong for some letters and unsettled for others.
3. Gospel differences. On their most straightforward readings, Mark and John put the Last Supper and the crucifixion on different days. Three explanations are live: a harmonisation; John moved the date for theology, as Ehrman and the Christian historian Michael Licona both think; or the earlier Gospels turned a farewell meal into a Passover meal, as the historian John Meier thinks.⁹¹⁰ Verdict: unless the harmonisation holds, the accounts as written do not agree on the day. Which one moved it, and why, is not established. The difference concerns the setting of an event both writers report.
4. Memory. Ehrman concludes that "Much of the gist is correct," while "a lot of details— and in fact entire episodes" are not.¹¹ Memory research supports that general picture but measures no Gospel episode. Verdict: a moderate conclusion. Which episodes fall on which side must be argued case by case, and the possible cost to Christian claims includes whole episodes, not only details.
5. How Jesus became God. Ehrman places belief in Jesus' divine status in the earliest resurrection belief and a pre-existent Christ in Paul.²² The disputes are what kind of divine being Paul meant and what Jesus claimed for himself. Ehrman's reading of Christ as an angel who was made God's equal only after his death rests on a contested reading of Galatians 4:14, an unsettled phrase in Philippians 2 and a very broad use of the word "divine." Gordon Fee, for one, reads the same verse as a step up from an angel to Christ.¹² Verdict: right that the language developed. The angel reading is possible, not established. Jesus' own claims are argued both ways from indirect evidence, and this article does not settle them.
6. Burial and visions. Ehrman doubts that Jesus "was decently buried on the day of his execution in a tomb that anyone could later identify."² The replies make burial before nightfall likely and burial by Joseph of Arimathea more likely than not, from Christian sources that share material. His vision model deliberately leaves open whether anything real was seen. Verdict: his specific doubt is not refuted, and the opposite judgment is better supported but not secure. His model offers an account of the spread of belief. If a known tomb was found empty, it needs an additional explanation for that fact.
7. Miracles. His argument that a miracle is "by definition" the least probable event confuses how probable an event is before the evidence with how probable it is after.¹³ The strongest Christian reply read here, Timothy and Lydia McGrew's probability analysis, shows how strong the evidence would need to be, but it reaches its answer on stated assumptions, such as the general reliability of the Gospels and Acts, that are themselves disputed.¹⁴ Ehrman's later rule, that historians as historians cannot assume God, describes what historians of every belief can agree on. It is not shown to limit what a reasonable person may conclude. Verdict: the "by definition" argument fails, and the rule decides nothing about whether Jesus rose. Whether the evidence is strong enough is the open question.
8. Canon and the end of the world. Ehrman's account of the canon includes age, apostolic origin and wide use as well as orthodoxy. Bishop Serapion's letter on the Gospel of Peter shows doctrine and authorship judged together, so Ehrman's claim that orthodoxy controlled the rest is one reading of entangled evidence.¹⁵ His view that Jesus expected the end within his own generation rests on a theme found across the earliest sources; R. T. France's reading of Mark 13 as the fall of the temple is a serious alternative for that chapter. Verdict: the canon account is largely fair. The apocalyptic reading is strong but not settled, and if it is right, it tells against Jesus' authority as a prophet.
9. Popular and academic books. In the passages compared, his trade books make the same arguments as his academic books, sometimes with firmer wording. The debate opening on recoverability is stronger than the book passages compared. Verdict: for these passages the trade books do not contradict the scholarship; they differ in confidence and emphasis.

Overall.  On the arguments tested, Ehrman's factual premises mostly hold in the form he states them in his considered work, and several of the conclusions built on them are weaker than his strongest wording.  The cost to Christian belief is real and specific.  Claims to certain knowledge of every original word, to every New Testament self-attribution being true, or to every Gospel detail being a precise report face distinct challenges on this evidence.  A case built on an approximately recovered text, on the seven letters of Paul that Ehrman accepts, and on the early and widespread claim that Jesus was raised keeps its premises against these arguments.  It is not proved by the replies either.  The burial, the apocalyptic sayings and the weighing of miracle claims remain open.

Part 1: Who Ehrman is, what he affirms, and how he is tested here

His own position

Ehrman trained in textual criticism under Bruce Metzger at Princeton Theological Seminary,¹ co-authored the fourth edition of Metzger's standard handbook on the text,¹⁶ and has written both technical monographs and best-selling trade books.  He describes leaving evangelical Christianity, and he says plainly what moved him:  "it was the problem of suffering, not a historical approach to the Bible, that led me to agnosticism."¹⁷  He adds that he has "never, in print or in an interview, indicated that I lost my Christian faith because of textual variants".¹⁷  So his biography is not an argument either way, and this article does not use it as one.

He affirms more than many readers expect:

1. Jesus existed. "Whether we like it or not, Jesus certainly existed." He calls this the conclusion of "virtually every sane historian on the planet -- Christian, Jewish, Muslim, pagan, agnostic, atheist".⁵
2. The Gospels can be used as historical sources. They are "riddled with problems," he says, but "historians can never dismiss sources simply because they are biased." The question is whether such sources can "yield historically reliable information, once their biased chaff is separated from the historical kernel."⁵
3. Much of the story is remembered correctly. "Much of the gist is correct," and "We can know a good deal about Jesus's historical life based on what our sources say."¹¹¹¹
4. Belief in the resurrection began early. He writes that "it is a historical fact that some of Jesus' followers came to believe that he had been raised from the dead soon after his execution," and "the apostle Paul, claims quite plainly to have seen Jesus alive after his death."⁴ Of Paul his textbook says, "there is no doubt that he believed that he saw Jesus' real but glorified body raised from the dead."¹⁸
5. Belief in Jesus' divinity began early. He treats the first resurrection belief as "the first understandings that Jesus was a divine being," and he finds a pre-existent divine Christ in Paul's letters, which date from about AD 48 to 62.²²
6. The New Testament text is unusually well attested and largely recoverable. The oldest recoverable text "is no doubt closely (very closely) related to what the author originally wrote."¹ He told an interviewer that if he and Metzger had to agree on the original text, "there would be very few points of disagreement — maybe one or two dozen places out of many thousands."⁶ Asked in debate how the gap between Tacitus or Suetonius and their earliest copies compares with the New Testament's, he answered "Ginormous doesn't cover it. The New Testament we have much earlier attestation than for any other book".¹⁹
7. Historical criticism need not end in unbelief. "It is my firmly held view that a historical understanding of the Bible does not necessarily lead to the kind of agnosticism that I myself have adopted."¹⁷

He has also changed his mind in public, three times on points that matter here.  In 1999 he wrote that Jesus "probably was buried";  in 2014 he wrote that he no longer thinks the burial and empty tomb are "relatively certain historical data".⁴²  In 2014 he also announced that he now reads Galatians 4:14 as calling Christ an angel, and that he had long said Jesus is portrayed as God in John "but not, definitely not," in Matthew, Mark and Luke, a view he then revised.²²⁰  A changed mind is not a fault.  It does mean that a quotation from one decade cannot stand for his view in another.

What is tested, and what is not

The article covers nine arguments Ehrman makes often against traditional Christian claims about the New Testament:  the text and its changes;  forged and disputed authorship;  differences between the Gospels;  memory and oral transmission;  how belief in Jesus' divinity arose;  the burial and the appearances;  whether historians can say a miracle happened;  and, more briefly, the canon and Jesus' expectation of the end, which serve as premises for the others.  A tenth part compares his popular and academic books on the same claims.  It leaves out his arguments from suffering (God's Problem, 2008), his later books on heaven, hell and the Apocalypse, and his autobiography, except where they bear on how to read him.  The letters of James and Jude, 2 Peter, and the Pauline letters case by case are not assessed here.  Where a completed article in this series has tested a question in depth, this article states its finding as that article's finding and links it.

Within each argument the examples are the ones Ehrman himself puts forward, or the case that best tests his central inference.  So the verdicts are about these arguments in these forms.  They do not add up to a score for his scholarship as a whole.

Three rules govern the quotations:

1. His books before his debates. A book is his considered statement. A debate transcript is unscripted speech, sometimes transcribed by an opponent, and is used only where it adds something and is labelled as such.
2. His words, not reports of them. Several sayings attributed to him in circulation could not be checked against a source. They are not used, and they are listed in the quarantine record.
3. No motives. Some critics say his conclusions follow from a commitment to materialism. He denies it: the reason historians cannot prove a miracle, he writes, "is not that historians are required to be secular humanists with an anti-supernaturalist bias."² Whether his method works is judged here on its merits, not on guesses about why he holds it. The same rule applies to his critics.

Part 2: The text: changed copies and recoverable wording

What he claims

Misquoting Jesus (2005) opens with the problem that turned Ehrman's view of the Bible:  "Not only do we not have the originals, we don't have the first copies of the originals. We don't even have copies of the copies of the originals, or copies of the copies of the copies of the originals."  The copies differ so often, he says, that "there are more differences among our manuscripts than there are words in the New Testament."¹  He then adds at once that "Most of these differences are completely immaterial and insignificant," and that "In some places, as we will see, we simply cannot be sure that we have reconstructed the original text accurately."¹¹

His case that the differences matter rests on a smaller set.  Scholars disagree, he writes, "not only about minor matters that have no bearing on the meaning of a passage ... but also about matters of major importance, matters that affect the interpretation of an entire book of the New Testament."¹  In his conclusion:  "It would be wrong, however, to say—as people sometimes do—that the changes in our text have no real bearing on what the texts mean or on the theological conclusions that one draws from them."¹  His academic monograph, The Orthodox Corruption of Scripture (1993), states the thesis behind that claim:  "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."²¹

His strongest formulation came in a 2009 debate with James White:  "Do we have a reliable text of the New Testament? Are there places where the Bible misquotes Jesus? The short answer is there is no way to tell."  He went straight on to particular passages:  "There are passages that scholars continue to debate, is this the original text or not? And there are some passages where we will never know the answer."¹⁹  Later the same evening he made his most serious point, which no count of later manuscripts can answer:  "We don't know how often the earliest scribes changed their text."¹⁹

The numbers, and what they measure

In the most careful recent count, a textual variant is any reading in any manuscript that differs from another manuscript at the same place, leaving out spelling.  Each different reading at a place counts once, however many manuscripts carry it, and the reading an editor thinks original counts too.²²  Some older counts tallied the same difference again in every manuscript that had it, so published totals are not all measuring the same thing.  Ehrman reports estimates of 200,000 to "400,000 or more".¹  Daniel Wallace, an evangelical textual critic and his frequent debate opponent, gives 300,000 to 400,000 against "only about 140,000 words in the NT."²³  Peter Gurry's 2016 estimate, built from the largest collation data then available, is "about 500,000" excluding spelling.  About half of these are singular readings, found in one manuscript only, and Gurry concludes that the number "reflects the frequency with which scribes copied more than their infidelity in doing so."²²²²

Three points follow:

1. The count grows with the evidence. A newly collated manuscript can add previously unknown readings and another witness; one that repeats known readings need not add a new variant. More copying creates more opportunities for variation, which is Gurry's point. The comparison with Tacitus and Suetonius concerns the age of their surviving copies. Asked how the gap between the originals of Tacitus or Suetonius and their earliest copies compares with the New Testament's, Ehrman answered "Ginormous."¹⁹
2. The count does not measure importance. Wallace sorts variants by whether they change meaning and whether they could be original; "Less than 1 percent of all textual variants belong to this group," but he adds that "There are still hundreds of texts that are in dispute."²³ Ehrman makes the same point from the other side: "The importance of theologically oriented variations, on the other hand, far outweighs their actual numerical count."²¹
3. Importance depends on more than the number of places. The comparison of variants with words is true and vivid, but it says nothing about which places matter. What matters at a given place is how early and how widely each reading is attested, which reading best explains how the other arose, whether the change looks deliberate, and what the passage teaches. Ehrman's own next sentence says that most differences do not matter; the examples below are the ones he says do.

Six places Ehrman says matter

These are the passages Ehrman himself puts forward, in Misquoting Jesus and in debate, as changes that make a difference.¹¹¹⁹  They were chosen because they are his strongest, not because they are easy.

The table follows the World English Bible where it prints the reading, except that its John 1:18 wording "only born Son" is shortened to "only Son";  "angry," "the only God" and "apart from God" render alternative Greek readings.²⁴²⁴²⁴  The rows are documented in Metzger's commentary and the current Greek editions.^252525252525  The commentary has no entry at John 1:1 or 20:28.²⁵  On Mark 1:41 the editions still divide:  the Society of Biblical Literature's Greek New Testament prints Ehrman's "angry,"²⁶ while the 2021 Editio Critica Maior of Mark keeps "moved with compassion" and lists "angry" only in Codex Bezae and some Old Latin copies.²⁷  Ehrman told White he would put "angry" in his own text.¹⁹  Metzger's committee used Mark 3:5, where Jesus looks around "with anger" (μετ’ ὀργῆς), against the theory that copyists softened anger, since they left that verse alone.²⁵²⁶  Ehrman uses the same verse for the other side:  Matthew and Luke drop Jesus' anger from Mark 3:5 and 10:14, and both also lack any emotion in the leper story, which fits a copy of Mark that read "angry";  "Matthew and Luke have no qualms about describing Jesus as compassionate, but they never describe him as angry."¹  Neither use of Mark 3:5 decides the reading at 1:41.  Hebrews 2:9 is not a fringe reading:  Metzger records "apart from God" in "a rather large number of Fathers, both Eastern and Western" and a few manuscripts, and the SBL edition prints it.²⁵²⁶  The World English Bible, which follows the Byzantine Majority text in the New Testament, has no "nor the Son" at Matthew 24:36 and footnotes the critical text's reading.²⁶²⁶²⁴

What the examples show, and what they cannot

Ehrman is right that some changes were deliberate.  Matthew 24:36 is the clearest case, and the reasons are evidence, not Metzger's authority.  The words "are lacking in the majority of the witnesses of Matthew, including the later Byzantine text," but "the best representatives of the Alexandrian, the Western, and the Caesarean types of text contain the phrase";  and Matthew's "only" ("but my Father only") fits a sentence that had just excluded the Son.²⁵  An accidental loss has no obvious trigger, while a copyist troubled that the Son does not know the day had an obvious reason to drop the words.  The alternative, that copyists added the words from Mark 13:32, is what the committee judged less likely.  This is an inference to motive from the spread of the readings and the shape of the sentence.  It is not an observation of anyone's intention, but it is the best explanation of the evidence.  Ehrman's academic claim is careful:  scribes changed texts "occasionally," and his overall thesis is that proto-orthodox scribes did so in the second and third centuries.²¹  That is a real finding about how early Christians handled their Scriptures, and an apologetic answer that offers only examples of spelling slips does not meet it.

The examples also have limits.

1. The reading Ehrman prefers survives. In every case above, the reading he thinks original is still present in surviving manuscripts, and he argues for it by the methods other textual critics use. That shows the evidence for each decision is available to be weighed. It does not show that the original is known. Mark 1:41, John 1:18 and Hebrews 2:9 are still disputed between the major editions.
2. The changes he documents mostly strengthened beliefs already held. Ehrman's own summary is that "Scribes altered their sacred texts to make them “say” what they were already known to “mean.”" He also found that they "were apparently more inclined to “correct” or “improve” a passage than to interpolate into it a notion that was previously wanting."²¹²¹ So these changes show which beliefs scribes already held, not that the beliefs entered Christianity through copying. That does not make the changes harmless. A change can carry a belief held in one place into a text read everywhere, and give it the authority of an apostle's words.
3. What the examples change, and what they do not. Four effects need to be kept apart: removing a central doctrine from the New Testament; changing which passages support it; changing what one author teaches; and changing one story. None of the six removes a central doctrine; each doctrine at stake is taught in other passages whose text is not in doubt. Several do the second and third. The Comma is the one verse that says the three "are one." John 1:18 decides whether that verse calls Jesus God. Luke 22:43-44 bears on how Luke portrays Jesus facing death, and Hebrews 2:9 on how its author describes that death. This is Ehrman's own position. Asked in 2007 why core beliefs were in jeopardy, he answered: "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. What he means by that (I think) is that even if one or two passages that are used to argue for a belief have a different textual reading, there are still other passages that could be used to argue for the same belief. For the most part, I think that's true." Then: "But I was looking at the question from a different angle. My question is not about traditional Christian beliefs, but about how to interpret passages of the Bible," and "the theology of entire books of the New Testament are sometimes affected by the meaning of individual passages."⁶

What the examples cannot show is what happened before the first copies.  Ehrman's point that later agreement "don't tell you what the early manuscripts did" is correct.¹⁹  The surviving papyri begin in the second and third centuries.  A change made in the first decades and copied into every surviving line would leave no trace.  The evidence does not show that such changes happened, and it cannot show that they did not.  Metzger and Ehrman's handbook states the working assumption of the field:  "most textual critics would agree that our manuscripts for each biblical book ... in one way or another go back to a text that was produced—either authored or edited—and published at some specific time and place".¹⁶  That is a reasonable inference from the agreement of independent lines of copying, not an observation of the original.  The same handbook makes a stronger claim about one line of copying.  P75, a papyrus of Luke and John, has "a form of text very similar to that of Vaticanus," a codex of about AD 350.  Dating P75 and P66 to about AD 200, the handbook concludes that "proof became available" that the text of Vaticanus and its allies "goes back to an archetype that must be put early in the second century."¹⁶  If that holds, one major line of the text changed little for a century and a half, and the unmeasured period shrinks to the decades between the writing of each book and that archetype.  The proof rests on the papyri's dates, which are disputed.  Brent Nongbri argues that P75's handwriting and format fit the fourth century as well, so that "we should seriously consider the possibility that p75 was also produced in the fourth century and hence that the “B Text” is also a fourth-century development."²⁸  On his dating, P75 and Vaticanus are two copies of about the same age, and their agreement says little about the second century.  The New Testament Manuscripts article treats this gap in detail.

His theological inference

Ehrman draws one further conclusion, and it is theological.  He came to realise, he writes, that "it would have been no more difficult for God to preserve the words of scripture than it would have been for him to inspire them in the first place," and "if he didn't perform that miracle, there seemed to be no reason to think that he performed the earlier miracle of inspiring those words."¹  His conclusion gives the argument in full:  "For the only reason (I came to think) for God to inspire the Bible would be so that his people would have his actual words; but if he really wanted people to have his actual words, surely he would have miraculously preserved those words, just as he had miraculously inspired them in the first place."¹  Set out as steps:

1. The only reason for God to inspire the Bible would be so that his people would have his actual words.
2. If God wanted his people to have his actual words, he would have miraculously preserved them.
3. He did not preserve them.
4. So he did not inspire them.

The form is valid:  if the three premises are true, the conclusion follows.  But the premises are judgments of different kinds, and the manuscripts bear only on the third.

1. Premise 3 depends on what having the words means. On Ehrman's own figures the words are had approximately: the oldest text is "very closely" related to the original, and he and Metzger would disagree at "maybe one or two dozen places out of many thousands."¹⁶ So premise 3 is true of exact and certain possession of every word, and false of approximate possession.
2. Premise 2 must then say that God would secure exact and certain possession for every later reader. Ehrman applies that standard even to language: God would have given his people the words "in a language they could understand, rather than Greek and Hebrew."¹ Applied to readers of translations, that standard would require guaranteed access across languages as well as copies. Ordinary translation does not provide word-for-word identity with a Greek or Hebrew text. Whether God would want more than an approximately recovered text is a question about God's purposes, and history cannot settle it either way.
3. Premise 1 is the one Ehrman marks as his own conclusion ("I came to think"). A Christian who holds that inspiration's purpose is a trustworthy message, rather than exact wording in every later reader's hands, rejects it. Ehrman himself says that "the vast majority of Christians" never held the view of inspiration he was taught at Moody Bible Institute.⁶

Christian statements of the doctrine differ, and the evidence bears on them differently.  The 1978 Chicago Statement, an evangelical statement on inerrancy, says that "inspiration, strictly speaking, applies only to the autographic text of Scripture, which in the providence of God can be ascertained from available manuscripts with great accuracy," and that "God has nowhere promised an inerrant transmission of Scripture."²⁹  Ehrman's evidence agrees with the second sentence and is compatible with substantial recovery.  His agreement with Metzger measures their judgments, not independently measured accuracy against lost originals.  The Westminster Confession of 1646 says more:  the original-language texts were "by his singular care and providence, kept pure in all ages."³⁰  Read as a promise that the exact wording can be identified with certainty in one printed text, as some defenders of the Textus Receptus read it, that claim requires textual decisions which the manuscript evidence leaves disputed;  the series' How Textual Criticism Works tests those positions.  Read as preservation within the manuscript tradition as a whole, it is compatible with the surviving evidence.  That compatibility does not verify it, especially for the decades before the earliest copies.

Assessment

1. Right: the copies differ widely; some changes were deliberate and doctrinal; counts of variants measure how often the text was copied more than how much it matters; and a single reading can change how one writer portrays Jesus.
2. Partly right: some original wording remains uncertain. The uncertainty sits at places that can be listed and argued, and a gap before the earliest copies cannot be measured either way.
3. Not established: that any central Christian belief depends on a doubtful reading. Ehrman's considered statements do not claim it. His debate line "there is no way to tell" is stronger than his books (Part 10).
4. Not a historical result: his conclusion that God did not inspire the text, which turns on premises about what God would do.
5. Cost to a Christian: certainty about every original reading is not secured by the surviving evidence. This does not prove that the exact original is absent from every surviving text. Inerrancy of the original documents, and the claim that their text is substantially recovered, survive this evidence, with a residue of undecided places and an early period no one can measure. A verse whose wording is in doubt, such as the Comma or John 1:18, cannot carry an argument by itself.

Part 3: Forgery and disputed authorship

What he claims

Ehrman separates two cases that are often confused.  The Gospels are anonymous;  when later readers named them after Matthew or John, "the author is not trying to deceive anyone."  That is a "false ascription," not a forgery.³¹  A forgery, in his usage, is a writing whose author falsely claims to be someone else:  "In other instances it is because an author lied about who he was, claiming to be someone he was not."³¹  He puts six of the thirteen letters that bear Paul's name, and both letters of Peter among others, in that second group, in his trade book Forged (2011) and in his monograph Forgery and Counterforgery (2013).³¹³¹

His argument has three steps:

1. Ancient people condemned forgery. He finds that "in virtually every instance in which an ancient author mentions forgery, he condemns it." He concludes that forgers "intended to deceive their readers" and that ancient authors "considered it a form of lying and deceit."³¹³¹
2. Several New Testament letters differ from their supposed authors in style, vocabulary, theology or setting, enough to show another writer.
3. Secretaries do not explain the differences. Paul used scribes: "There is no doubt that the apostle Paul used a secretary on occasion," and Romans 16:22 names one, Tertius.³¹²⁴ But "Did the secretaries contribute to the contents of the letter? ... Despite what scholars often claim, all of the evidence we have suggests that the answer is no."³¹ "I don't know of a single piece of evidence or a single analogy to suggest that Peter or Paul used a secretary who significantly—or insignificantly, for that matter—added to the contents of the letter."³¹

Step 2 carries the weight for any particular letter.  Steps 1 and 3 decide only how a letter shown to come from another writer should be described.

Ancient attitudes

Ehrman's monograph collects the ancient statements and deals with the exceptions.  A presbyter in Asia who wrote the Acts of Paul was removed from office, Tertullian reports, although he said he had done it "from love of Paul."³²  The strongest evidence on the other side is the late philosopher Iamblichus, who says the followers of Pythagoras had "a fine custom" to "ascribe and assign everything to Pythagoras."  Ehrman answers that Iamblichus wrote "eight hundred years after Pythagoras," about one school, with no earlier source for the practice.³²  That answer is fair:  a late report about one school cannot establish a general convention, although it shows that at least one ancient writer could praise the practice.  The series article on ancient citation and authorship conventions tested the ancient sources and reached the same result on the main point:  the texts examined condemn deceptive authorship.  It also found that they do not show how every first audience read every pseudonymous work, and that composition authorised by the named author was possible.

Four practices need to be kept apart:

1. False ascription: an anonymous work later credited to a famous name. No deception by the writer.
2. Impersonation: writing as someone else in order to be believed to be him. This is what Ehrman calls forgery, and the ancient statements condemn it.
3. School attribution: crediting a teacher with a follower's work, as Iamblichus describes. Rarely approved in the surviving sources, and never shown for the New Testament's audiences.
4. Help that does not change the author: a named author who used a scribe, an editor, co-senders or assistants. The ancient condemnations do not touch this, and it is where the real dispute lies.

Secretaries, co-senders and helpers

The surviving evidence supports different kinds of help to different degrees:

1. Recording, certain. Tertius "who write the letter" (Romans 16:22) is one example among many.
2. Polishing, attested among the elite. Forged grants "some evidence, though it is very limited, that secretaries occasionally were asked to improve the author's style."³¹ The monograph grants that "This evidence is notably sparse, but it does indicate that a secretary would occasionally edit an author's letter stylistically," and he notes that Cicero speaks of a letter of Pompey's apparently "written (or rewritten?) instead by Sestius."³² Cicero sent the Greek version of his own account of his consulship to his friend Atticus to have its barbarisms corrected.⁸
3. Composition by a friend in the author's name, with permission, attested for Cicero. Cicero asked Atticus to write to several people in his name and "say whatever you think fit," and told him what to say if they looked for Cicero's seal or handwriting. Ehrman calls these "short, stereotypical letters" and the practice a deceit, since Cicero "instructs Atticus to lie about the matter."³² Their length is not stated.
4. Co-authorship, argued but not described. E. Randolph Richards, whose 1991 study of ancient secretaries is the one Ehrman answers, argues that "secretaries were used as co-authors," because the conventions of ancient letters "almost necessitates the inclusion of a secretary's expertise."⁸ Ehrman answers that Richards gives no example, and that another scholar who collected eighteen co-authored letters admitted "we have no way of knowing how the letters were composed."³²
5. Composition by an ordinary secretary, not attested in Richards's evidence. Richards himself concludes: "Nowhere was there any indication that an ordinary secretary was asked, much less presumed, to compose a letter for the author."⁸

Richards supports the objection to treating composition as an ordinary secretary's routine task.  His qualifier distinguishes ordinary secretaries from friends and peers;  it does not exclude authorised composition by a colleague or show how long every such letter was.  His wider sentence in Forged, that "all of the evidence" says secretaries did not contribute to content, goes further than the Cicero, Atticus and Sestius evidence he himself reports.

Two analogies bear on the dispute.  The first is Josephus, a Judaean priest whose native language was Aramaic.  He says he wrote his Jewish War "with the aid of some assistants for the sake of the Greek," and the War itself says it is a Greek version of books "which I formerly composed in the language of our country."³³³⁴  Ehrman cites Josephus's admission only to show how few Judaeans could write literary Greek.³²  It also shows something else:  an Aramaic speaker could publish a polished Greek book with help and still claim authorship while acknowledging that assistance.  It is help with a book, acknowledged in public, by a man with Greek education.  It does not show that any New Testament letter was made that way.

The second is Paul's co-senders.  Letters from more than one sender were rare:  Richards found "6 among the 645 papyrus letters" he checked, and none like Paul's.⁸  Paul names co-senders in most of his letters (1 Thessalonians 1:1, 1 Corinthians 1:1, Philippians 1:1), so the habit may mean that others shaped the letters.  But a co-sender cannot by himself explain why a disputed letter sounds different.  Timothy is named as co-sender in Colossians, and also in 2 Corinthians, Philippians and Philemon, which every side accepts as Paul's and which sound like him.²⁴  This is the article's own inference from the openings, not a published finding.

A worked case: 1 Peter

1 Peter is tested here because Ehrman argues it in detail, because the main replies are specific, and because no other article in this series takes it up.  His case in Forgery and Counterforgery has five parts:³²

1. Who Peter was. A Galilean fisherman in a province where few could read, fewer could compose, and fewer still in Greek; Ehrman concludes that Acts 4:13 "is probably right: Peter was illiterate."
2. What the letter is. Polished Greek with rhetorical devices "a refinement not found among most koine writers," quoting Scripture in its Greek version throughout.
3. When it was written. "Babylon," the place the author writes from (5:13), is a code name for Rome that "makes sense only after 70 CE, years after Peter's death."³² Persecution simply for being a Christian also points to a later time.
4. "Through Silvanus." The phrase in 5:12 names the man who carried the letter, not a secretary.
5. Help would not save the claim. The secretary theory has no evidence behind it, and "If he did commission such a work, this other person would be the author."³²

The replies, and how they fare:

1. Silvanus as secretary fails. Richards, who defends secretarial help in general, agrees with Ehrman on 5:12: "It will be argued here that this formula identified only the letter-carrier and was not used to identify the secretary."³⁵²⁴ Christians who cite 5:12 as evidence of a secretary are using a reading that Richards rejects.
2. Peter's education is a rate applied to a man. Acts 4:13 shows how Luke describes the council's view of Peter and John, men without scribal training; the series' Who Could Read and Write? article finds that secure as a description of the narrative but not a record of Peter's schooling. Low literacy rates make it unlikely that a given fisherman composed literary Greek; they cannot exclude it. But no surviving record establishes Peter's ability to compose this letter. Speaking Greek, writing it and composing an argument in it are different skills.
3. The Greek has a measured counter-case. Jobes compares grammatical patterns with Polybius, Josephus, 1 Thessalonians and Hebrews 5-9. Nine of fourteen usable criteria lean toward Semitic influence; her combined score for 1 Peter is +0.16 after excluding its scriptural quotations, compared with -1.38 for Josephus and -1.68 for Polybius. She argues that its author probably spoke Greek as a second language.⁷ This challenges the inference from polished writing to a native Greek writer. It does not identify Peter: Jobes explicitly limits what syntax can establish.⁷ Nor does excluding direct quotations exclude a wider influence from the Greek Bible. Her criteria require further testing, and language background does not establish literacy or training. These limits prevent either side from treating style alone as decisive.
4. Josephus answers part of Ehrman's fifth point, not his second. A work produced with help with its Greek can still be its named author's, as the Jewish War was, so the rule that the helper "would be the author" is too strong for that kind of help. Josephus does not establish how substantive composition from scratch was attributed; his acknowledged assistance is not a record of how 1 Peter was produced. It leaves Ehrman's point that the letter reads as an original Greek composition from someone at home in the Greek Bible.
5. A late date would exclude any lifetime helper. If "Babylon" for Rome and persecution for the name both require a date after Peter's death, which the tradition places in the 60s, then a helper working for Peter in his lifetime is excluded. Ehrman's Babylon argument is more than a first-attestation argument: Rome, like ancient Babylon, destroyed Jerusalem's temple, making the name especially apt after AD 70. That is a reason for a later date, not proof that the earlier imperial oppressor could never be called Babylon. The persecution argument also needs to distinguish local hostility from a later legal practice. Neither dates the letter merely by naming Rome or mentioning suffering.
6. The persecution argument rests on a reading of Tacitus that is firmer than Tacitus. Ehrman calls persecution "for the name" "a much stronger argument against Petrine authorship" than his other dating points. "There is nothing to suggest that the mere name “Christian” was ground for persecution in apostolic times," he writes, and "As Tacitus makes abundantly clear, Nero rounded up the Christians of Rome and subjected them to brutal treatment and execution not because they were Christians per se, but because of arson". He adds that the word "Christian" is otherwise not attested "until closer to the end of the first century, in the book of Acts."³² Tacitus's own account says something different. Nero blamed the fire of AD 64 on a group "called Christians by the populace," and after the first arrests "an immense multitude was convicted, not so much of the crime of firing the city, as of hatred against mankind."³⁶³⁷ Suetonius lists the same punishment among Nero's measures for public order and does not mention the fire: "Punishment was inflicted on the Christians, a class of men given to a new and mischievous superstition."³⁸ On both Roman historians' accounts, Christians in Rome in the 60s were a group known by that name and punished as a group, and on Tacitus's account the charge that stuck was not arson. 1 Peter describes that kind of hostility, not a court procedure: its readers are "spoken against as evildoers," it sets suffering "for being a Christian" against suffering "as a murderer, or a thief," and it tells them to "Honor the king."²⁴²⁴²⁴ The limits are real. Tacitus wrote about fifty years later, and "called Christians by the populace" may give the crowd the name used in his own day. "Hatred against mankind" is a charge about character, not the bare name that Pliny punished in about 112. Tacitus does not say what the first prisoners "pleaded guilty" to, the fire or the faith. And Rome is not the provinces of Asia Minor to which the letter is addressed. So this evidence does not date 1 Peter to Peter's lifetime. It does show that open hostility to Christians as Christians is attested in Rome in AD 64, the place and the persecution in which, as Ehrman notes, tradition puts Peter's death, so the persecution argument cannot exclude a date in Peter's last years.

Verdict on 1 Peter.  Ehrman raises substantial objections about education, composition and date.  Jobes supplies a reason to resist equating this Greek with native Greek authorship, without establishing Peter's authorship.  The evidence tested here does not decide between Peter's composition, authorised help and later impersonation.  The dating arguments, if accepted, would favour the last;  of the two, the persecution argument is weaker than Ehrman states it, because Tacitus and Suetonius report Christians punished as a named group in Rome in AD 64.  "Forgery" is a serious possibility, not a demonstrated result.  Jobes's full 2005 commentary remains unread;  the counter-case tested here is her published 2003 syntax study.

Paul's disputed letters

The series article Did Paul Write the Disputed Letters? weighed each letter separately, asking two questions:  was the letter sent in Paul's lifetime with his approval, and are the words his own?  Its verdicts, which are that article's and are not re-argued here, differ from letter to letter:

1. 2 Thessalonians: probably Paul's, at low to moderate confidence.
2. Colossians: Paul's more likely than not, at low confidence; the words probably not his own.
3. Ephesians: unresolved; its own wording doubtful.
4. 1 Timothy and Titus: the words probably not Paul's; whether he authorised them is unresolved.
5. 2 Timothy: open.

On the wording of the letters to Timothy and Titus those verdicts agree with Ehrman.  They also show that "forged" is not the only explanation the evidence allows for every disputed letter, and that a single verdict across all six does not fit the evidence.

What it costs

If a letter is a forgery in Ehrman's sense, its claim to come from Paul or Peter is false.  That conflicts with any doctrine of Scripture that treats every self-claim as true, and it weakens the test of apostolic authorship that the early church itself used.  It does not touch the historical evidence that rests on the seven letters that, in Ehrman's words, "Paul certainly wrote".³¹  Those letters carry the earliest testimony to the death, burial and appearances of Jesus and to the worship of Jesus in the decades after his death.  The Paul's Letters as Historical Evidence article uses only those seven, for that reason.

Popular and academic

The trade book and the monograph make the same argument.  The monograph asks "Why did an author choose to lie about his identity?" and promises to show "That forgers were consciously lying about their identity in an effort to deceive their readers".³²  Forged says it is the "layperson's level" version of that study.³¹  The monograph is fuller on secretaries and adds the Sestius report and the reply to Richards.  The trade book's "all of the evidence" sentence is firmer than the monograph's "notably sparse."

Assessment

1. Right: ancient writers who mention forgery almost always condemn it; the claim that pseudonymity was an accepted, harmless convention is weak; the examined evidence does not establish composition as an ordinary secretary's routine role.
2. Substantial but contested: the case against Peter's authorship of 1 Peter; its language must be weighed against Jobes's contrary analysis, and its date remains argued.
3. Not established: that the disputed Pauline letters are forgeries as a group. The evidence differs by letter, and for several it is unresolved.
4. Cost to a Christian: real for any view that every New Testament authorship claim is accurate. The central historical case, which rests on letters whose authorship is not in doubt, is unaffected.

Part 4: Differences between the Gospels

What he claims

In Jesus, Interrupted (2009) Ehrman says that students who study the Bible historically find it "chock full of discrepancies, many of them irreconcilable."¹⁷  His method is to read "horizontally," setting the same story from two Gospels side by side.  He is careful to say that ordinary, "vertical" reading is not wrong:  "There is absolutely nothing wrong with reading the Gospels this way, as this is no doubt how they were written to be read."¹⁷  His point is that the comparison shows each writer shaping the story, and that some differences cannot be smoothed away.

His lead case is the day Jesus died.  In Mark, the disciples prepare the Passover "when they sacrificed the Passover," Jesus eats it with them, and he is crucified the next day (Mark 14:12;  15:25).²⁴  In John, the priests will not enter Pilate's headquarters the next morning so that they "might eat the Passover" (John 18:28), and Pilate condemns Jesus on "the Preparation Day of the Passover, at about the sixth hour" (John 19:14), the day the Passover lambs were killed.²⁴²⁴  Ehrman:  "I do not think this is a difference that can be reconciled."  He explains it as a choice:  "In other words, John has changed a historical datum in order to make a theological point: Jesus is the sacrificial lamb."¹⁷¹⁷  His textbook says the same thing more cautiously:  "This analysis gives just one example of how historical facts may have been changed to convey theological “truths.”"¹⁸

The day of the crucifixion, weighed

Three explanations are live, and each has a cost.

1. Harmonisation. John's "Passover" can mean the whole festival week, so that the priests in John 18:28 want to eat the festival sacrifices of the following days, and "Preparation" was the ordinary word for Friday, as Mark 15:42 explains ("the Preparation Day, that is, the day before the Sabbath").²⁴ The series' Bible Contradictions article judges this a real reconciliation, held by careful scholars, but notes that it requires "eat the Passover" in John 18:28 to mean something other than its most natural sense. Ehrman's reply is that Mark "tells us what he means by this phrase: it is the Day of Preparation “for the Sabbath”" and not the day before the Passover meal.¹⁷ John ties the word to the Sabbath too, a few verses later: the bodies had to come down "because it was the Preparation Day, so that the bodies wouldn’t remain on the cross on the Sabbath," and Jesus was laid in the tomb "because of the Jews’ Preparation Day."²⁴ The word had that meaning outside the Gospels. A decree of Augustus quoted by Josephus excuses Jews from court "on the sabbath day, nor on the day of the preparation to it, after the ninth hour."³⁹ So John's "Preparation Day of the Passover" (παρασκευὴ τοῦ πάσχα) can be read as "the Friday of Passover week." This does not settle the case. On John's own chronology that Friday was also the day before the Passover meal, so John 19:31 fits either reading of 19:14. The harder text for the harmonisation remains 18:28, where the priests want to "eat the Passover."
2. John moved the date. This is Ehrman's view, and the Christian historian Michael Licona, who defends the Gospels' general reliability, agrees on the direction: "John appears deliberate in his attempts to lead his readers to think the Last Supper was not a Passover meal," and "John may have displaced the celebration of the Passover meal to have occurred one day later than we find in the Synoptics."¹⁰ Licona treats the move as a device of ancient biography rather than an error. He also quotes Robert Stein's judgment that it "is doubtful that any of the explanations has a particularly high degree of certainty."¹⁰
3. The earlier Gospels moved it. John Meier, in the first volume of his study of the historical Jesus, A Marginal Jew (1991), thinks John's date is the historical one. He begins: "It would be foolish to claim that the choice between the two chronologies is a clear or easy one." He concludes that "a number of considerations lead me to favor the basic outline of the Johannine chronology as the most likely," and that "In a sense, the Synoptics were the first to take that route, by making an “extraordinary” meal fit under the rubric of Passover."⁹ Among his reasons: a Passover amnesty for Barabbas makes most sense on the morning before the Passover meal, not after it, and the Synoptic tradition below its explicit Passover references does not describe a Passover meal.

Two facts limit Ehrman's version.  First, the theology he says John wanted was available without changing any date:  Paul, writing some twenty years after the crucifixion and before any Gospel, already says "Christ, our Passover, has been sacrificed in our place."²⁴  A writer could see that symbolism in a true date.  Second, "John has changed a historical datum" assumes that Mark's date is the datum and that John knew it.  Meier's reading reverses both.

What this case establishes, then, is limited.  The accounts as written do not agree on the day unless the harmonisation holds.  Which account, if either, moved the date, and whether by mistake, by an accepted literary convention or for theology, is not established.  And on any of the three explanations both writers agree that Jesus was crucified under Pilate in Jerusalem at Passover.

Three more cases

These are the cases Ehrman uses most, and they are among the cases the Bible Contradictions article left unresolved or judged plausible but unproved.  That article states the texts and every serious answer for each;  only the result is given here.

1. The hour. Mark has the crucifixion at "the third hour," about nine in the morning (Mark 15:25); John has Pilate still judging "at about the sixth hour."²⁴²⁴ The best answer appeals to rough time-telling and John's "about," but a three-hour gap stretches it. Unresolved.
2. Galilee or Jerusalem. Ehrman calls the instructions after the resurrection "irreconcilable": in Mark the women are told the disciples will see Jesus in Galilee, in Luke the disciples stay in Jerusalem.¹⁷ The appearances can be put in one sequence, but the angel's message cannot be fully squared. Unresolved.
3. Jairus's daughter. In Mark the girl is dying when her father comes; in Matthew she has already died. It is Ehrman's "one simple example" of a minor irreconcilable difference.¹⁷ The usual answer, that Matthew compresses the story and states the outcome at the start, is plausible but unproved. The printed page also contains a slip, "In Mark's version Jairus comes to Jesus because his daughter has already died," where Matthew's version is meant. It is one wrong word in a passage that otherwise describes Mark correctly, and it has no bearing on the argument.

What these cases show

They show that some differences resist harmonisation.  That is a fair finding.  The Bible Contradictions article, which worked through 629 published allegations grouped into 266 cases, found most of them answerable but left 34 cases unresolved and judged 3 to be probable errors.  It warns that these counts describe the lists it tested and are not an error rate for the Bible.

What they cost depends on what is claimed, and on how the case comes out.  Three levels need to be kept apart:

1. Uncertainty. If the right explanation cannot be identified, as here, a reader cannot claim either that the accounts are reconciled or that one is wrong.
2. Probable incompatibility. If the harmonisation is rejected, the accounts as written disagree, and at least one is not a literal report of the day. A doctrine that every reported detail is a precise report, in the modern sense, cannot absorb that. The Chicago Statement allows "non-chronological narration" and "the topical arrangement of material";²⁹ whether that covers a writer who moves the day of an event is a further question the statement does not answer.
3. Demonstrated error. Not knowing which writer was wrong would not rescue two demonstrably incompatible factual claims: at least one would be mistaken. Here the unresolved question is whether the apparent incompatibility admits a defensible reconciliation, not merely which writer departed from the facts.

They do not show that the Gospels are unreliable as historical sources, and Ehrman does not say they are.  His own standard is that biased sources can "yield historically reliable information, once their biased chaff is separated from the historical kernel."⁵  He grades the sources rather than discarding them, "1. The earlier the better," and he judges John "less historically reliable than the others" for that reason.¹⁷  A disagreement about the day of an execution is a disagreement between two accounts that both report the execution.  It bears directly on the handling of details.  It does not directly negate the shared event, although a demonstrated error can also affect how much trust a reader places in a source elsewhere.

Assessment

1. Right: some Gospel differences are unresolved, including the day and hour of the crucifixion.
2. Not established: that John changed a date he knew; the opposite direction is defended by a major historian, and a harmonisation remains possible.
3. Not established, and not claimed by Ehrman: that such differences make the Gospels generally unreliable.
4. Cost to a Christian: a view that every Gospel detail is a precise chronological report has an unresolved problem here. The shared report of crucifixion remains; broader judgments about each source's reliability can still be affected.

Part 5: Memory and oral tradition

In Jesus Before the Gospels (2016) Ehrman applies memory research to the decades between Jesus and the Gospels.  The stories, he argues, were told and retold, and memory changes in the retelling.  His conclusion is two-sided:  "there are gist memories of Jesus recorded in the New Testament Gospels that are almost certainly accurate. At the same time, there are a lot of details— and in fact entire episodes—that are almost certainly not accurate."  And:  "Still, many of the broad outlines that are narrated in the Gospels certainly did happen. Much of the gist is correct."¹¹  He ends the book with "Memory is not only faulty: most of the time it is pretty good."¹¹

His best-known illustration is the psychologist Ulric Neisser's study of John Dean, the White House counsel whose detailed Watergate testimony could later be checked against Nixon's tapes.  Ehrman quotes Neisser's finding that "hardly a word of Dean's account is true" for one meeting, and stresses that Neisser did not think Dean was lying.¹¹  He then grants Dean the level above the words:  "whether Dean had a decent gist memory probably depends on how broadly one defines “gist.” He knew he had a conversation with Nixon. He knew what the topics were."¹¹  Neisser's own summary went one level further:  "Except where the significance of his own role was at stake, Dean was right about what had really been going on in the White House."⁴⁰  Ehrman does not quote that sentence.  It supports his own two-sided conclusion rather than contradicting it, so leaving it out does not reverse what the example shows, though it leaves the reader with a harsher picture of Dean than Neisser's.

The series article Oral Tradition and Memory tested the studies on both sides of this debate and found:

1. Memory studies measure particular tasks in particular samples, such as students recalling where they heard news of a disaster, or one witness recalling meetings. They show that confidence and vividness do not guarantee accuracy, and that central items are kept better than details. They supply no rate that can be multiplied across forty years of ancient retelling.
2. Paul's letters give early evidence of received traditions and contact with leaders, but no study and no text identifies which Gospel details are accurate.

The disputed step is the transfer.  The Dean case measures one eyewitness recalling one meeting after nine months.  Ehrman's application imagines a report "written by someone who had heard about the conversation from someone who was friends with a man whose brother's wife had a cousin who happened to be there."¹¹  How many links actually separate each Gospel from the events is the contested question;  Dean measures neither a short chain nor a long one.

Ehrman's own earlier book supplies the best evidence on the length of the chain, at least for the core of the tradition.  Arguing against those who deny that Jesus existed, he stresses that Paul spent fifteen days with Peter in Jerusalem, about AD 36, and saw James, the Lord's brother (Galatians 1:18-19).  He calls that visit "one of the most likely places where Paul learned all the received traditions that he refers to," and adds:  "But it defies belief that Paul would have spent over two weeks with Jesus’s closest companion and not learned something about him—for example, that he lived."⁴¹²⁴  Fourteen years later Paul met Peter and James again, with John, and he says of the message about Jesus' death and appearances, "Whether then it is I or they, so we preach."²⁴²⁴  For the traditions Paul passes on, such as the list of appearances in 1 Corinthians 15:3-7 and the words over the bread and cup in 1 Corinthians 11:23-25, the chain is one link from Paul back to Jesus' closest disciple and his brother.  The Oral Tradition and Memory article reaches the same result:  Paul's contacts provide "a concrete opportunity for exchange rather than an anonymous chain of unlimited length."

This answer has two limits.  It covers what Paul repeats, which is a short core and very few stories;  Ehrman holds that "Matthew, Mark, Luke, and John did not write the Gospels,"¹⁷ so the number of links behind a particular Gospel episode is still unknown.  And a short chain is not the same as an accurate one.  Dean was himself the eyewitness, with no links at all, and still got the words of the meeting wrong.  A short chain supports the gist and the core claims, which is what Ehrman also grants;  it does not certify the wording or the details.

Assessment.  Ehrman's conclusion is close to what the evidence supports:  he affirms an accurate core and denies the accuracy of some details and some whole episodes.  The second half is a real cost to Christian claims, since a doubted episode can be an important one.  But the memory research does not say which episodes belong to which group;  each has to be argued from its own evidence.  His long-chain illustration does not fit the core tradition Paul repeats, which by Ehrman's own account in Did Jesus Exist? came one link from Peter and James;  for individual Gospel episodes the length of the chain stays unknown.

Part 6: How Jesus became God

What he claims, and what he does not

Ehrman does not date the origin of belief in Jesus' divinity to John's Gospel or the Council of Nicaea in 325.  How Jesus Became God (2014) argues four things:

1. Jesus did not proclaim his own divinity. The Synoptic Gospels and their sources contain nothing like John's "I and the Father are one":²⁴ "In none of them do we find exalted claims of this sort."² Miracles, pronouncing sins forgiven and receiving homage "are compatible with human, not just divine, authority." "What we can know with relative certainty about Jesus is that his public ministry and proclamation were not focused on his divinity; in fact, they were not about his divinity at all. They were about God."²
2. Belief in the resurrection made the first Christology, and it was already divine. "My overarching contention is that belief in the resurrection—based on visionary experiences—is what initially led the followers of Jesus ... to believe that Jesus had been exalted to heaven ... These beliefs were the first Christologies—the first understandings that Jesus was a divine being."² In these "exaltation" views Jesus is "elevated to a divine status on a level with God himself, sitting at his right hand."²
3. The exaltation view "very quickly morphed" into an incarnation view, already in Paul. An incarnation Christology "maintains that Christ was a preexistent divine being who became human before returning to God in heaven."² "They hold to exaltation Christologies, and Paul holds to an incarnation Christology."²
4. For Paul, the pre-existent Christ was an angel. "In this verse Paul calls Christ an angel," he writes of Galatians 4:14, and Paul "is not contrasting Christ with an angel; he is equating him with an angel." On this reading "virtually everything Paul says about Christ throughout his letters makes perfect sense. As the Angel of the Lord, Christ is a preexistent being who is divine; he can be called God".²² The poem in Philippians 2 then shows "a preexistent divine being. But he was an angel or an angel-like being, who only after his act of obedience to the point of death was made God's equal."²

So the dispute is not about whether the first Christians thought Jesus divine within a few years of his death.  Ehrman says they did.  It is about two narrower things:  what kind of divine being the earliest Christians thought he was, and whether Jesus himself claimed it.

Galatians 4:14 and the angel

The verse reads:  "you received me as an angel of God, even as Christ Jesus" (ὡς ἄγγελον θεοῦ ἐδέξασθέ με, ὡς Χριστὸν Ἰησοῦν).²⁶²⁴  Paul is recalling how the Galatians welcomed him when he was ill.  The two "as" phrases can be read as a ladder (as you would welcome an angel, indeed as you would welcome Christ) or, as Ehrman now reads them, as two names for one figure.  He credits the reading to Charles Gieschen and Susan Garrett.

Ehrman's case for identity is grammatical.  When Paul writes "but as ... as," he says, "he is not contrasting two things; he is stating that the two things are the same thing," and he cites 1 Corinthians 3:1 and 2 Corinthians 2:17.²  The first parallel fits:  "as to fleshly, as to babies in Christ" gives two descriptions of the same people.  The second is weaker, because 2 Corinthians 2:17 repeats "but" before its second phrase ("But as of sincerity, but as of God"), which reads naturally as a step up.²⁶²⁴  The construction allows identity;  it does not require it.

The replies:

1. Gordon Fee agrees with Ehrman on half of the question. He thinks "the evidence seems strongly to favor Paul's having picked up a common phrase from the Septuagint," so that the phrase means the angel of God, the figure who in some Old Testament stories "turns out to be the Lord himself." But he judges that "it seems unlikely that Paul is intending an absolute identification" and concludes that "we seem in fact to be dealing with progression here rather than identification, which would mean that Christ is a full rung higher than the angelic theophanies of the OT."¹²
2. Chris Tilling, in the five-author reply How God Became Jesus (2014), says Ehrman "uncritically adopts a disputed understanding of Gal 4:14" and "uses his disputed reading of Gal 4:14 as the interpretive key for Paul's entire Christology!"⁴² His larger complaint is that Ehrman's whole scheme, exaltation first and incarnation later, is an artificial grid, and that his word "divine" is stretched to cover God, angels and deified kings alike.⁴²
3. The rest of Galatians ranks angels low: an "angel from heaven" who preached another gospel would be "cursed" (1:8), and the law, "ordained through angels," is the lesser arrangement (3:19).²⁴²⁴ That weighs against Paul classing Christ with ordinary angels in this letter. It does not decide the case, because Gieschen, whose grammar Ehrman follows, reads the figure as the unique divine Angel of the Lord, not an ordinary angel, and Ehrman himself notes that this Angel "is in fact called God" in the Hebrew Bible.²

So the verse is genuinely disputed.  Even on Ehrman's reading it does not show that Paul ranked Christ below God before the resurrection;  that claim rests on his reading of Philippians 2.

What Paul's undisputed letters show

Ehrman's claim that the angel reading makes "virtually everything Paul says" fit is a claim about the whole corpus, and has to be tested text by text.  The series article Early Belief in Jesus' Divinity examined the texts that decide the question, all from letters Ehrman accepts as Paul's, written about AD 48 to 62:

1. 1 Corinthians 8:6 reworks the Jewish confession of one God, the Shema, into "one God, the Father, of whom are all things ... and one Lord, Jesus Christ, through whom are all things".²⁴ The article judges a deliberate Shema echo probable. It notes that the Jewish philosopher Philo also gives God's Word a role in creation, so the creation phrase alone does not separate God from an agent.
2. Prayer and invocation directed to Jesus, including the Aramaic maranatha, "Our Lord, come," are attested in these letters, each practice dated from its own letter.
3. Philippians 2:6 says that Christ, "existing in the form of God, didn't consider equality with God a thing to be grasped".²⁴ The meaning of the Greek word rendered "a thing to be grasped" is disputed, and so is whether Christ had equality before his exaltation. This is the verse on which Ehrman's "only after" depends, and the evidence does not settle it.

The article favours the reading that Paul included Jesus within the identity of Israel's one God over the reading that he was an exalted agent, without calling it decisive.  Tilling, for his part, grants Ehrman's main dating point:  New Testament authors "believed in the preexistence of Christ; they clearly did (and Paul did as well, as even Ehrman acknowledges)!"⁴²

Did Jesus claim to be God?

Ehrman's argument is strongest in its narrow form.  The explicit self-claims of John's Gospel, such as "before Abraham came into existence, I AM," are absent from Mark, Matthew and Luke,²⁴ and he asks:  "Did they (all of them!) just decide not to mention the one thing that was most significant about Jesus?"²  An argument from silence has force when the silence concerns something the sources would be expected to report, and this one does.

The reply is that the Synoptic claims are indirect, and the question is what they amount to.  Four kinds of evidence carry it:

1. Forgiving sins. In Mark 2 Jesus tells a paralysed man his sins are forgiven, and the scribes object, "Who can forgive sins but God alone?" Jesus answers that "the Son of Man has authority on earth to forgive sins."²⁴ Ehrman reads this as a priestly prerogative: "When Jesus forgives sins, he never says “I forgive you,” as God might say, but “your sins are forgiven,” which means that God has forgiven the sins ... Jesus may be claiming a priestly prerogative, but not a divine one."² Michael Bird answers that Jesus was no priest and was nowhere near the temple, and that the scribes do not ask "Who can forgive sins but a priest alone?"⁴²
2. The Son of Man. Ehrman argues that in Jesus' sayings about the coming Son of Man there is not "any hint that Jesus is talking about himself," as in Mark 8:38.² Bird answers that other sayings use the phrase as a plain self-reference, and that in Matthew 19:28 the twelve sit on thrones beside the Son of Man while Jesus, on Ehrman's reading, has no throne at all.⁴²
3. The Son. Sayings such as "No one knows the Son, except the Father; neither does anyone know the Father, except the Son" (Matthew 11:27) claim a unique relation to God.²⁴
4. "I have come." Simon Gathercole, in the same reply volume, argues that Jesus' habit of summing up his mission as "I have come to ..." (Mark 2:17, Matthew 5:17, Luke 12:49) matches the way angels announce their errands: "I found twenty-four examples in the Old Testament and Jewish tradition of angels saying, “I have come in order to ...” as a way of summing up their earthly missions. A prophet or a messiah in the Old Testament or Jewish tradition never sums up his life's work this way."⁴² If he is right, the Synoptic Gospels already portray Jesus as coming from heaven, against Ehrman's view that they have no pre-existent Jesus. The argument concerns how the Gospels portray Jesus; whether the sayings are Jesus' own words, and whether "I have come" in them means more than coming forward with a mission, are further questions the comparison does not settle.

Two points limit both sides.  First, Bird's stated aim is to show that Jesus saw himself as "a divine agent with a unique authority and a unique relationship with Israel's God," not that the Synoptic Jesus said he was God.⁴²  The live alternatives are therefore a human, messianic agent (Ehrman) and a unique agent who acts with God's own authority (Bird).  Second, what Mark portrays and what Jesus said are different questions:  the scribes' objection shows how Mark understood the scene;  whether the scene happened as told is argued separately.  Both are different again from whether any such claim was true.  This article does not settle the historical question.

Ehrman's own view of the Gospels' portraits has moved.  Until 2013, he wrote, he would have said that Jesus is portrayed as God "in the Gospel of John but not, definitely not, in the other Gospels of Matthew, Mark, and Luke."²⁰  The rest of that post, which states his revised view, is not publicly readable and is not used here.

Assessment

1. Right: Christian language about Jesus developed; the Nicene vocabulary of one being with the Father is fourth-century; the explicit divine self-claims of John are absent from the earlier Gospels.
2. Not his position: that belief in Jesus' divinity was late. Ehrman dates divine status to the first resurrection belief and personal pre-existence to Paul.
3. Possible, not established: that Paul thought of Christ as an angel who became God's equal only after his death. It rests on a disputed reading of Galatians 4:14, an unsettled reading of Philippians 2, and a broad definition of "divine." The inclusion reading is favoured in this series, not proved.
4. Open, and not settled here: what Jesus claimed for himself. The evidence is indirect, and it can be weighed; this article reports the main arguments without deciding them.
5. Cost to a Christian: little on dating, which Ehrman's own account supports. Early belief is not proof that the belief is true, and Ehrman explains the belief by visions whose cause he leaves undecided (Part 7).

Part 7: The burial and the appearances

What he claims

In 1999 Ehrman wrote that "Jesus actually did die, he probably was buried, and that some of his disciples (all of them? some of them?) claimed to have seen him alive afterward."⁴  By 2014 he had changed his mind about the burial and said so on page 7 of How Jesus Became God.²  His target is specific.  He does not say Jesus was left unburied;  he writes that "As a historian, I do not think we can say definitively that this tradition is false," and that it is too much to say Jesus was eaten by dogs.  He doubts one thing:  "it is highly unlikely that Jesus was decently buried on the day of his execution in a tomb that anyone could later identify."²²  His reasons are Roman policy and Pilate's character ("If the Romans followed their normal policies and customs, and if Pilate was the man whom all our sources indicate he was");  an older tradition in Acts 13:29 in which the Jewish council buries Jesus, where "There is no word of Joseph here, any more than there is in Paul's letters";  and the storyteller's need for a known tomb, since "If there was no empty tomb, Jesus was not physically raised."²²²

On the appearances he proposes a small start:  "My tentative suggestion is that three or four people—though possibly more—had visions of Jesus sometime after he died," most likely Peter, Paul and Mary Magdalene.²  He uses "vision" neutrally:  "By “vision” I simply mean something that is “seen,” whether it is really there or not."  He thinks it would not be "a historical sin at all" to leave open whether the visions were veridical.²²  The belief then spread and the stories grew.  This is not the claim, sometimes attributed to skeptics, that every follower had the same hallucination.

The replies

1. Roman practice in peacetime Judea. Craig Evans argues that "Peacetime administration in Palestine appears to have respected Jewish burial sensitivities," citing Philo and Josephus, and that Pilate, who had backed down over the military standards in Jerusalem, would not lightly have defied Jewish law by leaving bodies up at Passover.⁴² This answers Ehrman's general appeal to Roman custom. It supports burial before nightfall. It does not by itself show burial in a known, private tomb rather than a common grave for criminals. Ehrman's side has a reply of its own: Jesus was executed on a political charge, and a later Roman legal source says burial was refused especially for treason. The Empty Tomb article weighs both points and finds that no source gives a rate.
2. Joseph and the formula. Dale Allison calls Ehrman's argument from the silence of 1 Corinthians 15 "ill-considered": "he was buried by Joseph" cannot be built as a short parallel to the two-word Greek "he appeared to Cephas."⁴³ The Empty Tomb article notes the limit of this reply: it answers a demand for an exact verbal parallel, not the broader point that a short formula leaves things out. Allison concludes that "a man named Joseph, probably a Sanhedrist, from the obscure Arimathea, sought and obtained permission from the Roman authorities to make arrangements for Jesus' hurried burial," while granting "the evidence is imperfect".⁴³

The steps, and how they depend on each other

The question is not simply burial or no burial, but a chain of steps, each resting on the one before:

1. Jesus' body was taken down and buried in some way before nightfall. Evans's evidence about peacetime practice and the Jewish law against leaving a body overnight make this likely.
2. It was buried by a named member of the council, Joseph of Arimathea. The Empty Tomb article judges this more likely than not, held with caution. The story needed a man with the power to obtain the body, which an inventor could also supply; what gives Joseph some weight is the awkward detail, a member of the council that condemned Jesus, from an obscure town. The earliest accounts naming him are Gospel narratives whose degree of independence is disputed.
3. The place was known to Jesus' followers. Likely if Joseph or another known person buried him.
4. Women found it empty. More likely than not, tentatively, on the same shared sources.

Each step depends on the ones before, so the chain is no stronger than its weakest link, and several steps share their sources.  No non-Christian source mentions Jesus' burial at all.  Ehrman's doubt targets steps 2 and 3.  The evidence specific to Jesus, chiefly the awkward details about Joseph, favours them modestly.  It does not make them secure, and the Empty Tomb article's judgments on them are that article's, reported here rather than re-argued.

In the 2006 debate with William Lane Craig, Ehrman offered a story in which relatives moved the body and soldiers threw it into a common grave.  He said plainly that it was "just one I dreamt up" and that he did not believe it;  his point was only that any such story is more probable than a miracle.¹³  It is not his historical reconstruction and should not be quoted as one.

Assessment

1. Not refuted: Ehrman's doubt about a decent same-day burial in an identifiable tomb. The replies make burial before nightfall likely and burial by Joseph more likely than not, on evidence that is all Christian and partly shared; the opposite of Ehrman's view is better supported, but not secure.
2. The vision model and the tomb. The model explains how belief could begin with a few experiences and spread. It is neutral on whether the experiences had a real external cause. It does not explain a missing body; a visionary explanation of belief could nevertheless coexist with a separately emptied tomb. If a known tomb was found empty, the model needs a further explanation for that fact, and so does any natural account.
3. What history alone can say. The Hallucination Theories article finds that grief experiences could explain a single early report but that nothing shows they did, and that the surviving evidence leaves the cause of the appearances undetermined. Ehrman, who refuses to say whether the visions were veridical, agrees with that limit.

Part 8: Can a historian say a miracle happened?

Two arguments, not one

Ehrman has argued that historians cannot establish a miracle in two different ways.  They are often run together, and they need separate answers.

The probability argument appears in his 1999 book, his textbook, the 2006 Craig debate and Jesus, Interrupted.  In the debate:  "Miracles are not impossible. I won't say they're impossible."  But "Historians can only establish what probably happened in the past, and by definition a miracle is the least probable occurrence. And so, by the very nature of the canons of historical research, we can't claim historically that a miracle probably happened. By definition, it probably didn't."¹³¹³  His textbook puts it as a contradiction:  historians "cannot show that miracles happened, since this would involve a contradiction—that the most improbable event is the most probable."¹⁸  He applies the rule to every religion's miracles alike, "Mohammed, Hanina ben Dosa, Apollonius of Tyana — and Jesus."¹⁷

The access argument runs alongside it.  In the same debate he said:  "Historians, I'm sorry to say, have no access to God."  Answering Craig's formula, he repeated:  "If he wants to mount mathematical evidence for what God probably did in the world, I have to say it's not going to be convincing to most mathematicians and certainly to most historians. Historians have no access to God."¹³  In How Jesus Became God (2014) this becomes a rule about the discipline.  History, he writes, "is based only on shared presuppositions," and belief in a miracle is "rooted in a particular set of theological beliefs ... Since historians cannot assume these beliefs, they cannot demonstrate historically that such miracles happened."²²  He insists that this is a limit, not a denial:  "I am not saying the opposite either—that historians can use the historical disciplines in order to demonstrate that Jesus was not raised from the dead."²  And he rejects the charge that his reason is unbelief:  it "is not that historians are required to be secular humanists with an anti-supernaturalist bias."²

The probability argument fails as stated

The probability of a hypothesis has to be judged on the evidence.  Bayes's theorem, which William Lane Craig set out in the 2006 debate, separates two things:  how probable the resurrection was before looking at the evidence (the prior), and how probable it is after (the posterior).¹³  The posterior also depends on how well the resurrection explains the evidence and how well every alternative explains it.  Craig's point:  "The probability of the resurrection could still be very high even though the Pr(R/B) alone is terribly low."  Ehrman's textbook sentence treats the two probabilities as one, which is why Craig said "there's no contradiction here at all because we're talking about two different probabilities".¹³¹³

A worked illustration shows the structure.  Suppose, only for illustration, that a resurrection has a prior probability of one in a million, that the evidence would arise half the time if it happened, and that every natural alternative together would produce that evidence one time in ten million.  The posterior is then about 0.83.  Change only the last number to one in a hundred thousand and the posterior falls to about 0.05.⁴⁴  None of these numbers is measured, and the illustration does not show that the real evidence can supply the ratio the first case needs.  It shows only that a very low prior does not decide the answer by itself;  the comparison with the alternatives does.

Two more defects follow.  Michael Licona notes that "Ehrman links probability to the frequency of an event,"⁴⁵ so the prior is set by counting how often dead men rise naturally.  Ehrman's own example is walking on water:  "none of us can, so let's say the chances are one in ten billion."¹³  But the Christian claim is not that Jesus rose naturally.  As Craig put it, "It's the hypothesis that Jesus rose supernaturally from the dead. It is not the hypothesis that Jesus rose naturally from the dead."¹³  The prior for that hypothesis depends on whether God exists and might act, which a count of natural events cannot measure.  Second, if historians cannot speak about God, they cannot say that an act of God is improbable either.  Licona puts it bluntly:  Ehrman's third argument "cuts the legs off his second".⁴⁵

What the reply has to show

Rejecting the "by definition" argument does not make a miracle easy to establish.  John Earman, a philosopher of science whose book on Hume's argument is titled Hume's Abject Failure, also wrote that the eighteenth-century defenders of miracles "knew that miracle claims could not be established without the help of very strong evidence."⁴⁶  Timothy and Lydia McGrew, defending the argument from miracles, state the condition precisely.  For a miracle to be more probable than not on the evidence, "the prior probability of the miracle must be greater than the prior probability that the miracle does not happen and that we also have the evidence in question."¹⁴  They also hold that "any real, nonzero prior improbability can be overcome by sufficient evidence."¹⁴  That sentence is about what probability allows, not about what evidence exists.  A low prior can be overcome only by evidence that is very much more likely if the resurrection happened than if it did not.

The McGrews then try to supply the numbers.  They estimate how much more likely the women's testimony, the disciples' testimony and Paul's conversion are if Jesus rose than if he did not, as factors of 10², 10³⁹ and 10³, "a Bayes factor of 10⁴⁴, a weight of evidence that would be sufficient to overcome a prior probability (or rather improbability) of 10⁻⁴⁰."  They state at once what this rests on:  "this conclusion is predicated on the assumption that in matters other than the explicit claims of miracles, the Gospels and the book of Acts are generally reliable ... This calculation tells us little about the evidence for the resurrection if those assumptions are false."¹⁴  Their largest factor comes from multiplying the testimony of thirteen disciples as if each were independent.  They grant that for the disciples "the independence assumption almost certainly breaks down" and that "the values we have supplied will certainly be contested," and they argue that "any dependence among the testimony of the witnesses actually favors our case."¹⁴

That is the strongest reply read for this article, and it moves the dispute to the right place.  The disagreement is not over whether a miracle can ever be probable.  It is over the size of the likelihoods:  how far the general reliability of the Gospels and Acts can be assumed, how independent the witnesses were, and how well grief visions, legend or other natural accounts explain the reports.  Those are historical questions, and the McGrews' answers to them are contested inputs, not measurements.

A second Christian reply is built to avoid the reliability premise.  Michael Licona's 2010 study admits as evidence only a "historical bedrock" of facts "so strongly evidenced that they enjoy a heterogeneous and nearly universal consensus."  He finds three:  Jesus died by crucifixion;  very soon afterward the disciples had experiences that led them to believe and proclaim that he had been raised and had appeared to them;  and within a few years Paul converted after an experience he took to be such an appearance.  "In fact," Licona notes, "Ehrman grants all three facts that belong to our relevant historical bedrock."⁴⁵  Ehrman does.  Of Paul's report of what the disciples claimed, he writes:  "I don’t think he is making it up. And he knew at least a couple of them, whom he met just three years after the event".¹⁷  Licona then weighs the resurrection against five rival accounts by five criteria of explanation, among them the hallucination accounts of Michael Goulder and Gerd Lüdemann.  He judges the resurrection "by far the best explanation of the historical bedrock," meeting all five criteria while "the strongest of the other five ... met only one."⁴⁵

This removes the McGrews' stated dependence on the general reliability of the Gospels and Acts.  It does not finish the argument.  Licona's criteria are qualitative, and how he scores them is his own judgment, with no numbers a skeptic can rerun.  The rival he scored strongest was Geza Vermes's view that historians cannot know whether Jesus rose, which is close to Ehrman's own position as a historian.  On three facts alone, the question narrows to a few visions against a resurrection, which the Hallucination Theories article leaves undetermined.  And the prior, which depends on whether God exists and might act, still has to be supplied before any weighing of explanations can give a probability.

The access argument is a rule for a discipline

Ehrman's access argument has a sound core.  The prior probability of a resurrection depends on whether God exists and what God would do, and historians as a group share no answer to that.  Licona, on the Christian side, concludes in the same discussion that "the probability of the resurrection of Jesus is inscrutable," meaning its probability before the evidence is weighed.⁴⁵

But the argument claims more than that core supports.  Most of the evidence, what the witnesses claimed, when, and whether visions or fraud explain it, is ordinary historical material, and assessing how well each hypothesis explains it does not require a view about God.  Historians infer the purposes of unseen agents from their effects all the time.  The likelihood of the evidence if Jesus rose does carry some assumptions about what a risen Jesus would do, and an Ehrman sympathiser can press them;  and they cannot be assigned numerical likelihoods simply by declaring that God could raise someone.  So "no access to God" bears chiefly on the prior, much less on the comparison of explanations.  And the 2014 rule that history "is based only on shared presuppositions" describes what historians of every belief can agree on.  As a description of consensus history it is fair.  It is a stipulation about the discipline, not a demonstrated limit on rational inference:  that people with different views of God reach different conclusions does not by itself make either conclusion unwarranted.  A historian can state a conditional result from specified priors and assumptions about the reports.  Whether the evidence favours a hypothesis depends on the likelihood comparison;  whether the resulting probability is high also depends on the prior.  Strict naturalism assigns a supernatural resurrection a prior of zero.  Ehrman himself draws the boundary:  historians "cannot prove or disprove whether God has performed a miracle in the past," and historians who are believers affirm miracles "not in the capacity of the historian, but in the capacity of the believer."⁴²

Assessment

1. Fails as stated: the claim that a miracle is "by definition" the least probable explanation. It conflates the probability before the evidence with the probability after, and it sets the prior by a frequency count that assumes no agent.
2. Right: a miracle claim needs very strong evidence, and the strength required depends on how likely the evidence is without the miracle.
3. Partly right: historians as a group cannot fix the prior, because it depends on whether God exists. The 2014 rule is a fair description of what can command agreement across worldviews, not a proof that no one may rationally conclude a miracle happened.
4. Open: whether the actual evidence is strong enough. The strongest Christian attempt to show it rests on stated assumptions that are themselves disputed, and neither Ehrman's rule nor Craig's formula settles the question alone.

Part 9: Two premises: the canon, and the end of the world

Who chose the books

Ehrman's account of the canon in Lost Christianities (2003) gives four criteria that "proto-orthodox" Christians used:  a book had to be ancient, apostolic, accepted widely across the churches, and orthodox.  He argues that orthodoxy controlled the others:  "To some extent, in fact, the other criteria were handmaidens to this one."⁴⁷  He rejects the popular story that Constantine chose the Bible:  "The historical reality is that the emperor Constantine had nothing to do with the formation of the canon of scripture: he did not choose which books to include or exclude, and he did not order the destruction of the Gospels that were left out of the canon (there were no imperial book burnings)."⁴⁸  The Constantine story is therefore not his, and refuting it does not touch his account.

The test of his account is a case where the reasons are recorded.  His own example is Serapion, bishop of Antioch about 200, who first let a church read the Gospel of Peter and later rejected it.  Ehrman reads the episode this way:  "How did he know that Peter had not written it? It was because the book contained something that looked like a docetic Christology, and obviously Peter could not have written such a thing."⁴⁷  Serapion's own letter, preserved by Eusebius, gives two reasons.  The first is tradition:  "For we, brethren, receive both Peter and the other apostles as Christ; but we reject intelligently the writings falsely ascribed to them, knowing that such were not handed down to us."  The second is content:  having borrowed a copy from those "whom we call Docetæ," he found "many things in accordance with the true doctrine of the Saviour, but some things added to that doctrine."¹⁵

So the case supports Ehrman in part:  doctrine triggered the inquiry and served as evidence that Peter was not the author.  It also shows a stated test of tradition that Ehrman's "It was because" leaves out, and a bishop who had allowed the book before examining it.  This episode alone cannot establish a general policy toward rejected books.  Ehrman's account handles other contrary cases by treating orthodoxy as necessary rather than sufficient:  the Shepherd of Hermas, an orthodox book, "could not pass muster in the Muratorian canon because it was, relatively speaking, a recent production."⁴⁷  That makes the account coherent, and it fits the cases.  Whether orthodoxy controlled the outcome, or was one of several tests that were applied together, is not something the surviving records can measure.

The series article The Biblical Canon agrees with much of this.  It finds the four Gospels and most of Paul received as Scripture by about 180-210 in Gaul and Carthage, before any council list, and finds no canon ruling at Nicaea.  Ancient readers could also separate the tests:  Dionysius of Alexandria accepted Revelation as inspired while denying that the apostle John wrote it.  Assessment:  largely fair;  his claim that orthodoxy controlled the other criteria is one reasonable reading of evidence in which the tests are entangled.

Did Jesus expect the end within a generation?

Ehrman's 1999 book presents Jesus as an apocalyptic prophet.  "Jesus appears to have thought that this coming judgment of God through the cosmic Son of Man is imminent. It is right around the corner. In fact, it is to happen within his own generation."  He cites Mark 9:1, "some standing here ... will in no way taste death until they see God's Kingdom come with power," and Mark 13:30, "this generation will not pass away until all these things happen."⁴²⁴²⁴  But his argument does not rest on any one saying:  "What matters for my purposes here is not whether Jesus spoke each and every one of these sayings exactly as they are recorded. What matters is that these themes are resounded repeatedly and independently in our earliest surviving sources."⁴  A line often quoted from the same book, "It turns out that Jesus was wrong," closes Ehrman's summary of Albert Schweitzer's view;  the next paragraph says no one now accepts Schweitzer's particular reconstruction.⁴  Ehrman's own inference is put more cautiously, as what people might claim "if Jesus was wrong."⁴

Dale Allison, a Christian historian who has written a book-length study of Jesus as a millenarian prophet, reaches the same conclusion:  "Jesus' generation, however, passed away. They all tasted death."  "Jesus the millenarian prophet, like all millenarian prophets, was wrong: reality has taken no notice of his imagination."  He then argues that Jesus nevertheless "says the only things worth saying".⁴⁹⁴⁹  Allison's agreement shows that the reading is not peculiar to skeptics;  it is not itself evidence for the reading.

The main alternative for Mark 13 is R. T. France's.  The chapter begins with Jesus predicting the temple's destruction, and the disciples ask "when will these things be?" (13:2-4).²⁴  France argues that verse 30's "all these things" answers that question, so the "generation" saying concerns the temple, which fell in AD 70;  that the coming of the Son of Man in verse 26 draws on Daniel 7, where one like a son of man comes "to the Ancient of Days" to be enthroned, not down to earth;  and that verse 32, "But of that day or that hour no one knows," turns to a different, undated event.²⁴²⁴  On his reading verses 24-27 describe "a change of government: the temple and all that it stood for is out, and the Son of Man is in."  France says openly that this reading removes "the single most embarrassing feature of chapter 13 for traditional Christian interpretation," the "very emphatic statement in v. 30," and that on it the events happened within the generation:  "They did!"⁵⁰⁵⁰  His candour about the relief his reading gives is not evidence that he adopted it for that reason;  his reasons are the textual ones just listed.

The reasons on the other side are also textual.  Verse 24 ties the cosmic signs closely to the crisis:  "in those days, after that oppression."  Verse 27 has the angels gather the elect "from the ends of the earth to the ends of the sky," which reads most naturally as a final gathering.  And Jesus tells the high priest that "You will see the Son of Man ... coming with the clouds" (14:62).²⁴²⁴  Mark 9:1 has its own complication:  Mark places the transfiguration directly after it ("After six days," 9:2), which allows a narrower reading of "see God's Kingdom come with power";  Ehrman reads the saying with 13:30 as a prediction of the final end.²⁴  Ehrman's wider point, that imminence appears across independent early sources, is supported by Paul's "we who are alive, who are left until the coming of the Lord" (1 Thessalonians 4:15), which supports an expectation of a coming within the lifetime of some contemporary believers, although an inclusive "we" alone does not prove that Paul predicted his own survival.²⁴

The series article Eschatology and Israel concludes that "this generation" either failed or referred to the temple, and that the text does not settle which.  Assessment:  Ehrman's reading is strong, because it rests on a theme attested across the earliest sources and not only on Mark 13.  France's alternative is serious for Mark 13 and weaker for the imminence theme as a whole.  If Ehrman is right, a prediction Jesus made failed.  That would tell against Jesus' authority as a prophet and against any claim that he could not err, and it would rightly affect how much weight a reader gives to his other claims.  It would not, by itself, decide the separate historical questions about the text, the authorship of letters or the resurrection.

Part 10: Do his popular books outrun his scholarly ones?

This comparison tests whether Ehrman's trade books claim more than his academic work.  It compares the same claim, in full context, across his trade books and his academic books.  Different audiences and different subjects are not contradictions, and a changed mind is not hypocrisy.  But the same main thesis can still be stated with more certainty, fewer caveats or a wider scope, so each pair is classified by how it differs.

The rows rest on the passages quoted in Parts 2 to 4 and 8, and on the literacy passages.¹⁸¹⁷³²  The literacy figures are models, not counts.  The study they rely on, William Harris's Ancient Literacy (1989), concludes that literacy in Italy was "likely to have been below 15%," and that it is unlikely that literacy in the western provinces "even rose into the range of 5—10%."⁵¹⁵¹  His "maybe 10 percent" for the empire is within those ranges.

The stronger formulations are elsewhere.  In the 2009 debate he opened with "there is no way to tell" whether we have a reliable text.  The next sentences narrow it to particular passages, so it is an overstated framing rather than a denial of his own "very closely," but it is stronger than the book passages compared here.¹⁹  The comparison of variants with words is more vivid than informative, but the title Forged accurately states the deliberate-deception claim made in both the trade book and the monograph.  Titles alone establish no difference between his popular and academic positions.

Assessment.  In the passages compared, the trade books do not contradict the academic work.  They are sometimes firmer in modality, as on the day of the crucifixion, and once broader than their own concession, as on secretaries.  This finding is limited to these passages.  It says nothing about how readers take the books, for which this article has no evidence;  the practical lesson is to quote his arguments rather than his titles.

Objections and responses

1. "Ehrman's conclusions follow from a materialist starting point." He denies it in print, and his stated reason for the miracle rule is the discipline's shared assumptions, not unbelief.² More to the point, an argument stands or falls on its premises. His probability argument fails for a stated reason (Part 8), and his other arguments succeed or fail on the manuscripts, the letters and the Gospels themselves. Guessing at motives adds nothing and invites the same move in reverse.
2. "His own text-critical work proves we have the original text, so his skepticism refutes itself." It supports substantial recovery, not self-refutation. His reconstruction of disputed places depends on the competing readings surviving, and he says the oldest text is "very closely" related to the original.¹ But "very closely" is not "exactly," some places stay disputed, and no method reaches behind the earliest copies. Nor is he inconsistent: he never claimed the text was unrecoverable, and Michael Bird's charge that he argues we "cannot realistically talk about recovering an original autograph" misstates his books.⁴² The fair conclusion is that the text is substantially recovered, not that every word is known.
3. "Every contradiction he cites has a harmonisation." Many do, and some harmonisations are good. A possible harmonisation is weaker than a probable one. For the day and hour of the crucifixion the best harmonisations require a less natural reading, and Licona, a defender of the Gospels, does not rely on them.¹⁰ Claiming that every difference is resolved overstates the evidence and is easy for a skeptic to check.
4. "John's date may be the right one, so Ehrman has the Gospels backwards." That is a serious possibility (Part 4), but it does not help a view that every detail is a precise report: if John is right about the day, the earlier Gospels are not. It helps against Ehrman's claim that John changed a known date for theology.
5. "Pseudonymity was an accepted convention, so 'forgery' is an anachronistic charge." The ancient sources examined in this series condemn deceptive authorship, as Ehrman says.³¹ The better reply is not that forgery was acceptable but that each disputed letter needs its own evidence, and that for several the evidence does not show forgery.
6. "Peter used a secretary, so his authorship of 1 Peter is safe." Not on the evidence available. "Through Silvanus" names the carrier, as Richards, who defends secretarial help in general, agrees,³⁵ and Richards's evidence does not establish full composition as an ordinary secretary's routine task. It does distinguish a trusted colleague from such a secretary. Help with the Greek is possible, as Josephus shows, but the dating arguments for 1 Peter, if correct, would exclude any help given in Peter's lifetime.
7. "If even one letter is forged, the Bible is discredited." This is a skeptic's inference, not Ehrman's argument. A forged letter would disprove the claim that every New Testament self-attribution is true. It would not disprove anything that rests on letters whose authorship is not in doubt.
8. "Ehrman admits the earliest Christians thought Jesus divine, so he concedes the Christian case." He admits the belief was early. He does not admit it was true, and early belief does not prove truth; sincere early beliefs can be mistaken. The early date removes one skeptical argument, that divinity was a late invention, and nothing more.
9. "Extraordinary claims need extraordinary evidence; this article lets the resurrection off lightly." The article agrees that a miracle claim needs very strong evidence, states the condition in probability terms, and reports that the strongest Christian attempt to meet it rests on disputed assumptions.¹⁴¹⁴ It concludes only that Ehrman's "by definition" argument does not settle the matter, and that the question is open.
10. "This article lets Ehrman off lightly on miracles: his rule is just methodological naturalism." His rule does exclude conclusions that require God, and the article says it is a stipulation, not a proof. But he applies it in both directions, and he says historians cannot disprove the resurrection either. The objection that matters is to the rule's scope, not to his honesty.
11. "The article underrates the decades of copying before the first manuscripts." The gap is real and is stated in Part 2: changes before the earliest copies cannot be measured either way. What the evidence does show is limited: the surviving copies come from several lines of copying that separated early, so an early change would have had to enter all of them to leave no trace. That is possible, and it cannot be ruled out.
12. "He keeps changing his mind, and his books contain errors such as the Jairus slip." He discloses his changes and gives reasons, which is what a historian should do. One wrong word on one page shows ordinary fallibility. Neither is evidence about his arguments.
13. "Most New Testament scholars disagree with him" or "most agree with him." Head counts are not arguments, and this article has no survey. Each claim has to be judged on its evidence.
14. "Giving up inerrancy means giving up Christianity." Ehrman's own story can suggest that, but his stated reason for leaving the faith was suffering, not the text, and he says historical criticism "does not necessarily lead" to agnosticism.¹⁷¹⁷ The arguments tested here bear hardest on particular doctrines about the Bible. The central historical claims stand or fall on separate evidence, and two of Ehrman's arguments that bear on them, the burial and the apocalyptic sayings, remain open.

Issues and unanswered questions

1. Debate wording is from transcripts. The White debate transcript was produced by White's ministry, and the Craig transcript is a joint copyright of both speakers. Neither was checked against audio. The article uses them only for points his books also make or for clearly labelled oral formulations.
2. Sources not obtained. The automatic captions of Ehrman's 2019 debate with Peter Williams, a recorded dialogue with Michael Licona whose link in older notes is dead, and his debates with Daniel Wallace. Karen Jobes's full 2005 commentary on 1 Peter and Richards's 2004 book on Paul's letter-writing remain unread. Jobes's published 2003 syntax study was obtained, and the publisher's 2022 excerpt confirms she retains that analysis; the full revised excursus is outside the excerpt. Darrell Hannah's Michael and Christ, which Tilling cites against the angel reading, was found only as an unauthorised upload and was not used. Of the reply volume How God Became Jesus, Simon Gathercole's chapter was read only for its "I have come" argument, and Charles Hill's not at all. Sayings attributed to Ehrman from these and other sources are kept out and listed in the quarantine record.
3. Second-hand pieces. The rest of Ehrman's 2014 blog post on the Synoptic Gospels is behind a membership wall. The Empty Tomb, authorship, early-belief and canon verdicts are those articles' findings, reported here and not re-argued.
4. Not assessed. 2 Peter, James and Jude; the six disputed Pauline letters case by case; Ehrman's books on suffering, heaven and hell, and the Apocalypse.
5. Selection. The examples were chosen because they are Ehrman's strongest or the best test of his central inference. They show how his arguments work and where they stop. They do not measure how often the New Testament is right or wrong, and no rate is offered.
6. An open question for the text. Whether any independent evidence, such as early quotations in second-century writers, can test the first decades of copying more directly than the surviving manuscripts can. The strongest manuscript evidence for an early, stable text, P75's agreement with Vaticanus, depends on a date for P75 that Nongbri has reopened.
7. An open question for miracles. No published analysis read for this article puts numbers on the likelihoods of the resurrection evidence without assuming the general reliability of the Gospels and Acts. Licona's 2010 method avoids that assumption by working from three facts Ehrman grants, but it weighs explanations qualitatively and assigns no numbers. A numerical analysis built on those three facts, with the vision hypothesis as the main rival, would move the question forward.

Verdict

On the nine arguments tested here, Ehrman's factual premises mostly hold in the form he states them in his considered work.  The copies of the New Testament changed, sometimes for doctrinal reasons;  ancient writers condemned forgery;  1 Peter presents substantial questions about education and composition, with a contrary syntax study to weigh;  some differences between the Gospels are unresolved;  Christian language about Jesus developed;  and historians of every belief cannot agree on a prior probability for an act of God.  On those points a Christian apologist should concede rather than contest.

Several of the conclusions he builds on those premises go further than the evidence tested here.  His argument that God did not inspire the text rests on premises about what God would do.  "Forged" is strong for some letters and unsettled for others.  That John changed a known date is one of three live explanations.  The angel reading of Paul is possible, not established.  His "by definition" argument against miracles fails, and his rule about historical method decides nothing about whether the resurrection happened.  Other questions he raises stay open after the replies:  how Jesus was buried, whether Jesus expected the end within his generation, what Jesus claimed for himself, and whether the evidence for the resurrection is strong enough.  In the passages compared, his trade books repeat his scholarship with firmer wording rather than contradicting it.

What Ehrman's work costs Christian belief is real and specific.  A believer claiming certain knowledge of every original reading must address actual textual uncertainty.  New Testament self-attributions and precise Gospel chronology face separate unresolved challenges;  uncertainty about either is not itself a demonstrated falsehood.  A believer whose case rests on an approximately recovered text, the seven letters of Paul that Ehrman accepts, and the early and widespread claim that Jesus was raised keeps those premises against these arguments.  Whether those premises are enough is a further question, which the open issues above bear on directly.

Sources