Hello brothers and sisters,
For two weeks now we’ve lived inside one paragraph. About ninety Greek words, tucked into the eighteenth book of Josephus’s Antiquities.
The first week we went hunting for a missing word and found that Josephus most likely wrote that Jesus “was called the Christ.” Last week we followed παράδοξος (paradoxos), the “startling deeds,” from the plagues of Egypt to a packed house where a paralyzed man came down through the roof. If you missed either one, I recommend getting caught up before you proceed.
This week there’s no new word. No new line of Josephus. Instead, I want to open up the toolboxes and show you how the scholars actually reach their verdicts, and where each toolbox breaks.
Because you may have noticed something that ought to bother a both/and guy like me. For two weeks I’ve been telling you that one reading is probably original and another probably isn’t. That is not how I talk about the Hebrew and the Greek Scriptures. So before we close this series, I owe you an explanation.
We’ll get there. But first, the prosecution.
Let’s dig in.
If you’re reading this in email, be aware that the text is likely to cut off without warning. For a smoother reading experience and all the features Substack has to offer (including audio voiceovers of my posts), you can go HERE or download the app.
The Case for Forgery
Ken Olson’s mature argument doesn’t open with Josephus at all. It opens on a battlefield.
In his Life of Constantine, written shortly after the emperor died in 337, Eusebius reports a speech the pagan emperor Licinius supposedly gave his troops before Constantine crushed him. Licinius tells his men that the battle will settle whose gods are real. If Constantine’s strange God wins, everyone should worship Him. Eusebius assures us he heard all this from men who were standing there (Life of Constantine 2.5.5).
Modern historians aren’t buying it. Averil Cameron and Stuart Hall, who translated the Life, put it plainly: “more probably it is his own invention.”
The Greeks had a word for writing a speech in someone else’s voice. προσωποποιία (prosōpopoiia), “face-making.” Ancient historians did it constantly, and Eusebius was no exception. In his Demonstration of the Gospel he writes a whole mocking speech in the voice of Jesus’s disciples, as if they really were the frauds his opponents said they were.
So Olson asks the obvious question. What if the Testimonium is one more face Eusebius made? His answer: “The most likely hypothesis is that Eusebius either composed the entire text or rewrote it so thoroughly that it is now impossible to recover a Josephan original.”
He brings receipts. “Doer of startling deeds” turns up nowhere in surviving Greek before Eusebius, who uses it ten more times outside the paragraph. (Schmidt counts nine. The difference is one line in Eusebius’s commentary on the Psalms.) “And ten thousand other marvels” (καὶ ἄλλα μυρία) shows up in Eusebius eight more times. “The tribe of the Christians” twice more.
And the content fits Eusebius’s argument like a key in a lock. His opponents, pagan and Jewish, said Jesus was a γόης (goēs), a sorcerer, and His disciples were frauds. Eusebius’s central answer in that book: if they were frauds, why didn’t they abandon Him after the cross? The paragraph in Josephus says exactly that. “Those who had loved him at the first did not cease.”
But Olson’s sharpest point isn’t a word count. It’s about method.
Arguments about negative tone and ironic or ambiguous readings are almost entirely subjective. Our ability to perceive them depends on who we think wrote the text in the first place.
And he sees Schmidt’s timeline argument coming before Schmidt ever made it. Suppose Eusebius soaked up Josephus’s phrases, and even its apparent Christology, so deeply that he echoed them for decades. That, Olson writes, “seems not only improbable but comes near to removing the hypothesis of authenticity from any possibility of falsification.”
In plain English: an explanation that can absorb every piece of evidence can’t be tested by any of it.
That’s a serious charge. Hold on to that word, falsification. It’s coming back.
Toolbox One: The Fingerprint
T. C. Schmidt answers with a magnifying glass.
His method is style. Every phrase of the paragraph gets laid beside the nearly half million words Josephus left behind, using the big searchable databases that earlier generations of scholars could only dream about. Does Josephus build sentences this way? Does he pair these words? Does he use these little particles, the ones a writer sprinkles in without thinking, the way you and I say “well,” “you see,” or “anyway”?
Some of it is wonderfully precise. Schmidt finds that Josephus uses a word he never uses anywhere else about once every eighty-seven words. The paragraph is about ninety words long. So if Josephus wrote it, you’d expect it to contain about one word found nowhere else in his work. It contains exactly one: “Christians.” And what other word would anyone use?
He looks at placement too. The paragraph sits in a run of five stories about uproar and scandal under Pilate: riots, a seduction in a temple of Isis, a swindle. Would a Christian forger really slip his Lord into a lineup like that?
And you already know his answer to Olson’s word counts from last week. Eusebius only starts using “doer of startling deeds” on his own, apart from his quotations of Josephus, twenty years and more after he first quoted it. Schmidt reads that as a man who absorbed a phrase, not one who invented it.
That’s a lot of evidence. But notice what kind of evidence it is.
Schmidt asks, “Does this sound like Josephus?” Olson asks, “Does this sound like Eusebius?” They’re holding the same ninety words up to two different lights. And each one thinks the other is seeing what he expected to see.
Toolbox Two: The Family Tree
Jesús Fernández-Pedrera Correa thinks both of those lights are the wrong instrument.
He names the problem with style arguments bluntly.
First, the sample is tiny. Measures of an author’s habits need thousands of words to mean much, and here we have about ninety.
Correa’s paper actually says a hundred and twenty-seven, but that’s the length of Whiston’s English. The Greek is shorter, which only sharpens his point.
Second, the reasoning tends to run in a circle:
The elements identified as un-Josephan are, with striking regularity, precisely the elements a Christian reader finds theologically loaded.
Read that next to Olson’s complaint about “subjective” readings of tone. Olson says the authenticity camp hears irony because it expects Josephus. Correa says the forgery camp hears Christian theology because it expects a forger. Each side accuses the other of finding what it came looking for. And both of them have a point.
So Correa moves style to the back of the line and reaches for the oldest tool in textual criticism: the family tree of the manuscripts.
Here’s his key insight, and I think it’s the most useful sentence anyone has written on this paragraph in years. All our Greek manuscripts agree. People usually treat that as a point for authenticity. Against the idea that the whole paragraph was invented, Correa agrees it is. But for any single line in it, he says, that agreement proves nothing:
A tradition that is uniform is evidentially poor at the point of its uniformity; a tradition that has fractured into languages and ecclesiastical worlds that cannot correct one another is evidentially rich at the points where the fractures show.
Think of it like a family resemblance. If every cousin at the reunion has the same nose, that tells you the nose goes back to a common grandparent. It doesn’t tell you what combination of genetics gave it to Grandpa. Every Greek copy of the Antiquities descends from the same lost ancestor. They can tell us what that ancestor said. They can’t tell us how it came to say it.
The translations are different. Latin, Arabic, and Syriac grew up in churches that couldn’t read each other’s books. Where they split from the Greek, they might be telling us something the Greek family has forgotten.
And Correa runs a test that I love, because it’s the kind of thing good science does. He brings in a negative control: the Old Slavonic version of Josephus’s Jewish War, which contains long passages about Jesus and John the Baptist found in no Greek copy. Nobody takes them seriously anymore. They’re medieval. And Correa points out what they look like: expansion, extra detail, more Christian content. That’s what Christian scribes do when they tinker.
At the messianic line, Jerome, Agapius, and Michael do the opposite. They pull back. “Believed to be.” “Perhaps.” “Thought to be.”
Whatever produced that, it wasn’t the ordinary habit of Christian copyists.
Cousins, Not Strangers
Then Alice Whealey walks in and rearranges the family tree.
In a 2008 article in New Testament Studies, she showed that Agapius’s Arabic and Michael’s Syriac both descend from the same source: the Syriac translation of Eusebius’s Church History. So two of Correa’s three witnesses aren’t strangers who happen to agree. They’re cousins, branches off the same Eusebian trunk. (Correa admits as much up front.) Whealey also shows that some of the “less Christian” touches in Agapius, the ones that excited scholars for decades, come from the Syriac source he was paraphrasing, not from Josephus.
Score one for caution.
But then Whealey turns up what I think is the most important fact in this whole debate.
Michael’s Syriac says Jesus “was thought to be the Messiah.” Jerome’s Latin says he “was believed to be the Christ,” credebatur esse Christus. In that era Latin and Syriac writers didn’t read each other’s books, except through Greek translations. Whealey’s conclusion: “it is scarcely credible that the writers could have independently modified the Testimonium in this same way.” So both must go back to a Greek text that read “something like ‘he was believed to be the Christ.’”
Now look at the dates. The earliest manuscripts of Jerome’s book come from the sixth or seventh century. Whealey points out that this makes them “several centuries older than the earliest Greek manuscripts” of either Josephus’s Antiquities or Eusebius’s Church History. And Jerome most likely took the paragraph from Eusebius.
Whealey even suggests the Greek word that once stood there: ἐνομίζετο (enomizeto), “he was supposed,” the very word Luke uses for Joseph being “supposed” to be Jesus’s father. That would explain Michael. It would explain Jerome. It would explain why Origen, who read this book of Josephus a century and a half before Jerome, said Josephus didn’t believe Jesus was the Christ. And it would explain why a Latin adaptation of Josephus from around 370 never mentions the line at all, even while it plays up the rest of the paragraph.
See what just happened? The family-tree method, corrected by Whealey, gives us fewer independent witnesses than Correa counted. But it gives us a sharper result. At the one line that matters most, two branches that never spoke to each other agree with each other against every surviving Greek manuscript.
And that creates a real problem for Olson. In his reading, the paragraph is built to arrive at a conclusion: “that, in fact, this one was the Christ.” But if the copy behind Jerome, most likely Eusebius’s own Church History, said “believed to be the Christ,” then the softer reading is at least as old as Eusebius’s own text. The line Olson treats as the punch line may not be what Eusebius wrote.
Remember Olson’s word, falsification? This is what a real test looks like. Nobody’s ear for tone is involved, mine included. Two translations from two churches either agree or they don’t. Here they agree, and their agreement could have pointed toward Olson.
But it points away from him.
The Friendly Fire
One more voice, because the sharpest critique of Schmidt doesn’t come from the forgery camp at all.
Steve Mason is one of the most respected Josephus scholars alive. Schmidt cites him more than once. Early this year Mason published a long review of Schmidt’s book in the Irish Theological Quarterly, and on the question we’ve cared about for three weeks, his verdict is generous: “His case for authenticity in Part 1 seems to me strong, and something that others will need to deal with.”
Then he takes the rest of the book apart.
Schmidt’s boldest claim is that Josephus knew some of the men who accused Jesus, most likely the high priest Ananus II, son of the Annas we meet in the Gospels. The key evidence is two little words in the paragraph: “the leading men among us,” παρ᾽ ἡμῖν (par’ hēmin). Schmidt reads that as Josephus vouching for people he knew personally.
Mason says no. “Among us” is simply how Josephus talks to Greek and Roman readers. It means we Judaeans, over against you. A third of the times he uses it are in Against Apion, his defense of “our” laws against Greek critics. And the timeline doesn’t cooperate either. Josephus had about six months of relative freedom in Galilee before the Romans captured him, not the years Schmidt imagines, and Ananus II was killed in the winter of 68, not 69.
Then Mason lands his deepest point. Even if Josephus wrote every word, we still can’t see behind it to his sources. Josephus rewrites everything in his own voice. “We cannot reconstitute it,” Mason says, “any more than we can the eggs that go into a cake.”
He’d push back on me too, by the way. Mason thinks Schmidt hears more reserve in the paragraph than Josephus put there. In his view, “the picture of Jesus as wonder-worker and truth-teller hardly suggests negativity.” His own guess is that Josephus, the Judaean expert in Rome, is simply explaining where those strange Christians came from. In Mason’s paraphrase: “That is who your Christiani are. They do not eat babies or hold orgies.” And he adds, honestly, “I would not go to the wall for this proposal.”
Notice the shape of that. Mason agrees with Schmidt on the core question and disagrees hard about what can be built on it. He closes by thanking Schmidt for the clarity of his case. That’s what good scholarship looks like. Nobody wins by shouting. Everybody gets sharper.
Why I Don’t Read Josephus Both/And
Now, the explanation I promised.
For three weeks I’ve weighed readings of Josephus and picked winners. “Probably original.” “Probably altered.” If you’ve been reading me for a while, you know that isn’t how I talk about the Hebrew and the Greek Scriptures. When the Masoretic Text and the Septuagint differ, I go looking for how both can be true, because I believe that in most cases they both can be.
So why the difference?
Because Scripture is God-breathed. The Holy Spirit inspired it, and I believe God deliberately preserved it in both the Hebrew and the Greek. So in any case where the two can be reconciled, I reconcile them.
Josephus spent a good part of his career paraphrasing Scripture, and I have tremendous respect for that. But his work is not, in itself, Scripture. It deserves exactly the scrutiny we’d give any other historical document.
That isn’t a double standard. It’s one standard applied to two different kinds of text. My both/and is a conviction about what God has done with His Word. It was never a claim that no copyist anywhere ever slipped.
In fact, I’ve argued the opposite on a number of occasions, even in Scripture. Because, you see, Scripture itself shows us the fault lines.
Here’s one quick question to illustrate the point: who killed Goliath?
David, of course (1 Samuel 17). We know that. But read this:
And there was a battle in Gob with the foreigners and Elhanan son of Jaare-oregim the Bethlehemite struck down Goliath the Gittite and the wood of his spear was like a weaver’s beam. (2 Kingdoms 21:19, LES; 2 Samuel 21:19 in English Bibles)
And now Chronicles, telling the same story:
There was again war with the foreigners, and Elhanan the son of Jaur struck Lahmi, brother of Goliath the Gittite, and the wood of his spear was like the crossbeam of a weaver. (1 Chronicles 20:5, LES)
Here’s what matters. This isn’t a Hebrew-versus-Greek problem at all. The Hebrew and the Greek of Samuel agree with each other: Goliath. The Hebrew and the Greek of Chronicles agree with each other: Lahmi, Goliath’s brother. Both streams God preserved carry both books faithfully.
Look at Samuel’s Hebrew. Elhanan is בֶּן יַעְרֵי אֹרְגִים (ben Ya’arei Oregim), “son of Jaare-oregim.” Then the verse ends with Goliath’s spear like a מְנוֹר אֹרְגִים (menor oregim), a “weavers’ beam.” Oregim, “weavers,” twice in one verse, once as part of a name nobody else in the Bible carries. Now set the middle of the verse beside Chronicles:
Samuel: בֵּית הַלַּחְמִי אֵת גָּלְיָת (beit hallachmi et Golyat), “the Bethlehemite ... Goliath”
Chronicles: אֶת לַחְמִי אֲחִי גָּלְיָת (et Lachmi achi Golyat), “Lahmi, the brother of Goliath”
Nearly the same letters. The likeliest story is that a copyist’s eye slipped: “weavers” crept up from the end of the line into the father’s name, and “Lahmi, the brother of” blurred into “the Bethlehemite.” And since the Greek of Samuel carries the same reading, the slip happened before the Greek translators ever unrolled that scroll.
That’s what reconciling looks like for me. I don’t pretend Samuel’s copyists never slipped. I let Scripture interpret Scripture, and 1 Samuel and Chronicles hand us the missing piece. David killed Goliath. Elhanan killed his brother. God didn’t leave the question hanging. He preserved the answer, in both languages, in the other book.
Now watch what Josephus does with the same verse. When he retells it in his Antiquities, he quietly drops the problem: “Nephan his kinsman fought in a single combat with the stoutest of all the Philistines, and slew him” (Antiquities 7.302, Whiston). No Goliath at all. Just “the stoutest.”
That’s a historian smoothing his source, which is exactly what historians do. It’s also exactly why his books aren’t Scripture. And it’s why, when we come to his paragraph about Jesus, we don’t have a Chronicles to check him against. There’s no second stream God promised to keep. There’s one family of Greek manuscripts descending from a single lost ancestor and a handful of translations, plus a church historian who may or may not have had a heavy hand.
So we do with Josephus what you’d do with any ancient witness. We weigh him.
What We Can Actually Know
So where does all that weighing leave us?
Here’s the surprise. For all their fighting, the toolboxes mostly point the same way.
Josephus wrote something about Jesus here. Mason calls Schmidt’s case for that strong. Correa defends a core notice. And two books later, when Josephus mentions “the brother of Jesus who was called Christ” (Antiquities 20.200), he talks like a man who has already introduced that Jesus.
Olson, with a few allies like the late Louis Feldman, still gives the whole paragraph to Eusebius. But as Mason notes, fewer scholars today than in the past think it was invented wholesale.
The flat “This was the Christ” is the line most likely to have been touched. Schmidt, Correa, and Whealey come at it with different tools and land in the same place: Josephus wrote something softer, “called” or “believed to be.” Correa gives that conclusion his highest grade of confidence.
Past that, honest people disagree. Correa rates the resurrection line only a moderate case. He thinks “if indeed one ought to call him a man” can’t be settled from outside evidence at all. And whether Josephus ever met anyone who stood at Jesus’s trial, Mason says, is something we simply can’t know.
Three grades of confidence. Fairly sure, probably, and we can’t say. That’s what honesty sounds like with a human document.
And I’ll tell you, I find that freeing.
My faith doesn’t rest on Josephus. It never did. If every scholar in this post woke up tomorrow convinced that Eusebius invented the whole paragraph, the tomb would still be empty. Peter’s “You are the Christ” would still stand. The Gospels would still be what they are. Josephus is a witness we’re glad to have on the stand. But Scripture isn’t a witness in the case. It’s the Word.
That’s exactly why we can afford to be honest about Josephus. We don’t need him to say more than he said.
And I think the scholars in this post model something worth taking home. Olson made the best case he could for a position most of his colleagues reject. Schmidt took it seriously enough to call it the strongest argument against him. Correa named the blind spot in style arguments, then built a control to test his own method. Whealey corrected the scholar whose discovery she was building on. Mason praised the book he was dismantling.
Mason puts the standard well. The test of new work in history, he writes, “is not that it settles a problem once and for all. The test is that [it] investigates a problem methodically, clarifying options and supporting evidence.”
That’s not a bad test for us either. Steel-man before you swing. Know which tool you’re holding and where it breaks. Match your confidence to your evidence. And when a headline promises you a two-thousand-year-old manuscript, ask to see the manuscript.
Thanks for walking through this one paragraph with me for three weeks. It’s a small window. But it looks out on the most important life ever lived.
Questions for Reflection
Olson argues that how we hear a text’s tone depends on who we already think wrote it. Where might that same trap catch you when you read Scripture, a news story, or someone you disagree with?
Most of us inherited our Bible from a translation committee and never think about the manuscripts behind it. Would knowing more about how God preserved His Word strengthen your trust in it, or unsettle it? And why?
Mason praised the very book he was dismantling. What would it look like for your church to disagree that way, on doctrine, on politics, on anything?
If Josephus disappeared from history tomorrow, what in your faith would actually change?
Sources
Ken Olson, “A Eusebian Reading of the Testimonium Flavianum”, in Eusebius of Caesarea: Tradition and Innovations, ed. Aaron Johnson and Jeremy Schott (Washington, DC: Center for Hellenic Studies, 2013), 97-114. Earlier form: “Eusebius and the Testimonium Flavianum,” Catholic Biblical Quarterly 61 (1999): 305-322.
T. C. Schmidt, Josephus and Jesus: New Evidence for the One Called Christ (Oxford: Oxford University Press, 2025; online edn, Oxford Academic, May 5, 2025), doi.org/10.1093/9780191957697.001.0001. Free to download. Pages drawn on here: 105, 127-129, 225-227.
Jesús Fernández-Pedrera Correa, “The Testimonium Flavianum in Four Transmission Lines: Greek, Latin, Arabic, and Syriac,” Zenodo preprint, August 15, 2026, doi.org/10.5281/zenodo.21945637. Not yet peer reviewed.
Alice Whealey, “The Testimonium Flavianum in Syriac and Arabic,” New Testament Studies 54 (2008): 573-590, doi.org/10.1017/S0028688508000301.
Steve Mason, “The Historical Jesus: New Evidence or New Techniques? Article Review of Schmidt T. C, Josephus on Jesus: New Evidence for the One Called Christ,” Irish Theological Quarterly 91.1 (2026): 74-98, doi.org/10.1177/00211400251411327.
Averil Cameron and Stuart G. Hall, Eusebius: Life of Constantine (Oxford: Clarendon, 1999), as quoted in Olson.
Flavius Josephus, Jewish Antiquities 7.302; 18.63-64; 20.200. English quotations are from William Whiston’s 1737 translation (public domain).
Septuagint quotations marked LES are from Rick Brannan, Ken M. Penner, Michael Aubrey, Israel Loken, and Isaiah Hoogendyk, The Lexham English Septuagint, Second Edition (Bellingham, WA: Lexham Press, 2020). Hebrew: Westminster Leningrad Codex. Greek Old Testament: Rahlfs-Hanhart.
The LXX Scrolls is free to read and always will be. If this work has been worth something to you, there are a few ways to say so:
Buy the ebooks. Completed teaching series are available as polished ebooks under the Two Witnesses, One Truth series. Buying through Curios will support this work most directly but they’re also available on Amazon (and elsewhere) if you’re loyal to a particular ereader.
Become a supporter. A monthly or annual pledge through Substack helps me to bring the Septuagint to those who never knew they needed it.
Send a one-time tip. If this post has blessed you and you want to express that directly, you can Buy Me a Coffee.
Thank you for being part of this journey, your support makes this work possible
© 2026 LXX Scrolls. All rights reserved.





