In Part One of this series, “What Clement Actually Says,” I ended with a puzzle. A letter from the Roman church to Corinth, written around AD 95, was read aloud in Corinthian worship, treated as Scripture in church after church, and cited for centuries as an authority — and today it is not in the Bible. How does a book fall out of the Bible?

The obvious suspect is a bonfire. The early church, and the empire that eventually adopted it, could be ruthless with books. So the natural guess is that somebody condemned Clement — a council, a bishop, an emperor — and the letter was suppressed. I went looking for that condemnation. It does not exist — no condemnation, no burning order, no anathema, no index of forbidden books. Clement was dropped, not condemned — and the difference turns out to be the real story of how the New Testament got its table of contents.

Was it ever in?

One thing to be clear about before anything else: this isn’t a letter, and Clement didn’t write it. What we call Second Clement is a sermon — a homily, meant to be read aloud to a congregation — written by an unknown Christian in the middle of the second century, decades after Clement of Rome was dead. It travels under his name because later generations filed it there; the text itself never claims him. Whether that even counts as forgery is a fair question. Nobody signed his name to it. They just shelved it with his work, and the label stuck.

Start with how high the letter stood, because “fell out” only means something if it had somewhere to fall from.

Around AD 170 — some seventy-five years after the letter was written — Dionysius, bishop of Corinth, wrote to Rome and mentioned that Clement’s letter had been read aloud in his church that week, as it had been since it arrived, and would go on being read. Eusebius says it was read publicly “in very many churches” (HE 3.16). Clement of Alexandria, around AD 200, quotes it constantly and calls its author “the apostle Clement.” In the fifth century, Codex Alexandrinus bound 1 and 2 Clement right after Revelation, between the same covers as Matthew and John. The Apostolic Canons, a late-fourth-century church rulebook, list “the two epistles of Clement” among the “venerable and holy” books (canon 85).

No council ever formally canonized it — but no council formally canonized Matthew either; that is not how it worked yet. For three hundred years the letter functioned as Scripture: read in worship, cited as authority, copied into Bibles. The edges of the big Bibles were soft — Codex Sinaiticus tucks Barnabas and the Shepherd in after Revelation the same way. The shelf existed; its contents were still settling.

The church knew how to ban a book

In AD 303, Diocletian ordered Christian scriptures burned — the church was the victim of book-burning before it was ever a perpetrator. A generation later Constantine ordered the works of Arius burned and made concealing one a capital crime. In 398, the books of the Eunomians were burned “under the eyes of the judges.” In 435 the Theodosian Code ordered Nestorius’s writings burned and his followers renamed “Simonians” (16.5.66). When the ancient world wanted a book dead, it said so — by name, in law, with penalties attached.

Clement’s name appears in none of it. The machinery of suppression existed, it was used repeatedly, and it was never aimed at this letter. Whatever happened to Clement, it was not that.

The snub of silence

What happened instead shows up in the lists.

Around AD 325, Eusebius of Caesarea sorted the Christian books into four tiers: the recognized; the disputed (James, Jude, 2 Peter, 2 and 3 John); the spurious (the Acts of Paul, the Shepherd, the Apocalypse of Peter, Barnabas, the Didache); and the outright heretical forgeries (the Gospels of Peter, of Thomas, and their kin). He had a slot for everything — the honored, the arguable, the second-rate, and the damned. First Clement is in none of them. And this is the same Eusebius who elsewhere speaks warmly of the letter, and a few chapters later vouches for it as genuine and widely read (3.38), while shrugging at 2 Clement. He did not reject Clement. He just did not file him. It is the snub of silence, and you start seeing it everywhere in this process.

Then, in AD 367, Athanasius of Alexandria sent out his annual Festal Letter — the thirty-ninth — with the first surviving list of exactly our twenty-seven books: these are the canon, he wrote; let no one add to them or subtract from them. He kept a second shelf — Wisdom, Sirach, the Didache, the Shepherd — for books appointed to be read to new converts, and a third category for the apocrypha, “an invention of heretics.” Clement is in none of the three. Not canonized, not second-shelf, not condemned. Simply nowhere. David Brakke, Athanasius’s editor and translator, reads the letter as episcopal boundary-making — a bishop drawing a line that also fenced in his own authority. Whatever the motive, by 367 the most influential list ever drawn proceeded as if the letter did not exist.

It lingered on the edges — Alexandrinus and canon 85 are later than 367, and from elsewhere — but lists converged slowly, region by region, and the direction was set.

Runner-up, not defendant

So why did it fail? Not on content. Nobody ever found Clement’s theology wanting; his theology was the emerging orthodoxy’s own — order, succession, repentance, resurrection. Part One’s verdict was that Clement is orthodox in instinct and pre-canonical in tools, and the centuries acted only on the second half of that sentence.

It failed on the two tests that governed admission. The first was apostolicity: a book had to come from an apostle, or from next door to one. Clement’s letter is anonymous and corporate — it speaks for “the church of God which sojourneth in Rome” — and its author belonged to the generation after the apostles, whose deaths the letter itself records. The later tradition that he had personally known Peter and Paul was, as I argued in Part One, embroidery — and embroidery is not documentation. The second test was catholicity: a book had to be used by the whole church, everywhere. Clement’s letter was used widely, very widely. Wide is not everywhere.

The Shepherd of Hermas and the Epistle of Barnabas died the same death, of the same non-crime. Clement was a runner-up, not a defendant. There was no trial, because there was never a charge.

How does a process like that actually run? Lee Martin McDonald, who has spent a career on the formation of the canon, describes it in centuries, not sessions: books rose or settled by use in the churches, long before any list ratified the outcome. Harry Gamble makes the same point from the book-trade side — collections circulated, wore out, got recopied or didn’t. Bart Ehrman’s Lost Christianities tells the broader version: the New Testament is the portion of early Christian literature that won. Was politics involved? Of course — but not the committee kind. No council ever voted Clement off the island. The pressure came from a hundred local churches slowly converging on the same shelf, with an occasional shove from the side. In the mid-second century Marcion published a canon — one Gospel, ten letters of Paul — and an old theory (Harnack’s) made him the canon’s inventor, who forced the church to build a counter-canon. Scholars have since demoted him to catalyst, not creator: Gamble notes Paul’s letters were already circulating as a collection before Marcion, and Metzger’s textbook puts it in one sentence — Marcion “did not create the idea of a canon, but accelerated a process already begun.”

Fairness requires giving the other side its hearing, and it has a serious spokesman. Michael Kruger of Reformed Theological Seminary argues in Canon Revisited that the canon is essentially self-authenticating — the church did not create the list so much as recognize books that carried their authority with them. On Kruger’s telling, Clement’s fade is not politics at all; it is the criteria working, and the letter lacked what the twenty-seven have. The two stories agree on every fact and differ only on the explanation — and on either story, Clement’s exit was quiet. Nobody in the ancient world seems to have noticed a decision about Clement being made. That is what attrition looks like from the inside.

Meanwhile, the actually condemned books survived

Here is the irony the whole business turns on. The books that were named, condemned, and ordered burned — a good many of them survived anyway. By being buried.

In December 1945, near Nag Hammadi in Egypt, a sealed jar came out of the ground holding thirteen leather-bound codices, fifty-two treatises — among them, in Codex II, the only complete copy of the Gospel of Thomas. Thomas had been a condemned name for sixteen centuries; the fathers quoted it mainly to refute it. The find made Elaine Pagels’s The Gnostic Gospels possible, and Karen King has spent a career on what these books do to the tidy picture of one orthodox stream and a few heretical puddles. The codices were copied near the Pachomian monasteries, and James M. Robinson, the find’s great editor, proposed that monks buried their library in response to Athanasius’s letter of 367 — the same letter that completed Clement’s quiet exit. It is the leading hypothesis. It is not a settled fact: Nicola Denzey Lewis points out that we know of no other case of monks hiding books from their own church, and the codices were produced over decades, not in one panicked season. The honest version: probably a monastic library, buried in the fourth century for reasons we cannot prove, with Athanasius’s letter as the leading suspect. Even the suspicion completes the picture: in the very years Clement was being silently dropped, somebody in Egypt may have been digging a hole for the loudly condemned — and the condemned are the ones we got back whole.

Egypt did it twice. In the winter of 1886–87, a French archaeological mission opened a Christian monk’s grave at Akhmim and found a parchment codex buried with him: the Gospel of Peter, the Apocalypse of Peter, and a Greek copy of 1 Enoch — three texts condemned for well over a thousand years, treasured by some monk enough to take them into eternity with him. Survival by burial, not by blessing.

The empire could burn what it could find. It could not find Egypt.

The verdict

So Clement’s letter was never executed. It suffered something quieter and, for a book, sadder: it was gradually not copied. A canon forms the way a shoreline does — mostly by slow water, not by explosion. The books that made it were read, and read everywhere, and read as apostolic, until Athanasius could write down in 367 what the churches had already decided without ever quite deciding it. Clement missed on paperwork: no apostle’s name on a title page that wasn’t even there, a home in many churches but not in all. Its theology was fine. Its author may well have run the church in Rome. None of that was the test.

I am an atheist, so I don’t believe the twenty-seven were handed down from heaven either. But the Clement case is the cleanest window I know into what the canon actually is: not a library assembled by conspiracy, not a thunderbolt from Sinai, but a human institution’s long memory, accidents of survival included — which books a monk in Egypt loved enough to bury, and which letter a thousand copyists, one by one, stopped recopying, because the list on the bishop’s desk no longer had a line for it.

Nobody threw Clement out of the Bible. The door just closed, quietly, and he was on the other side of it.

Read more

• The letters themselves, in J. B. Lightfoot’s translation: 1 Clement · 2 Clement (Early Christian Writings)
• For how the early church itself used and judged these letters: Eusebius, Church History, Books 3 and 4.
• For the modern scholarly picture: Bart D. Ehrman, Lost Christianities, and his Loeb edition of The Apostolic Fathers.