How Was the Bible’s Canon Decided? Who Chose the Books, and Why

In Dan Brown’s The Da Vinci Code, a character explains the Bible with the breezy confidence of a man who has done no homework: “The Bible, as we know it today, was collated by the pagan Roman emperor Constantine the Great,” who “commissioned and financed a new Bible, which omitted those gospels that spoke of Christ’s human traits and embellished those gospels that made him godlike”; the same character claims that Jesus’s divinity itself was settled at the Council of Nicaea in 325 by “a relatively close vote.”¹ The novel sold eighty million copies, and its account of the canon—an emperor and his bishops choosing, suppressing, and voting—is now what a great many people believe actually happened.

It is worth saying plainly at the outset: almost every detail of that story is false. The Council of Nicaea did not vote on the canon, did not discuss the canon, and left no record of anyone raising the subject; its business was the relationship of Christ to the Father in the Arian controversy.² The truth is at once less cinematic and more interesting. The canon was not decided at a moment; it emerged over centuries. And the better way to describe what the church did is not that it created a canon but that it gradually recognized one—sifting the writings already circulating under its hand, testing them against criteria, and slowly arriving at a list. How that happened, what drove it, and what it does and does not prove, is the subject of this essay.

Stating the Question Carefully

“Canon” comes from the Greek kanōn, a measuring reed or rule—a standard. To speak of the biblical canon is to speak of the list of books the church received as the rule of its faith. The historical question is: by what process did that list take shape, and on what grounds?

Two things must be distinguished from the start, because confusing them is the source of most of the heat in this debate. The first is the historical question—how the canon was formed—which is the question this essay can actually settle on the evidence. The second is the theological question—whether the books in that canon are in fact inspired and true. These come apart cleanly. Showing that the early church recognized twenty-seven books by reasonable criteria is a claim historians of any or no faith can assess. Showing that those books are God-breathed is a different argument, resting on different grounds. A reader should not expect the second conclusion to fall out of the first. As with the question of manuscript transmission—which establishes preservation, not truth—recognizing the canon historically is not the same as proving its contents inspired. The honest apologetic gain here is narrower: it dismantles the conspiracy story, and shows the canon to be the fruit of a discriminating, centuries-long reception rather than a fourth-century political fiat.

The argument, stated as the church’s defenders have put it, runs like this. Premise 1: if the church recognized a set of writings that already carried authority, rather than conferring authority on them by decree, then the canon reflects something the church found rather than something it manufactured. Premise 2: the historical evidence shows the core books were received, copied, and cited as Scripture long before any council ratified a list. Conclusion: therefore the canon is best understood as recognition, not invention. The logic is straightforward; the work is in defending the premises and marking the conclusion’s limits.

How the Canon Actually Formed: A Timeline

The process has no single founding moment, but it has a discernible shape. Within the first century, Paul’s letters were already being collected and exchanged among churches. The exchange is visible inside the letters themselves. Colossians closes with an instruction: “And when this letter has been read among you, have it also read in the church of the Laodiceans; and see that you also read the letter from Laodicea” (Colossians 4:16). Many critical scholars doubt that Paul wrote Colossians and date it to the years or decades after his death, while others defend his authorship; on either view it is usually dated within the first century, so the verse presupposes, as a practice its first readers would recognize, letters in Paul’s name being read in one congregation and passed on to the next within the first century—the seed of a collection. It shows circulation, not yet a fixed collection, let alone a canon. Remarkably, 2 Peter 3:15–16 (first-century on the traditional dating, early second-century on many critical datings—early on either) refers to Paul’s letters and groups them with “the other Scriptures,” an early sign that apostolic writings were being read as authoritative alongside the Old Testament.³

One of the clearest early marks of that authority is liturgical: the apostolic writings were read aloud in the churches’ worship from the first generation—and, by the mid-second century, demonstrably alongside the Old Testament. Colossians directs that the letter be read publicly and exchanged between congregations (4:16, quoted above); the Pastoral Epistles urge attention to “the public reading of Scripture” (1 Timothy 4:13); and Revelation pronounces a blessing on “the one who reads aloud the words of this prophecy” in the gathered assembly (Revelation 1:3). By around AD 150 the practice is explicit: Justin Martyr describes the Sunday gathering, where “the memoirs of the apostles or the writings of the prophets are read, as long as time permits” (First Apology 66–67), placing the apostolic “memoirs” on a footing with the prophets. To read a text in worship was itself a public act of recognition—to treat it as functioning Scripture. This attests early functional authority; it does not by itself prove a closed canon at that date, since which books were read could still vary from region to region.

By around AD 180, the boundary around the four Gospels was firm enough that Irenaeus of Lyons could treat any deviation from exactly four as self-evidently absurd. In Against Heresies he insists “it is not possible that the Gospels can be either more or fewer in number than they are,” comparing them to the four zones of the world and the four winds—a strange-sounding argument to modern ears, but one that only works if a fourfold Gospel was already an established fact his readers would not dispute.⁴ Around the same period—most scholars say roughly AD 170–200—an anonymous list now called the Muratorian Fragment names most of the books we know: the four Gospels, Acts, thirteen letters of Paul, Jude, two (perhaps three) letters of John, and Revelation. It is mutilated at both ends, and it also accepts the Wisdom of Solomon and, with hesitation, the Apocalypse of Peter while omitting Hebrews, James, and 1–2 Peter—but it is the oldest surviving canon list we possess, and it already looks substantially like the New Testament.⁵

In the early fourth century—and crucially, before Nicaea—the historian Eusebius of Caesarea gives us a snapshot of where things stood. In his Church History (3.25) he sorts the writings into the “recognized” (homologoumena), the “disputed” (antilegomena), and the “spurious” (notha). The recognized books include the four Gospels, Acts, Paul’s letters, 1 John, and 1 Peter; the disputed include James, Jude, 2 Peter, 2–3 John, and (with hesitation on either side) Revelation; the spurious include works like the Shepherd of Hermas and the Apocalypse of Peter.⁶ This is the decisive evidence against the conspiracy story: a careful sorting by inherited usage was already underway before Constantine’s council met.

The first surviving list that matches our twenty-seven exactly comes in 367, in the 39th Festal Letter of Athanasius, bishop of Alexandria—an annual Easter letter in which he enumerates the books and calls them “fountains of salvation,” warning, “let no man add to these, neither let him take ought from these.”⁷ The language is telling: Athanasius writes as one recording a settled deposit, not legislating a new one. Finally, the regional Councils of Hippo (393) and Carthage (397) listed and affirmed the canon for the churches of North Africa.⁸ By then the core had been read as Scripture for the better part of three centuries. The councils did not vote books into existence; they ratified a consensus long since reached.

Premise 1: The Criteria Were Principled, Not Arbitrary

The heart of the “recognition” case is that the church was not choosing favorites but applying tests. Three criteria recur in the sources. The first and most important was apostolicity: did the writing come from an apostle or a close associate of one, and so trace back to the founding generation? The second was orthodoxy, or what the fathers called the “rule of faith”: was the book consonant with the teaching the churches had received? The third was catholicity, or widespread use: was the book received and read across the scattered churches, not just prized by one faction in one region?⁹

These criteria behave like the marks of a discovery, not an invention. A community that asks “is this writing genuinely apostolic, and has it been received everywhere?” is acting like someone trying to find out a fact, not like someone free to decide one. The Edinburgh-trained scholar Michael J. Kruger argues that the canon is in this sense “self-authenticating”: the books that were received bore divine qualities, apostolic origins, and corporate reception that together account for their acceptance without appeal to any external authority that conferred status on them.¹⁰ Kruger’s model is developed within the Reformed tradition; Catholic and Orthodox Christians also hold that the church recognized rather than created the canon, but locate that recognition in the church’s authoritative judgment—a difference this essay does not adjudicate. One need not adopt any of these theological models to grant the historical point underneath them: the church behaved as though it were measuring candidates against a standard it did not feel free to invent.

Premise 2: A Functioning Core Long Before Any Council

The second premise is where the historical record does its heaviest lifting, and it is largely uncontested. Long before any list-affirming council, the central books were already functioning as Scripture. Irenaeus around 180 quotes the four Gospels and Paul as binding; the Muratorian list around the same time already contains the great majority of the New Testament; and the earliest fathers cite these books the way they cite the Old Testament—as authority, not as candidates under review. The disputes that did occur clustered at the edges. What was contested were a handful of shorter or later-attested books; what was never seriously in question were the four Gospels, Acts, and the letters of Paul—precisely the doctrinal core. A skeptic and a believer surveying the second-century evidence will agree on this much: the center was fixed early, and only the margins were argued over.

Treating the Disputed Books Honestly

That brings us to the books the early church genuinely doubted—the antilegomena, the “spoken against.” Honesty requires naming them and not minimizing the doubt. Hebrews was questioned in the West because its authorship was uncertain (it does not name Paul, and many doubted he wrote it). James, 2 Peter, 2 and 3 John, and Jude were received slowly and unevenly; 2 Peter is the most weakly attested book in the entire New Testament in the early centuries. Revelation, oddly, was widely accepted in the West but doubted for a long time in the Greek East, partly because of its use by millenarian movements.¹¹

It would be dishonest to pretend these doubts were trivial or quickly resolved—in some regions they persisted into the fifth century and beyond. But two things should be said. First, the very existence of a category of “disputed” books shows the church was not rubber-stamping; it was sifting, sometimes for generations, before settling. That is what careful recognition looks like, not what a coordinated invention looks like. Second, none of the disputed books carries a central doctrine the rest of the New Testament does not teach independently; the faith does not stand or fall on 2 Peter or Jude. The shape of the evidence—a stable, early core and a slowly resolved fringe—is the signature of discovery, not decree.

The Strongest Objections, Stated at Full Strength

The strongest modern challenge does not come from Dan Brown but from the agnostic New Testament scholar Bart D. Ehrman, whose Lost Christianities (2003) reframed the question for a wide readership; here is its case at its best.¹² Ehrman’s case is not the crude “Constantine voted” myth—he has explicitly debunked that himself. It is subtler and harder. He argues that the earliest Christianity was not a single stream with a few heretical eddies but a roiling diversity of Christianities: Ebionites who kept the Law and denied Jesus’s divinity, Marcionites who severed Christianity from Judaism entirely, and various Gnostic groups, each with its own gospels and revelations, each claiming apostolic pedigree. On this telling, what we call “orthodoxy” was simply the party that won; the canon is “history written by the winners,” and the rejected gospels—Thomas, Peter, Mary, Judas—were silenced not because they were demonstrably late but because they lost a power struggle.

The factual kernel of this objection is accurate, and an honest apologist should grant it. The early Jesus movement was more diverse than the tidy Sunday-school picture suggests; there really were competing groups with competing texts; and the proto-orthodox did prevail. What does not follow is the inference that prevailing is all that orthodoxy was.

A second, sharper version of the objection presses on the criteria themselves. The criterion of orthodoxy, a critic will say, is circular: a book counted as orthodox if it agreed with the views of the group doing the judging, so “testing against the rule of faith” just means “keeping the books that already agree with us.” And the criterion of apostolicity, the critic adds, was applied inconsistently—Hebrews was kept despite uncertain authorship, while other early texts were excluded. So the “criteria” look less like neutral tests and more like rationalizations for choices made on other grounds.

A Reply—and What It Can and Cannot Carry

Take the rejected gospels first, because here the reply is strongest. The decisive point is chronology, not theology. The canonical four Gospels date to the first century, within living memory of the eventsSee the evidenceThe evidenceSourceThe four Gospels date to the first century; the rejected gospels — the Gospel of Peter, Judas, and the rest — are, as we have them, second-century or later. The Gospel of Judas (a Sethian Gnostic text, published by the National Geographic Society in 2006 from the Codex Tchacos) is generally dated to c. mid-second century — far too late to come from Judas. The Muratorian Fragment (on the majority dating, c. AD 170–200), the oldest surviving canon list, already looks substantially like the New Testament.Who holds itThe decisive point is chronology, not theology — and the first-century dating of the four is held by the great majority of critical scholars (even Ehrman affirms it). Thomas as we have it is a second-century Greek composition, though whether it preserves an early independent core is contested. (This shows the canonical four are the early ones; it does not claim the church’s every judgment was infallible — some authorship attributions are contested.). The Gospel of Thomas as we have it, the Gospel of Peter, the Gospel of Judas, and the rest are second-century or later compositions—the Gospel of Judas, recovered and published in 2006, is a Sethian Gnostic text generally dated to around the mid-second century, far too late to come from Judas.¹³ These works were excluded first and foremost because they failed the test of apostolic origin on historical grounds: they were written generations after the apostles whose names they borrow, and often betray that lateness in their developed Gnostic theology. The four canonical Gospels were already being received across the churches before most of these rivals were even composed. The exclusion was not, at root, the silencing of equally ancient witnesses; it was the recognition that the late pseudonymous works were exactly that.

The Gospel of Thomas is the one rival most often pressed as an early, independent witness, so it repays a further word—and here the reply is literary as much as theological. Thomas survives in full only in a fourth-century Coptic translation, with three earlier Greek fragments, and a strong line of scholarship argues that at points it preserves the distinctively Lukan and Matthean reworking of sayings the Synoptics share—which would mean it is drawing on the finished Gospels rather than on an independent stream. This is genuinely contested, and the honest thing is to say so. Mark Goodacre (Thomas and the Gospels, 2012) makes the case for dependence in detail, and Simon Gathercole argues that Thomas knows the Synoptics—their editorial fingerprints included—without being a mere copy of them; April DeConick and Stephen Patterson, on the other side, defend an early and independent core. What the dispute does not turn on is theology: even on the more cautious readings, Thomas as we have it is a second-century Greek composition that stands downstream of the first-century Gospels, not a first-century rival to them.

It is worth gently correcting a specific exaggeration at this point. The Da Vinci Code has its character claim that the early church weighed “more than eighty gospels” and that Constantine chose among them; both halves are mistaken. Scholars who actually catalogue the surviving gospel and gospel-like writings count only a few dozen—on the order of forty known titles, many of them preserved merely as names or in scraps (see Bart D. Ehrman and Zlatko Pleše, The Apocryphal Gospels: Texts and Translations [New York: Oxford University Press, 2011]; and Hans-Josef Klauck, Apocryphal Gospels: An Introduction [London: T&T Clark, 2003])—not eighty, and the great majority are second-century or later. And Constantine selected nothing: the four-Gospel core was already functioning as Scripture generations before he was born, and, as we saw at the outset, the Council of Nicaea he convened never took up the canon at all (see Darrell L. Bock, Breaking the Da Vinci Code [Nashville: Nelson, 2004]). Correcting the number and the emperor is not a debating point scored but simply the record set straight.

The accurate part of the objection has already been granted: early Christianity was diverse, and the boundaries were drawn more sharply over time, partly in response to second- and third-century conflicts. What that concession does not establish is that the settlement was arbitrary or that the losing texts had an equal historical claim. Diversity at the start is fully compatible with a principled sorting later. That some groups lost does not by itself show the winners were wrong to keep the early apostolic writings and set aside the late forgeries; that is a separate claim, and the chronological evidence cuts against it. So the “history written by the winners” objection does not hold: the Gospels the church kept are the first-century ones, and the rival gospels it set aside are, as we have them, second-century or later.

On the charge that orthodoxy is a circular criterion, the reply is to grant the possibility of abuse while showing that it did not decide the outcome. Yes, “agrees with the rule of faith” could in principle be used to launder a predetermined result. But the rule of faith was not drawn up to fit a list of books. It was a short summary of the apostolic preaching, handed down in the churches, and it is attested independently of any list. Around AD 180 Irenaeus sets it out—one God the Father Almighty, one Christ Jesus the Son of God “who became incarnate for our salvation,” and the Holy Spirit—as the faith the church “has received from the apostles and their disciples” (Against Heresies 1.10.1), and he says that many nations of barbarian believers hold it “without paper or ink,” “in the absence of written documents” (3.4.2). A little later Tertullian names it the regula fidei, the “rule of faith,” and states its content in the same creedal shape (Prescription Against Heretics 13; On the Veiling of Virgins 1). Irenaeus writes in the same generation as the Muratorian Fragment on its majority dating, and long before it on the minority fourth-century dating; either way, the standard did not depend on the list, and these summaries do not name which books belong in the canon. That answers one form of the charge—that the criterion was reverse-engineered from a predetermined list. It does not answer the deeper form—that the rule itself was simply the winning party’s view: the argument does not prove the criterion infallible, and it does not by itself show that the summary faithfully preserves what the apostles taught; that is a further historical question.¹⁴ And in practice the criteria worked together and constrained one another. Apostolicity and catholicity are not circular in that way—they ask when a book was written and where it was received, questions the judges were trying to get right on historical grounds, even where their tools sometimes failed. The proof that the criteria had teeth is that the church kept books it would have found convenient to drop. Revelation was doubted by serious orthodox figures for centuries yet retained; Hebrews was kept despite its anonymity; 2 Peter survived a long shadow of doubt. A purely results-driven process would have been tidier than the messy, hesitant one the sources actually record.

The same discipline shows in what the church refused. Around AD 200, Serapion, bishop of Antioch, visited the congregation at Rhossus, on the coast near Antioch, found it divided over a book circulating as the Gospel of Peter, and—not having read it—allowed it to be read. When he later obtained a copy and read it through, he found much in it “in accordance with the true doctrine of the Saviour, but some things added to that doctrine,” and wrote to refute it, stating the principle he was applying: “we 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.” At about the same time Tertullian reports that the presbyter in Asia who composed the Acts of Paul—the book that told the story of Thecla, then being cited as a precedent in a dispute over church practice—was convicted, confessed that he had written it “from love of Paul,” and was removed from his office; a century later Eusebius still classed the Acts of Paul among the spurious writings, even though some, Origen among them, had continued to quote it with approval. Neither a pious motive nor a congregation’s attachment was allowed to settle the question. Two limits belong beside these cases. A critic can fairly point out that Serapion’s verdict was partly doctrinal—the very criterion the circularity charge targets; but his stated ground was also one of provenance, a book “not handed down to us,” and he reached his verdict by obtaining and reading the text, not by reflex. And these episodes show that the church tested books for apostolic origin and orthodoxy and was willing to reject even one already in use; they do not by themselves prove that every judgment it made was right.¹⁵

And here the last word on this point should go to the critic before the reply. Ehrman would press: you keep appealing to “apostolic origin,” but several of your own canonical books are pseudonymous by the judgment of most critical scholars—the Pastoral Epistles, 2 Peter, perhaps Ephesians—so the church enforced a criterion it could not actually verify and sometimes got wrong. That is the objection at its sharpest, and it must be acknowledged that the church’s judgments about authorship were made with the limited tools of antiquity and are regarded as insecure by many modern critics—though those judgments are themselves contested by other scholars, evangelical and Catholic, and the argument here does not turn on any single attribution being correct.¹⁶ The reply is bounded but real: even granting disputed authorship for a few books, the argument of this essay does not require that every canonical attribution be correct. It requires that the canon was formed by a discriminating reception keyed to perceived apostolic connection and widespread use—and that is true whether or not every individual judgment about authorship would survive a modern seminar. The process was principled even on the critics’ own assumptions about authorship.

Denominational Differences: A Fair Statement, Not a Verdict

One genuine and unresolved difference must be presented carefully, because it touches a live disagreement among Christians, and this essay will not adjudicate it. On the twenty-seven-book New Testament, Catholics, Eastern Orthodox, and Protestants are unanimous. The difference lies in the Old Testament. Catholic and Orthodox Bibles include a set of books and additions—Tobit, Judith, Wisdom, Sirach, Baruch, 1 and 2 Maccabees, and parts of Daniel and Esther—that they call deuterocanonical (“second canon”) and that Protestants call the Apocrypha and place outside the canon (Orthodox canons are broader still).¹⁷

The roots of the disagreement are ancient and serious on both sides. These books appeared in the Greek Septuagint used by many early Christians, and were cited by numerous Fathers; yet Jerome, the great translator of the Latin Vulgate, distinguished them from the books of the Hebrew canon and judged them useful for edification but not for establishing doctrine.¹⁸ His younger contemporary Augustine of Hippo, by contrast, counted Tobit, Judith, the two books of Maccabees, Wisdom, and Ecclesiasticus within “the whole canon of Scripture,” while also advising readers to weigh a book by how widely the churches received it; a generation earlier Athanasius, in the same Festal Letter that gives the first surviving list of the twenty-seven New Testament books, had placed Wisdom, Sirach, Esther, Judith, and Tobit among books “not indeed included in the Canon, but appointed by the Fathers to be read.”¹⁹ The Reformers, following Jerome’s distinction, adopted the shorter Hebrew canon; the Catholic Church, pointing to the North African councils of Hippo (393) and Carthage (397), whose lists had already included these books, and to the Council of Florence, whose Bull of Union with the Copts (1442) listed them among the books it “accepts and venerates,” formally affirmed them as canonical Scripture at the Council of Trent in 1546.²⁰ Both traditions can point to early precedent; this is not a case of one side inventing books and the other being faithful. It is a real and reasonable disagreement about which Old Testament canon the church inherited, and it is not the burden of this essay to settle it. What matters for our question is that the disagreement is confined to the extent of the Old Testament and leaves the New Testament canon—and the entire conspiracy-theory framing—untouched.

The Honest State of Scholarship

Among historians who actually work on the canon, the broad outline is not seriously contested, and it crosses the ideological spectrum. There is wide agreement that the core books were received early, that no ecumenical council “created” the canon, that Nicaea did not address it, and that the process was one of gradual recognition extending over centuries. Even Ehrman affirms the basic chronology and rejects the Constantine myth.²¹ The live debates are more refined: how early the four-Gospel collection truly crystallized; whether the Muratorian Fragment is late-second-century (the majority view, defended by scholars like Joseph Verheyden) or fourth-century (a minority view argued by Geoffrey Hahneman and Albert Sundberg);²² how much weight to give “intrinsic” versus “community” factors in explaining reception; and how to characterize the diversity of earliest Christianity. These are real disagreements among serious people. But they are disagreements about the mechanism and timing of recognition, not about whether the canon was voted into being by an emperor. On that, the field is settled, and the popular myth has no defenders among specialists.

Conclusion: What This Does and Does Not Establish

The historical case establishes a great deal, and it is worth saying clearly. The canon was not handed down by Constantine, not voted on at Nicaea, and not imposed by a late council on a church that had no Bible until then. The core of the New Testament was received as Scripture within the first and second centuries, sifted by criteria of apostolic origin, orthodoxy, and widespread use, and only formally ratified once a consensus had long existed. The disputed books were genuinely disputed—evidence of care, not of fabrication—and the differences that remain among Christians concern the extent of the Old Testament, not the New Testament. The conspiracy story collapses under the weight of the actual sources.

But the limits must be stated as plainly as the strength. Showing that the church recognized these books does not, by itself, show that they are inspired or true. A discriminating, principled, centuries-long reception is exactly what we would expect if these were the genuine apostolic writings—but it is also, a skeptic may insist, compatible with their being merely the earliest and best-attested human documents of a new religion. The historical argument clears the ground; it removes the charge that the Bible is a fourth-century political invention and shows the canon to be a thoughtful inheritance. Whether the books it preserves are the word of God is a further question, argued on further grounds—how early and how close to the events the Gospels stand, and whether the Scriptures cohere as the unified testimony they claim to be, and, decisively, whether Jesus rose from the dead. Recognizing the canon is where the conversation about its truth can honestly begin—not where it ends.