Codex Bezae: Why Its Gospel and Acts Readings Differ

Scrivener describes Greek on the left and Latin on the facing right-hand page. Each page normally has thirty-three lines arranged in sense units rather than mechanically filled to the same width. The Gospels run Matthew, John, Luke, and Mark. A surviving Latin fragment of Third John and the text of Acts extend the collection beyond the four Gospel narratives. His account also records missing material and later replacement leaves. [SRC-01]

This physical arrangement matters when using an edition. The sigla for the Greek and Latin traditions can make the two sides look like separate entries in a list of witnesses. They are distinct texts, but here they occupy the same surviving object. Agreement between them needs to be evaluated within that relationship. Conversely, a disagreement can be informative: a bilingual page is not a guarantee that every Greek expression received an exact matching Latin expression.

Cambridge’s 2017 account dates the codex around 400 and identifies it as MS Nn.2.41, or D05. [SRC-02] Scrivener’s 1864 title instead assigned it to the beginning of the sixth century. That contrast records a change in scholarly dating; it is not evidence of two different manuscripts. Dating must be cited with its method and source, especially when an older edition supplies the transcribed readings. [SRC-01]

One of Scrivener’s substantial examples occurs after Matthew 20:28. The manuscript introduces a longer instruction about humility and taking one’s place at a meal. The invitation to avoid an honored seat recalls the banquet teaching in Luke 14:8–10. Scrivener sets the additional Greek lines beside evidence from ancient versions instead of simply dismissing them because they are longer. [SRC-01]

The important first observation is literary: material resembling a teaching known in one Gospel appears at another location in another Gospel. That gives comparison something concrete to explain. Did a scribe expand one account from another? Did an earlier form of a Gospel contain the material? Did a remembered teaching enter the transmission through a different route? The resemblance identifies the problem; it does not by itself choose among every possible history.

Scrivener’s discussion also prevents a common exaggeration. A reading can be unusual in Greek manuscripts while having parallels in ancient translations. Describing it as entirely without any other witness would erase the comparison on which his argument depends. The right unit of inquiry is the wording, its location, and the witnesses actually cited for it, rather than the exciting size of the addition. [SRC-01]

Scrivener’s survey of Luke includes expansions that make a scene more explicit. At the burial in Luke 23:53 he records an addition concerning a stone that twenty men could scarcely move. His next examples include material in the resurrection narrative associated with the question of moving the stone. These are claims about the readings his edition reports, not independent reports from the burial site. [SRC-01]

Such details can make a narrative easier to picture. That vividness is precisely why their textual position matters. An additional number may become the most memorable part of a story even though the critical question is whether it belonged to the earliest recoverable wording. A reader should resist turning a manuscript’s descriptive expansion directly into an archaeological measurement or an eyewitness headcount.

The same discipline protects the text from an opposite simplification. A longer reading is not disproved merely because it sounds explanatory. A convincing judgment must compare the surrounding passage, other witnesses, and the direction in which a change could plausibly arise. The presence of a smooth explanation is relevant to that reasoning, but it is not an automatic rule that the shorter version must always win.

Acts is central to the manuscript’s reputation. Hixson describes its text as approximately eight percent longer than the standard comparison text. [SRC-02] Scrivener’s detailed treatment of Acts 6 shows why a percentage is only a starting point. He compares vocabulary, grammar, small omissions, and additional phrases within a single chapter rather than treating all differences as one event. [SRC-01]

In the discussion of Acts 6:10, the account of opponents unable to answer Stephen receives fuller explanation: their inability is connected to the truth confronting them and his boldness. Scrivener compares the Greek manuscript with other witnesses, including versional evidence. The extra wording affects how a reader understands the opponents’ resistance; it is more than a change in spelling, yet it does not introduce an entirely different Stephen narrative. [SRC-01]

There are at least two questions to keep apart. The literary question asks what this expanded form makes explicit. The historical question asks when and where that wording entered the transmission. One can answer the first by comparing the sentences. The second requires a much wider pattern of evidence. A single chapter cannot establish the complete ancestry of a manuscript, and a whole-book percentage cannot decide the history of one clause.

Scrivener does not find the same pattern everywhere. His treatment of Mark emphasizes that the text can be comparatively restrained and can share shorter forms with other major witnesses. The manuscript’s reputation for expansive readings must therefore be tested book by book. A description built mainly from its Acts text will mislead if applied without qualification to every Gospel sentence. [SRC-01]

Western is a conventional comparative label, not a passport naming the exact city where every reading originated. In this article it identifies the familiar discussion around the manuscript’s textual character and Gospel arrangement. The useful work lies beneath the label: which witnesses agree here, where do they diverge, and is the shared wording distinctive enough to suggest a relationship?

A further distinction concerns changes within the manuscript itself. Scrivener separates the main text, later corrections, replacement material, and liturgical annotations. Those layers need not come from the same time or serve the same purpose. A correction to a word, a replacement for a lost page, and a direction for public reading should not be merged into a single claim that someone rewrote the Bible. [SRC-01]

Hixson reported reversed Greek ink traces on Latin folio 455r, transferred before the facing Greek leaf disappeared. Reversing the image and comparing letter forms yielded thalassa(n), meaning sea, in Acts 10:6. The final letter was reconstructed because it had not transferred. He also checked traces in person in February 2017. [SRC-02]

The brackets make a valuable boundary visible. They distinguish a proposed completion from the letters represented by surviving traces. A reconstruction can be persuasive without every letter being directly visible, provided the reader is told which parts depend on context. Concealing the bracket would overstate the recovery; refusing all reconstruction would discard a reasoned use of the evidence.

The example also shows how modest a significant discovery can be. It does not need to produce a whole new page of readable Greek to improve knowledge. The relationship between surviving material and a lost leaf is itself historically informative. This account relies on Hixson’s published report; it does not claim that the original traces or manuscript photographs were independently examined here.

Scrivener prints Theodore Beza’s letter of 6 December 1581 to Cambridge. Beza describes the manuscript’s association with St Irenaeus’s monastery at Lyon and acknowledges striking disagreements, particularly in Luke. His reasons for preserving the book include its antiquity and the possibility of comparing it with other evidence. He does not simply call it a fabrication by ancient heretics. [SRC-01]

That letter is evidence for what Beza said when giving the manuscript, not a continuous account of every earlier owner. Scrivener connects its emergence with the disturbances at Lyon in 1562, but the reconstruction belongs to his historical argument. The violent and confessional setting should neither be hidden nor made to prove that the text itself was manufactured during the Reformation. [SRC-01]

For a Catholic reader, the episode invites intellectual fairness. One can examine the responsibilities of collectors and the losses suffered by religious houses while also recognizing the value of preserving a difficult witness. The manuscript’s usefulness does not depend on sharing its donor’s theology. Nor does disagreement with that theology authorize ignoring the evidence that survived through his custody.

Pius XII encouraged Catholic textual criticism and distinguished the Vulgate’s ecclesial authority from a claim that comparison with original-language texts was unnecessary. [SRC-03]

Applied to Bezae, that distinction creates room for honest reading. A Catholic need not make every unusual wording original in order to honor Scripture. Neither must the existence of variation become a verdict that the entire New Testament is unknowable. The responsible task is smaller and more demanding: identify a particular reading, understand what changes, and weigh the evidence for it.

There is also a spiritual discipline in refusing to turn an interesting detail into more than it can carry. An explanation added to a scene may help a later reader imagine it, while still requiring a judgment about its textual history. Prayer and study need not compete for possession of that scene. A person may receive the Gospel with trust and remain willing to correct a mistaken assertion about one of its manuscripts.

The verdict here concerns the documented object and its distinctive bilingual readings. It does not pronounce every variant original, reconstruct a complete family tree, or make a historical edition the last word in scholarship. The manuscript is valuable precisely because its differences can be named, compared, and discussed in public. [SRC-01]