Four Dates Every Manuscript Claim Must Separate

“This manuscript dates to the fourth century.”

The sentence sounds like one fact. It may contain four different claims.

It may mean that the physical book was manufactured in the fourth century. It may mean that the wording copied into it originated then. It may mean that modern scholars have assigned the handwriting to that period. Or it may mean that the manuscript first appears in a record from that period.

Those claims can coincide. They often do not.

A page has a material history. A text has a transmission history. A date assignment has an argumentative history. A manuscript’s ownership has a documentary history. When those histories are compressed into one bare date, an inference can acquire the appearance of direct evidence.

The proposition of this article is:

Responsible manuscript history must distinguish at least four chronological claims: the production of the physical artifact, the age of the textual form it contains, its assigned scholarly date, and its first securely documented provenance.

The distinction does not make ancient manuscripts useless. It makes our description of them true.

Date One: Production of the Artifact

The first question concerns the material object.

When were these sheets of papyrus or parchment prepared, written, assembled, and made into this roll or codex?

For a printed book, a title page may name a printer and year. For a manuscript, a colophon may record when the copying was completed. A dated letter, legal document, or administrative register may supply its own chronological setting. In such cases the date can be directly attested, though even a colophon must be examined as part of the artifact rather than trusted mechanically.

Many biblical manuscripts contain no such date.

Their production is then estimated from evidence such as:

• handwriting;
• page design;
• quire construction;
• ruling patterns;
• abbreviation and numeral systems;
• inks and pigments;
• the physical writing surface;
• textual apparatus that cannot predate its invention;
• comparison with independently dated artifacts.

That evidence may strongly favor one period. It may exclude others. It may yield a broad range. What it does not do is turn an undated codex into a book bearing a publication statement.

The production date belongs to the object, not automatically to every wording inside it.

Date Two: The Textual Form

A copy made yesterday can contain words written centuries ago.

That is the ordinary purpose of copying. The physical age of a King James Bible printed in 2026 is not the age of its translation. A seventeenth-century manuscript of Augustine does not make Augustine a seventeenth-century author. A medieval Homer does not make the Iliad medieval.

The same rule applies to biblical manuscripts.

A late-antique Greek codex may preserve a translation made much earlier. It may also preserve later revision, harmonization, correction, or expansion. The date of the parchment does not settle which layer a particular reading represents.

The textual historian must therefore ask a second set of questions:

• Is this wording shared by an earlier fragment?
• Is it quoted by a writer whose date supplies an independent boundary?
• Does it belong to the translator’s normal vocabulary and syntax?
• Does it appear to revise an older Greek form toward Hebrew?
• Is it confined to later Christian witnesses?
• Is it written by the original scribe or a later corrector?
• Does it depend upon a New Testament quotation or another parallel passage?

Answers to those questions reconstruct the age and history of the wording. They do not date the animal skin on which the surviving copy was written.

This is why “Vaticanus is fourth-century” and “this reading is fourth-century” are not equivalent statements. Even if the conventional date assigned to the codex is correct, the reading could be older than the artifact, contemporary with its production, or introduced by a later hand.

Date Three: The Scholarly Assignment

An assigned date is a reasoned estimate.

The term does not imply deceit or guesswork without evidence. It identifies the logical status of the conclusion. Scholars inspect features of an undated object, compare them with other evidence, and assign the object to a period in which those features are judged to fit.

Palaeography—the comparative study of handwriting—is often central. Codicology adds the archaeology of the book itself: support material, format, gatherings, ruling, binding, sequence, and physical alterations. Historical arguments may provide termini. Scientific testing may date organic material within calibrated probability ranges.

These methods can converge. They can also carry different degrees of precision.

The official Codex Sinaiticus Project says that the codex is generally dated to the fourth century on the basis of palaeographical analysis.¹ Brent Nongbri’s renewed examination of the actual arguments concludes that the responsible range should be broader—approximately the early fourth through the early fifth century—and explains why a habitual date around the middle of the fourth century is too precise.²

That disagreement does not show that one side has a dated fourth-century invoice while the other ignores it. It concerns the range warranted by comparative and internal evidence.

The proper report is therefore not:

Sinaiticus was written in AD 350.

---

1. Codex Sinaiticus Project, “Date”. The project identifies palaeographical analysis as the basis of the usual fourth-century assignment. ↩
2. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34. Nongbri argues for an evidentiary range of approximately AD 300–425 and proposes accelerator mass spectrometry radiocarbon analysis as a possible means of narrowing it. ↩

It is:

Sinaiticus is conventionally assigned to late antiquity, most commonly the fourth century; a recent reassessment argues that the evidentiary range should extend from the early fourth into the early fifth century.

---

1. Codex Sinaiticus Project, “Date”. The project identifies palaeographical analysis as the basis of the usual fourth-century assignment. ↩
2. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34. Nongbri argues for an evidentiary range of approximately AD 300–425 and proposes accelerator mass spectrometry radiocarbon analysis as a possible means of narrowing it. ↩

The second sentence says both more and less. It says more about the evidence and less than the evidence cannot prove.

Date Four: Documented Provenance

Provenance is the documented history of an artifact’s location, ownership, custody, or examination.

A clear library inventory may place a manuscript on a particular shelf. A traveler may describe an identifiable object. Correspondence may record a collation. A receipt may document a loan. A sale or donation may transfer custody.

The earliest secure record establishes that the object existed by that time. It does not tell us how long it had already existed.

Codex Vaticanus is probably represented in a Vatican inventory of 1475 and is unmistakably identified by the more specific 1481 description of a three-column parchment Bible bound in red.³ That is strong fifteenth-century provenance. It is not a fifteenth-century production date.

Codex Sinaiticus presents a different documentary history. A description written by Vitaliano Donati in 1761 may refer to it. The identification is possible rather than certain. The manuscript is unmistakably documented when Constantine Tischendorf obtained and later published forty-three leaves beginning in 1844.⁴

Again, the correct conclusion is limited:

• Vaticanus is securely documented in the Vatican collection in the fifteenth century.
• Sinaiticus is possibly noticed in 1761 and securely documented from 1844.

Neither statement proves manufacture at the time of first documentation. Neither supplies an unbroken chain of custody back to the proposed production period.

Documented provenance and assigned production must stand beside one another without being confused.

Termini Are Boundaries, Not Secret Dates

Two Latin expressions are useful if kept plain.

A terminus post quem is a boundary after which an artifact must have been produced. If a codex incorporates a system demonstrably invented after a certain time, the codex in that form cannot predate that invention.

A terminus ante quem is a boundary before which an artifact or wording must already have existed. If an unmistakable inventory records a manuscript in 1481, that physical manuscript existed by 1481.

Neither boundary necessarily identifies the year of production.

Suppose an undated manuscript contains an apparatus invented sometime between AD 300 and 325 and is unmistakably listed in an inventory of 1481. Those facts yield a broad interval. Handwriting and codicology may narrow it. They do not convert either endpoint into the actual year in which the scribe worked.

The same discipline applies to a textual form. A writer who quotes a distinctive wording in AD 200 establishes that the wording existed by then. He does not prove whether it arose in 199, 150, or 50.

Chronological boundaries prevent impossibilities. They do not manufacture precision.

One Object, Four Records

Every significant manuscript claim in this series will use the following record.

This table prevents a bare century label from doing several jobs at once.

The Strongest Objection

One might object that this vocabulary overloads an ordinary sentence. Specialists know that “fourth-century manuscript” normally means “a manuscript assigned to the fourth century.” Must every paragraph carry an evidentiary warning label?

No.

Once the basis and limits have been stated, a writer may use reasonable shorthand. Repeating the full record every time would bury the argument.

But shorthand becomes dangerous when chronology is itself doing argumentative work. If a manuscript is preferred because it is “the oldest,” if a reading is declared pre-Christian because it appears in a late-antique codex, or if a gap in provenance becomes an accusation of forgery, the evidence categories must be explicit.

The issue is not whether scholars may abbreviate. It is whether the abbreviation conceals the premise being disputed.

There is an equal danger on the skeptical side. Saying that a date is assigned does not make it arbitrary. Saying that first provenance is late does not make manufacture late. Saying that a reading’s ancestry is reconstructed does not make it imaginary.

An estimate may be excellent. A reconstruction may be compelling. It remains an estimate or reconstruction, and its strength is shown by the reasons offered for it.

Historical Evidence and Scriptural Authority

Confessional bibliology does not need a false historical certainty in order to defend the Hebrew text.

The authority of the Hebrew Old Testament does not rise or fall with our ability to name the workshop and year of every Greek codex. Nor does an assigned early date give a translation authority over its source. Historical inquiry determines what an artifact can responsibly be said to witness. Scripture determines the order of authority between the fountain and the stream.

That distinction frees the argument from two temptations.

We need not exaggerate uncertainty until every ancient book becomes unknowable. We need not exaggerate confidence until conventional dates become eyewitness certificates.

The manuscript may be a venerable witness. It is still a witness. Before it testifies, the court must know whether its date belongs to the skin, the script, the wording, the catalogue, or the scholar’s reconstruction.

The next article will address a related mistake: treating the earliest surviving witness as though it fixes the date when the text was composed.

Notes and Further Reading

For the distinction between a physical witness and the reconstructed textual history carried by it, see Emanuel Tov, The Text-Critical Use of the Septuagint in Biblical Research, 3rd ed. (Winona Lake, IN: Eisenbrauns, 2015), and Brent Nongbri, “Palaeographic Analysis of Codices from the Early Christian Period: A Point of Method,” Journal for the Study of the New Testament 42.1 (2019): 84–97.

Source-control note: The existence and visible features of named artifacts are direct evidence (E1). Genuine chronological boundaries are E2. Palaeographical and codicological production ranges are conventional assignments (E3). Inventory, travel, loan, and custody records are documented provenance (E4). The age of a textual form copied into a later object is a reconstruction when no independent early witness fixes it (E5). Unattested origin stories remain hypotheses (E6).

Previous in the series: “From Jewish Translations to Christian Pandects.”

Next in the series: “‘Attested By’ Is Not ‘Composed In.’”