“Attested By” Is Not “Composed In”

The oldest surviving copy of a text is almost never the text’s birthday certificate.

A family Bible printed in 1850 may contain a translation made in 1611. A medieval copy may preserve an ancient sermon. A quotation in one author may reproduce language he inherited from another. The first artifact that happens to survive is a chronological boundary, not necessarily the first time the words existed.

Yet discussions of the Septuagint often slide between those categories. An early Greek fragment appears, and an entire Greek Old Testament is projected backward from it. A late Christian codex is the first substantial witness to a book, and every reading within it is treated as though it arose with the codex. Or, in reaction, the first surviving witness is used as a ceiling beyond which the text is not allowed to be much older.

All three moves claim more than survival can prove.

The proposition of this article is:

An earliest witness establishes that an artifact or wording existed by a certain point; it does not, by itself, fix the date of origin, prove the scope of a larger collection, or justify an arbitrary rule placing composition no more than one century earlier.

“Attested by” is a disciplined conclusion. “Composed in” requires another argument.

What an Earliest Witness Actually Establishes

Suppose a fragment contains five verses of Deuteronomy in Greek.

Direct inspection can establish that those verses are present on that fragment. If the material carries an independent date, or if a well-supported date range is assigned by palaeography and archaeology, the fragment can establish that this Greek wording existed by that period.

That is a terminus ante quem: a boundary before which the wording must already have come into existence.

It does not establish:

• the exact year of the translation;
• the identity of the translator;
• the date of every other Greek biblical book;
• the contents of the collection to which the fragment once belonged;
• the wording of passages not preserved on it;
• the existence of a complete, one-volume Greek Bible;
• the absence of revision between the first translation and this copy.

The artifact answers the question it is capable of answering.

P.Ryl. Greek 458 is a physical witness to portions of Deuteronomy in Greek. The University of Manchester conventionally assigns the papyrus to the second century BC.¹ The direct fact is the Greek Deuteronomy text on the papyrus. The century is a scholarly assignment. The legitimate conclusion is that a Greek form of these passages is attested in an artifact assigned to that period.

That is already important evidence. It needs no enlargement into a complete Alexandrian pandect.

The Copy May Be Much Later Than the Wording

No serious manuscript discipline assumes that every text originated when the surviving copy was made.

Copying is evidence of prior existence. A scribe normally works from an exemplar, dictation, memory, or some combination. The wording before him may be a week old or a thousand years old. The physical copy alone does not measure the interval.

This matters especially for Vaticanus and Sinaiticus.

If these codices were manufactured in late antiquity, their Greek Old Testament books did not therefore originate in late antiquity. Early fragments, Jewish quotations, translation vocabulary, and revision history can support much earlier textual forms. A fourth-century copy can preserve a pre-Christian translation.

The skeptical error would be:

The manuscript is late; therefore the wording must be late.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩

That is no better than the opposite error:

The manuscript is early; therefore every wording in it must be equally early.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩

The first denies transmission. The second denies revision.

A later artifact can preserve an early reading and a late reading on the same page—one in the main text and another in a correction. Even the main text can combine old translation with later harmonization. The age of each layer must be argued from its own evidence.

The Copy May Also Be Close to the Origin

The fact that a surviving witness can be much later than composition does not mean it always is.

A draft may survive in the author’s own hand. A presentation copy may be produced immediately. A translation may be copied within a generation. An early papyrus can stand near the beginning of a text’s transmission.

The distance cannot be decided in advance.

This is why an earliest-witness rule must remain asymmetrical:

• The witness proves that the text is no later than the witness’s secure chronological placement.
• The witness alone does not prove how much earlier the text may be.

Additional evidence may narrow the interval. A datable historical allusion may supply a lower boundary. A quotation may supply an upper one. Linguistic features may fit one setting better than another. Several independent witnesses may show that a reading had already spread widely.

But those are additional arguments. They do not come free with the age of the oldest surviving scrap.

Why the One-Century Rule Fails

It may seem prudent to say that an undated text should be placed no more than one century before its earliest surviving witness.

The rule has an appearance of moderation. It is actually arbitrary.

Why one century rather than fifty years, two centuries, or five? Manuscript survival does not follow a fixed schedule. Climate, war, reuse, burial, library practice, copying frequency, religious change, and sheer accident determine what remains.

The rule also fails every time a work is independently known to be older than its earliest surviving manuscript. If a later copy explicitly preserves a work quoted centuries earlier, the physical manuscript cannot limit the work to the preceding hundred years. The quotation has already carried the terminus ante quem backward.

A simple modern control exposes the mistake. Imagine that every physical copy of the Authorized Version printed before 1900 were destroyed, while the 1611 title-page record, contemporary quotations, and printers’ accounts remained. The earliest surviving complete Bible might then be a twentieth-century object. No rational historian would date the translation to the nineteenth century merely because the older copies disappeared.

The absence of an artifact is not a metronome.

Therefore this series rejects the rule:

No text may be dated more than a century before its earliest surviving copy.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩

The approved formula is:

This wording is attested in witness X, whose production is directly dated or conventionally assigned to range Y. An earlier origin is proposed for reasons Z, at stated confidence.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩

The reasons must do the work.

Applying the Rule to Greek Scripture Before Christ

The pre-Christian Greek biblical fragments correct two opposite overstatements.

First, they disprove the claim that no Greek translation of Hebrew Scripture existed before Christ. P.Ryl. Greek 458, Papyrus Fouad 266, and the Greek Leviticus fragments from Qumran are direct material evidence of particular Greek scriptural texts, with their chronological placement established or assigned through the relevant papyrological and archaeological methods.²

Second, they do not prove one complete, homogeneous Septuagint Bible corresponding to Vaticanus, Sinaiticus, Rahlfs-Hanhart, or the Göttingen editions.

A fragment of Deuteronomy attests Deuteronomy. A fragment of Leviticus attests Leviticus. Agreement between fragments may establish shared translation history. It does not supply the missing leaves of Isaiah, Psalms, Daniel, or the books later included in Christian codices.

The correct conclusion is stronger because it is exact:

Particular Greek translations of particular biblical books are physically attested before Christ; the extent, wording, and collection history of the larger Greek corpus must be demonstrated book by book.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩
2. Israel Antiquities Authority, <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q119-1?locale=enUS”>4Q119, 4QLXXLeviticus and <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q120-1?locale=enUS”>4Q120, 4QpapLXXLeviticus; consult the formal editions for the archaeological and palaeographical arguments. For Papyrus Fouad 266, see Françoise Dunand, Papyrus grecs bibliques (Papyrus F. Inv. 266) (Cairo: Institut français d’archéologie orientale, 1966). ↩

That statement grants every artifact its full value and no more.

Applying the Rule to Later Christian Codices

Vaticanus and Sinaiticus reverse the scale of the problem. Instead of tiny fragments, they preserve large Christian collections.

Their breadth tempts the reader to retroject their unity.

If a late-antique codex contains Genesis through Daniel in Greek, the physical object directly attests that Christian collection at the codex’s production. It does not prove that all its books had been translated together, that every reading is pre-Christian, or that a materially identical Jewish pandect existed centuries earlier.

At the same time, the later date of the codex does not imprison every reading in late antiquity. An earlier papyrus may confirm the same wording. A Jewish author may quote it. A translator’s consistent habits may reveal a stable form. A later revision may be distinguishable from the Old Greek beneath it.

The responsible question is not:

How old is the codex?

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩
2. Israel Antiquities Authority, <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q119-1?locale=enUS”>4Q119, 4QLXXLeviticus and <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q120-1?locale=enUS”>4Q120, 4QpapLXXLeviticus; consult the formal editions for the archaeological and palaeographical arguments. For Papyrus Fouad 266, see Françoise Dunand, Papyrus grecs bibliques (Papyrus F. Inv. 266) (Cairo: Institut français d’archéologie orientale, 1966). ↩

It is:

What is the earliest secure evidence for this particular wording in this particular biblical book?

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩
2. Israel Antiquities Authority, <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q119-1?locale=enUS”>4Q119, 4QLXXLeviticus and <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q120-1?locale=enUS”>4Q120, 4QpapLXXLeviticus; consult the formal editions for the archaeological and palaeographical arguments. For Papyrus Fouad 266, see Françoise Dunand, Papyrus grecs bibliques (Papyrus F. Inv. 266) (Cairo: Institut français d’archéologie orientale, 1966). ↩

The artifact’s date remains relevant. It simply does not answer every textual question.

Survival Is Not a Random Sample

The manuscripts on our desks are not a balanced survey of everything once copied.

Papyrus survives well in some dry regions and badly in damp ones. Parchment was scraped and reused. Worn books were discarded, repaired, rebound, or broken for binding material. Large codices were expensive and vulnerable. Political and ecclesiastical centers preserved some collections and lost others.

Christian institutions transmitted most of the substantial Greek Old Testament evidence now extant. That fact is historically important. It does not mean that Christians invented all the texts they copied. Nor does it mean that Christian copying left every Jewish form unchanged.

Survival favors some places, materials, communities, and accidents.

Therefore absence must be stated negatively:

No earlier witness is presently known.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩
2. Israel Antiquities Authority, <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q119-1?locale=enUS”>4Q119, 4QLXXLeviticus and <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q120-1?locale=enUS”>4Q120, 4QpapLXXLeviticus; consult the formal editions for the archaeological and palaeographical arguments. For Papyrus Fouad 266, see Françoise Dunand, Papyrus grecs bibliques (Papyrus F. Inv. 266) (Cairo: Institut français d’archéologie orientale, 1966). ↩

It must not be converted without further evidence into:

No earlier witness ever existed.

---

1. University of Manchester Library, P.Ryl. Greek 458, catalogue record and images. The date is a palaeographical assignment attached to a directly inspectable Greek Deuteronomy artifact. ↩
2. Israel Antiquities Authority, <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q119-1?locale=enUS”>4Q119, 4QLXXLeviticus and <a href="https://www.deadseascrolls.org.il/explore-the-archive/manuscript/4Q120-1?locale=enUS”>4Q120, 4QpapLXXLeviticus; consult the formal editions for the archaeological and palaeographical arguments. For Papyrus Fouad 266, see Françoise Dunand, Papyrus grecs bibliques (Papyrus F. Inv. 266) (Cairo: Institut français d’archéologie orientale, 1966). ↩

The first reports the archive. The second reconstructs all that has been lost.

The Strongest Objection

If the earliest witness supplies no fixed composition interval, does historical dating become limitless?

No.

An origin claim can be constrained by converging evidence:

1. External attestation: quotations, allusions, catalogues, and independently dated fragments.
2. Internal boundaries: references to events, institutions, vocabulary, or technologies.
3. Translation profile: repeated lexical and grammatical habits that identify a coherent rendering.
4. Textual relationships: shared readings and shared errors that establish descent or revision.
5. Material evidence: substrate, format, ink, and manufacturing features.
6. Historical fit: a proposed date and setting must explain the evidence better than rivals.

These controls can yield a strong range. What they may not do is borrow certainty from a rule invented for convenience.

Historical restraint is not historical paralysis.

Honest Uncertainty Is an Argumentative Strength

Confessional bibliology loses nothing by saying, “This is attested by the second century BC, but its exact origin is uncertain.”

The authority of the Hebrew text does not depend upon making every Greek translation late. The translation remains derivative whether it was made in 250 BC, AD 250, or any point between. Antiquity can increase a version’s historical interest. It cannot transform the stream into the fountain.

That is why we can correct Ruckman and Williams when they deny genuine pre-Christian Greek evidence without yielding the confessional order. We can also correct the opposite exaggeration that treats a few early fragments as proof of an entire, uniform Greek Old Testament with authority over Hebrew.

Both exaggerations ask survival to say what it cannot say.

The earliest witness gives us a floor of certainty: the wording existed by this point, under the qualifications attached to the witness’s own date. Beyond that floor, composition requires argument.

The next article examines the method most often used to assign the dates of undated biblical manuscripts: palaeography, a real and necessary discipline whose conclusions must be reported with real limits.

Notes and Further Reading

For the methodological distinction between a surviving witness, the Old Greek reconstructed from witnesses, and a hypothetical Hebrew parent text, see Emanuel Tov, The Text-Critical Use of the Septuagint in Biblical Research, 3rd ed. (Winona Lake, IN: Eisenbrauns, 2015), and Peter J. Gentry, “The Septuagint and the Text of the Old Testament,” Bulletin for Biblical Research 16.2 (2006): 193–218.

For the transition from individual Jewish Greek translations and revisions to later Christian codices, see Karen H. Jobes and Moisés Silva, Invitation to the Septuagint, 2nd ed. (Grand Rapids: Baker Academic, 2015), and Robert A. Kraft, “The Transmission of the Greek Jewish Scriptures”.

Source-control note: The words physically preserved in a named artifact are direct attestation (E1). A genuine terminus ante quem is a secure boundary (E2), but an artifact dated only by handwriting carries an E3 assigned range rather than a dated production statement. The proposed origin of a translation or reading before its earliest witness is a textual-historical reconstruction (E5). A fixed one-century interval unsupported by independent evidence is not an evidence class and is excluded.

Previous in the series: “Four Dates Every Manuscript Claim Must Separate.”

Next in the series: “Palaeography: A Real Tool with Real Limits.”