Palaeography: A Real Tool with Real Limits

Handwriting is evidence of time. It is not a clock.

A trained palaeographer can see features that an untrained reader misses: how a pen enters and leaves a stroke, whether a letter is made in one movement or several, the angle of a crossbar, the spacing of a line, the treatment of numerals, the rhythm of a practiced bookhand. Comparison can place an undated hand among other hands and exclude implausible periods.

But a scribe did not wake on the first day of a new century and begin forming alpha differently.

Styles overlap. Older scribes continue older training. Younger scribes imitate prestigious models. Documentary hands and formal bookhands change at different rates. Two people writing at the same desk can look generations apart.

The proposition of this article is:

Palaeography can assign useful comparative date ranges, but its precision depends upon securely dated comparanda and converging evidence; it cannot turn an undated hand into a dated colophon.

To deny the method would be foolish. To hide the limits of its conclusion would be equally so.

What Palaeography Does

Palaeography is the disciplined study of historical handwriting.

It helps scholars:

• identify scripts and scribal conventions;
• distinguish hands within a manuscript;
• read damaged or unfamiliar writing;
• compare formal and informal styles;
• locate writing within broad geographical and chronological traditions;
• estimate the production period of undated manuscripts.

Dating is only one part of the discipline.

The basic procedure is comparative. An unknown hand is placed beside samples whose date is better established. The scholar studies repeated features rather than one attractive resemblance.

Among the features considered are:

No single row supplies a date by itself. A date assignment arises from a pattern.

The Problem of the Comparison Set

An unknown manuscript can be dated comparatively only as well as the objects used to compare it.

The ideal comparandum is:

1. securely dated by evidence independent of handwriting;
2. similar in material and writing function;
3. geographically and culturally relevant;
4. extensive enough to display the hand’s normal range;
5. available in images good enough for stroke-level examination.

Ancient Greek literary manuscripts rarely provide that ideal set.

Administrative documents, contracts, petitions, receipts, and private letters may carry dates because their business required them. Literary books often do not. A formal biblical majuscule must therefore be compared partly with dated documentary hands and partly with literary manuscripts whose own dates may also depend on palaeography.

This creates two hazards.

The first is circularity. Manuscript A is dated from B; B is dated from C; C’s date ultimately rests upon its resemblance to A. The chain may contain genuine expertise, but no independent anchor.

The second is mismatch. A rapidly written contract and a monumental Bible can be contemporary without looking alike. Function, training, speed, pen, surface, and intended audience shape the hand.

Brent Nongbri’s methodological work calls attention to the small number and uneven character of securely datable literary comparanda for early Christian codices.¹ His point is not that comparison cannot work. It is that the confidence and width of the proposed range must reflect the actual comparison set.

Style Is Not a Single March Through Time

Some palaeographical schemes are presented as an orderly development: a script forms, matures, and declines. Such categories can describe real tendencies. They become dangerous when converted into a ruler marked with exact decades.

Handwriting does not evolve in one straight line.

A scribe can be:

• elderly or young;
• conservative or innovative;
• copying slowly or rapidly;
• trained in one region and working in another;
• imitating an older exemplar;
• shifting between formal and informal modes;
• writing with a damaged pen or on an uneven surface.

A long-lived scribe can carry the habits of one generation into the next. A prestigious script can be deliberately maintained. Several styles can coexist.

The resulting question is not merely, “When was this letter form popular?” It is, “How long is this form attested, in what writing environments, and with what combination of other features?”

The narrower the date, the heavier that burden becomes.

Sinaiticus and Three Contemporary Hands

Codex Sinaiticus provides a nearly perfect warning against overprecision.

The official project identifies three undisputed scribes and allows that a fourth hand may be distinguishable.² Those scribes worked within the same production enterprise. Their writing is therefore contemporary in the practical sense relevant to the codex.

Yet their styles do not look equally old.

Milne and Skeat observed that the hands conventionally called A and B look markedly more archaic than the hand called D. If the codex did not physically demonstrate that the scribes belonged to the same production, D could easily be placed many decades later on stylistic grounds. Nongbri uses that observation to show why graphic difference is not automatically chronological difference.³

This does not refute palaeography. Palaeography helped distinguish the scribes.

It refutes a false inference:

Different graphic appearance necessarily means different generation.

---

1. 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. See also Malcolm Choat, “Dating Papyri: Familiarity, Instinct and Guesswork,” in the same issue, 58–83. ↩
2. Codex Sinaiticus Project, “Production”. The project describes three undisputed scribes and the possibility of a fourth distinct from Tischendorf’s original fourth-scribe proposal. ↩
3. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34, especially the discussion of Milne and Skeat’s observation about scribes A, B, and D. ↩

The codex itself controls the comparison. What might have become a chronological sequence is revealed as scribal variety within one project.

The lesson extends beyond Sinaiticus. A letter can look old because the scribe’s training is conservative. Another can look young because his execution is more fluent or innovative. Style supplies probability, not a mechanical timestamp.

Broad Range and Exact Year

There is a large difference between:

This hand fits Greek biblical majuscule of late antiquity.

---

1. 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. See also Malcolm Choat, “Dating Papyri: Familiarity, Instinct and Guesswork,” in the same issue, 58–83. ↩
2. Codex Sinaiticus Project, “Production”. The project describes three undisputed scribes and the possibility of a fourth distinct from Tischendorf’s original fourth-scribe proposal. ↩
3. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34, especially the discussion of Milne and Skeat’s observation about scribes A, B, and D. ↩

and:

This hand was written in AD 350.

---

1. 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. See also Malcolm Choat, “Dating Papyri: Familiarity, Instinct and Guesswork,” in the same issue, 58–83. ↩
2. Codex Sinaiticus Project, “Production”. The project describes three undisputed scribes and the possibility of a fourth distinct from Tischendorf’s original fourth-scribe proposal. ↩
3. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34, especially the discussion of Milne and Skeat’s observation about scribes A, B, and D. ↩

The first is a comparative placement. The second names a year.

An exact year normally requires evidence of another kind: a reliable colophon, a dated document, an archival association, an historically fixed event incorporated at production, or a scientific result narrow enough to bear that precision. Even then, the evidence must be interpreted.

Palaeography commonly works best with ranges.

The range should become broader when:

• securely dated comparanda are few;
• similar hands persist for a long period;
• the manuscript uses a deliberately formal script;
• the proposed date depends upon a linear developmental scheme;
• different features point in different directions;
• comparison is based on photographs of unequal quality;
• the hand is represented by only a few letters.

It may become narrower when several independent features converge and the comparison set is strong.

“Fourth century” can be a responsible palaeographical conclusion. “About 350” may be reasonable shorthand in one context. But the latter must not be allowed to masquerade as a production statement written by the scribe.

What Codicology Adds

Palaeography studies the writing. Codicology studies the manufactured book.

It examines:

• the papyrus or parchment;
• the size and sequence of leaves;
• the construction of gatherings;
• prickings and ruling;
• page and column design;
• binding evidence;
• repairs, trimming, and replacement leaves;
• inks, pigments, and later reinforcement;
• the relationship among text, corrections, and marginal additions.

Codicology can confirm that apparently different hands belong to one production. It can reveal that a leaf was added later. It can identify disruptions in quire sequence. It can show that a manuscript now bound as one object once had another form.

It can also supply chronological evidence. A format or manufacturing practice may fit one period better than another. A later hand may establish that the book existed by the time of that intervention. A textual apparatus integral to the original production may give a terminus post quem.

But codicology also assigns and infers. A quire pattern is no more a calendar than a letter form.

The most responsible date comes from convergence:

script + book construction + internal boundaries + external records + material evidence.

---

1. 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. See also Malcolm Choat, “Dating Papyri: Familiarity, Instinct and Guesswork,” in the same issue, 58–83. ↩
2. Codex Sinaiticus Project, “Production”. The project describes three undisputed scribes and the possibility of a fourth distinct from Tischendorf’s original fourth-scribe proposal. ↩
3. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34, especially the discussion of Milne and Skeat’s observation about scribes A, B, and D. ↩

What Material Testing Can—and Cannot—Do

Scientific testing can add an independent chronological line.

Radiocarbon analysis measures the decay of carbon in organic material. Applied to parchment, it dates the death of the animal whose skin became the writing surface. Applied to papyrus, it dates the plant material.

It does not directly date:

• the day the scribe wrote;
• the age of the exemplar copied;
• the composition of the text;
• every ink or correction on the page.

Material can be stored before use. Old parchment can be reused. Calibration curves can produce broad or multiple date ranges. Contamination and pretreatment matter. Sampling consumes material, though modern accelerator mass spectrometry requires far less than older methods.

Nongbri argues that Sinaiticus is a particularly promising candidate because the radiocarbon calibration curve for the proposed AD 300–425 range may permit a useful narrowing of the parchment date.⁴ That recommendation is a proposal for testing, not a published carbon date for the codex.

Recent work on other manuscript corpora shows why external calibration matters. A 2025 study combined new radiocarbon dates with machine-assisted handwriting analysis for selected Dead Sea Scrolls and found that some conventional palaeographical placements needed revision.⁵ The lesson is not that an algorithm has replaced the palaeographer. It is that handwriting chronologies improve when anchored to independently dated material.

A Reporting Standard

Every palaeographical claim should answer five questions.

The resulting sentence should sound like this:

The original hand is conventionally assigned to the fourth century on comparative palaeographical and codicological grounds; a narrower mid-century date is disputed because the relevant forms persist into the early fifth century.

---

1. 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. See also Malcolm Choat, “Dating Papyri: Familiarity, Instinct and Guesswork,” in the same issue, 58–83. ↩
2. Codex Sinaiticus Project, “Production”. The project describes three undisputed scribes and the possibility of a fourth distinct from Tischendorf’s original fourth-scribe proposal. ↩
3. Brent Nongbri, “The Date of Codex Sinaiticus,” Journal of Theological Studies 73.2 (2022): 516–34, especially the discussion of Milne and Skeat’s observation about scribes A, B, and D. ↩
4. Nongbri, “Date of Codex Sinaiticus,” 529–34. His proposed broad range is approximately AD 300–425; the value of a carbon result would depend upon sampling, pretreatment, and calibration. ↩
5. Mladen Popović et al., “Dating Ancient Manuscripts Using Radiocarbon and AI-Based Writing Style Analysis,” PLOS ONE 20.6 (2025). The study concerns selected Dead Sea Scrolls, not Vaticanus or Sinaiticus, and is cited here only as a methodological control. ↩

That sentence neither dismisses the discipline nor worships it.

The Strongest Objection

Palaeographers may rightly object that outsiders often discover “uncertainty” and then use it selectively. They accept an early date when it favors their case and demand impossible proof when it does not.

That abuse is real.

No manuscript date in this series will be rejected merely because it is palaeographical. A well-supported fourth-century assignment remains evidence for a fourth-century production range. A lack of day-and-year precision does not make a codex medieval or modern.

The rule must work both ways:

• A conventional date may not be stated as a demonstrated year.
• The uncertainty within a conventional date may not be inflated into permission to choose any century.

The disciplined critic asks how wide the evidence allows the range to be. The partisan asks only which endpoint helps him.

The Confessional Boundary

The authority of Scripture does not come from palaeography.

Palaeography can help the church identify what a manuscript is, distinguish its hands, and estimate when it was copied. Those are ministerial historical services. The method cannot confer inspiration on a Greek translation, depose the authentic Hebrew, or decide the canon by the shape of a letter.

Nor must confessional bibliology fear honest historical method. A truthful argument has no need to call palaeography a fraud. It needs only to prevent a probability from being promoted into revelation.

The next two articles will apply this standard to the great codices themselves. Vaticanus and Sinaiticus deserve direct inspection, careful date assignments, and exact provenance records. They do not deserve dates stronger than their evidence.

Notes and Further Reading

For the physical and codicological evidence of Sinaiticus, see the Codex Sinaiticus Project’s physical description. For a modern codicological and palaeographical study of Vaticanus, see Jesse R. Grenz, The Scribes and Correctors of Codex Vaticanus: A Study on the Codicology, Paleography, and Text of B(03) (PhD diss., University of Cambridge, 2021).

Source-control note: Visible letter forms, layout, and codex construction are direct observations (E1). An apparatus that must postdate its invention can yield a boundary (E2). The date inferred from comparative handwriting and book construction remains an assigned range (E3). A proposed developmental sequence of scripts is a reconstruction (E5). Radiocarbon results would date the organic support within a calibrated range, not the wording’s composition; no such result is claimed here for Vaticanus or Sinaiticus.

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

Next in the series: “Codex Vaticanus: Assigned Date, Documented History.”