Why Scribes Omit as Well as Add

Article 9 of 27 in the series “When the Canons Collide.”

A rule favoring shorter readings becomes one-sided if omission is treated as a rare exception. Copyists lose text through repeated beginnings or endings, skipped lines, damaged exemplars, cramped abbreviation, faulty memory, and deliberate pruning. Empirical work on particular early papyri has also found strong omission patterns in the scribes studied. Those findings do not create a universal rule that the longer reading is better. They remove the old privilege by which addition was assumed and omission had to plead for admission.

Controlling question: What evidence can distinguish a secondary omission from a secondary expansion when both are known scribal acts?

How Texts Shrink

Omission is at least as concrete.

Homoeoteleuton and homoioarcton

When two words or lines have similar endings (homoeoteleuton) or beginnings (homoioarcton), the copyist's eye can jump from the first occurrence to the second. Everything between disappears. The broader phenomenon is often called parablepsis, a looking beside or skipping past.

Haplography

One of two adjacent similar letters, syllables, words, or constructions is written only once. What looks like a tidy shorter reading may be a mechanical loss.

Line or page omission

A copyist can skip a line, turn two leaves, lose a marginal insertion, or copy from a damaged exemplar. Large omissions require material as well as psychological explanations.

Deliberate pruning

A scribe or corrector may remove repetition, a perceived contradiction, an apparent theological problem, or material judged secondary. The critic who allows theological addition must also allow theological omission unless evidence rules it out.

Excerpt and lectionary effects

Material can disappear at a reading boundary or in transmission influenced by excerpted use. The direction is not always expansion.

Memory and dictation

A copyist working by memory or hearing may preserve the sense while dropping smaller elements. Shortening need not leave an obvious visual trigger.

The symmetry is important. Piety can add a title or remove an apparent difficulty. Liturgical use can add a doxology or omit context. Harmonization can lengthen by importing words or shorten by replacing a longer unique expression with a compact familiar one.

The Empirical Challenge

For generations, the preference for brevity was often supported by a broad picture of scribes as accumulators. Ernest Colwell and especially James Royse pressed the discipline toward study of identified scribes. Royse analyzed the singular readings and corrections of six extensive early papyri—P45, P46, P47, P66, P72, and P75—and found a marked tendency toward omission in that sample. His work challenged the idea that early scribes characteristically lengthened the text.¹

The right lesson is not “prefer the longer reading.” That would simply reverse an overgeneralization. Royse's result is conditioned by his method: he examines particular papyri, classifies singular readings, distinguishes meaningful additions and omissions, and reasons from those scribes. Questions remain about how singular readings represent total habits and how results transfer to other witnesses. Nevertheless, the study establishes a methodological demand:

Claims about scribal tendencies should be indexed to actual scribes, corpora, mechanisms, and types of variation.

---

1. Royse, Scribal Habits, 705–736, 897–906. See also Peter Malik, P.Beatty III (P47): The Codex, Its Scribe, and Its Text (Leiden: Brill, 2017), for the value of studying one codex and scribe in depth. ↩

Gurry's study of Matthew applies the same discipline at the authorial corpus level. Rather than assume a global preference for shorter readings, he tests proposed examples and emphasizes the circumstances in which omission or addition is likely. Eldon Epp's mature formulation makes the qualification explicit: prefer shorter or longer according to circumstances.

A canon that can point either way according to circumstances is not useless. It is no longer a canon of length. It is a demand to reconstruct the circumstances.

A Formal View of the Problem

Let S be the shorter reading and L the longer. The question is not whether S is short. It is whether the observed evidence E is more probable if S was prior than if L was prior:

Likelihood comparison: P(E given S is prior) divided by P(E given L is prior).

---

1. Royse, Scribal Habits, 705–736, 897–906. See also Peter Malik, P.Beatty III (P47): The Codex, Its Scribe, and Its Text (Leiden: Brill, 2017), for the value of studying one codex and scribe in depth. ↩

To estimate that ratio even informally, the critic needs more than length. If S is prior, what expansion produced L? If L is prior, what omission produced S? Does the letter sequence create an eye-skip? Does the scribe make comparable omissions? Does the longer wording resemble a marginal gloss or parallel? How are the readings distributed? Which proposed history requires fewer unsupported steps?

The answers may make one direction substantially stronger. But length alone does not supply them. The probability belongs to a causal model, not to the number of words.

The Asymmetry of the Traditional Slogan

The shorter-reading canon often gains rhetorical force from an asymmetry:

• Addition is treated as an ordinary scribal tendency.
• Omission is treated as an exception needing a special trigger.

Royse's work shows why that asymmetry cannot simply be assumed. More fundamentally, both directions require a cause. “A scribe expanded” is not an explanation until the proposed source, motive, and form of expansion are specified. A familiar phrase in another Gospel may supply such a source. A marginal note in a witness may supply it. Mere possibility does not.

Similarly, a shorter reading should not be rejected merely because omission is possible. A visual trigger, scribal pattern, or disrupted syntax can make omission a better explanation. Both proposals bear a burden.

Large Passages Expose the Rule’s Limits

At Mark 16:9–20 and John 7:53–8:11, calling one form “longer” and another “shorter” scarcely begins the inquiry. Large-unit variants involve manuscript endings, blank space, multiple placements, lectionary practice, patristic discussion, versional transmission, narrative structure, and long ecclesial use. A canon designed as a local presumption becomes especially weak when the variation has its own reception history.

The longer ending of Mark is not twelve extra words that look like a gloss. The pericope adulterae is not a simple expansion in one stable location. Whatever conclusion one reaches, length cannot substitute for explaining the whole evidence.

This point generalizes. The greater the textual and ecclesial consequence of a proposal, the more inadequate a one-feature presumption becomes.

When Brevity Is Legitimately Probative

Shortness can provide meaningful evidence when several conditions converge:

1. the longer form contains material traceable to a specific parallel, gloss, title, or liturgical formula;
2. there is no plausible visual, material, or contextual cause of omission;
3. the relevant scribe or branch shows a documented tendency to expansion of that kind;
4. the shorter form is independently and continuously attested;
5. the longer form appears at different locations or in unstable forms, suggesting accretion;
6. the shorter form leaves coherent grammar and sense without requiring editorial repair.

Even then, the conclusion is a strong historical inference, not a self-authenticating decree. The criteria work cumulatively because the mechanism has been specified.

The Verdict on Lectio Brevior

The shorter-reading canon is defensible only as a highly qualified presumption. It is strongest as a warning to inspect expansions, weakest as a default vote, and dangerous as a warrant for excision. It cannot rank itself against earlier attestation, harder sense, broader reception, or a clear omission mechanism.

The practical rule should therefore be rewritten:

Do not prefer a reading because it is shorter. Ask whether the longer reading has a more evidenced history of addition than the shorter reading has of loss.

---

1. Royse, Scribal Habits, 705–736, 897–906. See also Peter Malik, P.Beatty III (P47): The Codex, Its Scribe, and Its Text (Leiden: Brill, 2017), for the value of studying one codex and scribe in depth. ↩

That formulation is less memorable. It is also more honest.

What This Article Establishes

• Omission has visual, material, memorial, and intentional mechanisms that can be tested locally.
• Scribal-habit studies must remain indexed to their own scribes, corpora, and classifications.
• The true comparison asks which proposed cause better predicts the whole observed evidence.

Conclusion

Brevity becomes probative only when the longer form is better explained as growth than the shorter form is explained as loss. That is a comparison of histories, not a count of letters. Once the question is stated this way, lectio brevior has become a modest servant: it points to a line of inquiry, accepts correction from measured habits and material causes, and yields when a stronger omission account fits the facts.

Previous: Is the Shorter Reading Better? · Next: Is the Harder Reading Better?