What Is a Textual Canon? From Maxim to Verdict

Article 3 of 27 in the series “When the Canons Collide.”

The word canon sounds firmer than most textual canons are. In this setting it does not name the inspired canon of books. It names a maxim such as prefer the shorter reading, prefer the harder reading, or prefer the reading that best explains the others. Each maxim links a fact about a reading to a claim about its history. The link may be reasonable, but it is not contained in the fact itself. A short reading is short; it is not thereby original. This article sorts the main kinds of canons and identifies the judgments hidden inside their familiar wording.

Controlling question: What kind of rule is a textual canon, and what must be supplied before it can yield a verdict?

From Maxim to Verdict

A textual canon is a bridge between a fact and a conclusion. The fact may be that one reading is shorter, one is grammatically rougher, or one appears in earlier witnesses. The conclusion is that one reading is more likely prior or original. The canon supplies the “because”:

Prefer reading A because it is shorter.

Prefer reading B because it is harder.

Prefer reading C because its witnesses are earlier and more diverse.

The word because can hide several different logical relations. Distinguishing them is essential.

Deductive rules

A deductive rule guarantees its conclusion if its premises are true. “All manuscripts copied from exemplar E after E acquired error x will reproduce x, unless a further change occurs” approaches a conditional deductive statement. But the usual canons are not like this. A shorter reading can be secondary. A harder reading can be a blunder. An old manuscript can preserve a late error.

Statistical generalizations

A statistical canon states that one kind of event occurs more often within a defined population. For example, a controlled study might find that a particular scribe makes more significant omissions than additions among his singular readings. This is valuable. Its force, however, is limited by the population, definitions, and conditions studied. A tendency in six early papyri is not a law for every scribe, century, genre, and variation unit.

Defeasible presumptions

Most canons function as defeasible presumptions: absent countervailing evidence, a feature counts in favor of a reading. “Prefer the harder reading” means roughly: if the alternatives are otherwise comparably attested and if the difficulty is the kind scribes tend to smooth rather than the kind they tend to create by accident, then difficulty gives some reason for priority.

The italicized conditions do most of the work. “Otherwise comparable” requires a judgment about external evidence. “The kind scribes tend to smooth” requires a causal classification. “Some reason” leaves its weight unspecified. The maxim is a prompt to investigate, not a machine that produces a verdict.

Abductive criteria

The master canon—prefer the reading that best explains the origin of the others—is abductive. Abduction compares causal stories and chooses the best explanation. It is indispensable in historical inquiry because the copying event is not available for observation. Yet an abductive conclusion depends on the set of hypotheses considered, background knowledge, judgments of simplicity and plausibility, and the relative weighting of evidence. More than one explanation may fit the surviving effects.

Jurisdictional claims

Finally, a textual decision becomes jurisdictional when it answers not merely “What history seems likely?” but “What words shall stand as Scripture for the Church?” The latter claim cannot be smuggled into the former. It needs theological premises concerning what Scripture is, how God preserves it, how the Church recognizes it, and what burden of proof applies to disturbing its settled text.

A Taxonomy of the Main Canons

The following table groups familiar criteria by the question they try to answer.

No row is worthless. Each names a real consideration. The problem appears when a consideration is inflated beyond its evidential reach.

“Other Things Equal” Is Rarely Empty—but Rarely Given

Defenders of the canons often answer counterexamples by adding “other things being equal.” This is logically proper. A presumption should operate only when stronger considerations do not defeat it. But in textual criticism the other things are rarely equal, and equality is not observed directly.

Suppose reading A is shorter, has broader numerical support, and fits the author's normal grammar. Reading B is harder, occurs in two early witnesses, and could explain A as a scribal smoothing. There is no common unit in which “three points for A” and “two points for B” can be added. Age, number, difficulty, and style are not commensurable quantities. Their weights depend on a proposed copying history.

The critic might say the early witnesses are unusually reliable and the difficulty is exactly what later scribes would remove. Another critic might say those witnesses share an early error and the broad transmission preserves the sensible original. Both appeal to recognized considerations. Neither has violated the slogan “other things equal,” because each denies that the decisive things are equal.

Thus the phrase does not solve the conflict. It marks the place where adjudication is needed.

The Difference Between a Check and a Source

A useful distinction can prevent many category errors. Some criteria function well as checks on a conclusion without functioning well as independent sources of the conclusion.

Knowledge that a scribe often skips between similar endings can check a proposed omission. If the manuscript loses a clause at precisely such a place, the observation makes accidental omission plausible. But the tendency alone cannot tell us which reading God inspired. It works within a larger identification of the text and transmission.

Likewise, authorial style can expose an unlikely conjecture. Yet if the author's style profile was built from a critical text produced by the same criteria, using that profile to choose contested readings risks feedback. The style test is strongest where the corpus is textually secure and the feature is distinctive. It is weakest where disputed readings themselves create the profile.

The canons are most useful as disciplined questions:

• Could this be an accidental omission?
• Is this kind of expansion documented in this witness?
• Does the reading have early, independent, and continuous attestation?
• Does the immediate syntax constrain the alternatives?
• Is a proposed motive necessary, merely possible, or contradicted by the scribe's behavior?

They become dangerous when converted into hidden warrants:

• It is shorter, therefore original.
• It is awkward, therefore apostolic.
• It is found in an old codex, therefore superior.
• It differs from the parallel, therefore unharmonized.

Canons Do Not Interpret Themselves

Even identifying the property to which a canon applies may require interpretation.

Which reading is “shorter” when one omits a clause but adds a title? Do orthographic expansions count? Is a transposition longer? Which reading is “harder”—harder for a modern reader, an ancient Greek reader, a particular scribe, or a theologian? Is an apparent disagreement between Gospels a difficulty, a recognized citation convention, or a false problem created by modern expectations? Is a witness “early” by the date of its physical artifact, by the age of the reading it may preserve, or by the reconstructed age of its textual ancestry?

Eldon Epp's historical survey of the canons is valuable precisely because it shows how their wording, number, and application have changed. His deliberately qualified length criterion prefers the shorter or longer reading according to circumstances.¹ Peter Gurry's study of Matthew likewise argues against a default preference for brevity and calls for author- and corpus-specific evidence.² James Royse's study of six extensive early papyri found omission to be a substantial scribal tendency in that defined evidence set, challenging the inherited assumption that scribes characteristically lengthened the text.³

These studies do not abolish textual criticism. They make it more empirical. They also confirm the present argument: the old slogans possess no self-applying authority. Their force must be earned in a defined case.

The First Audit Question

Before applying any canon, ask:

What kind of inference is being made, and how far does the evidence actually carry it?

---

1. Eldon Jay Epp, “Traditional ‘Canons’ of New Testament Textual Criticism: Their Value, Validity, and Viability—or Lack Thereof,” in The Textual History of the Greek New Testament: Changing Views in Contemporary Research, ed. Klaus Wachtel and Michael W. Holmes, Text-Critical Studies 8 (Atlanta: Society of Biblical Literature, 2011), 79–127. ↩
2. Peter J. Gurry, “On Not Preferring the Shorter Reading: Matthew as a Test Case,” in Studies on the Intersection of Text, Paratext, and Reception: A Festschrift in Honor of Charles E. Hill, ed. Gregory R. Lanier and J. Nicholas Reid (Leiden: Brill, 2021), 122–141. ↩
3. James R. Royse, Scribal Habits in Early Greek New Testament Papyri, New Testament Tools, Studies and Documents 36 (Leiden: Brill, 2008), especially 705–736 and 897–906. Royse studies P45, P46, P47, P66, P72, and P75; his result should not be universalized beyond the evidence he analyzed. ↩

If the answer is “a modest historical presumption,” then state it modestly. If the answer is “the inspired wording that displaces the Church's received text,” then identify the additional theological premises. Do not allow the second claim to ride unnoticed on the first.

The remainder of this series is an extended application of that audit.

What This Article Establishes

• Canons may be descriptive observations, statistical tendencies, defeasible presumptions, or abductive criteria.
• The phrase other things equal marks the need for adjudication; it does not perform that adjudication.
• A useful check on a proposed reading is not automatically an independent source of the authentic text.

Conclusion

A canon earns its place by making inquiry more exact. It should force the critic to name a mechanism, test an alternative, or expose a hidden assumption. It ceases to help when its name replaces the work it was meant to prompt. The first audit question must therefore be asked every time: what fact does this canon identify, what inference is drawn from it, and what further premises make that inference stronger than its rival?

Previous: Who Is the Court? Ministerial and Magisterial Textual Criticism · Next: R. L. Dabney’s 1871 Indictment of the Critical Canons