How Textual Critics Actually Work
External evidence, internal evidence, and the question that holds them together
On this page
This is the workshop session. The previous lessons asked why textual criticism matters, how the New Testament was copied, and how printed Greek editions developed. Now we put those pieces to work.
When two readings compete, the controlling question is not, “Which one do I prefer?” It is:
Which reading best explains how the other reading arose?
That question brings two kinds of evidence together. External evidence asks what the surviving witnesses say. Internal evidence asks what a copyist was likely to do and what the author was likely to write. Neither side should operate alone.
1. Start with the master question
Suppose a verse survives in two forms, A and B. Ask the historical question in both directions:
- If A came first, can we give a plausible account of a copyist producing B?
- If B came first, can we give a plausible account of a copyist producing A?
Sometimes both paths are plausible. Often one path fits familiar copying habits while the reverse requires a string of unlikely events. The reading that explains the rise of its rival has the stronger claim.
This test disciplines our instincts. Familiarity is not evidence of origin. Neither is a reading’s usefulness in a theological argument. The church does not honor Scripture by choosing the wording it finds most convenient. It listens to the witnesses and asks what history best explains them.
2. Two kinds of evidence, one decision
| External evidence asks | Internal evidence asks |
|---|---|
| Which witnesses contain each reading? | What change would a copyist be likely to make? |
| How early are those witnesses? | Is there a common copying accident that explains the difference? |
| Are they independent, or are many descended from one stream? | Which wording fits the author’s vocabulary, style, argument, and context? |
| Is the reading distributed across languages and regions? | Does one reading explain why the other would have seemed difficult? |
If the two sides converge, an editor can decide with considerable confidence. If they pull in different directions, the editor must weigh them carefully and explain the judgment publicly. A critical edition’s apparatus exists in part so that other readers can inspect and challenge that work.
3. External evidence: witnesses are weighed
Counting manuscripts is useful, but a raw total can mislead. Imagine a medieval scriptorium making fifty copies from one exemplar. Those fifty copies preserve evidence, but they are not fifty independent witnesses to a first-century reading. They are many descendants of one branch.
Three considerations are especially important.
Date
An early manuscript is closer in time to the composition of the New Testament and has passed through fewer copying generations. That gives it importance, not automatic authority. An old manuscript can preserve a secondary reading, and a later manuscript can preserve an ancient one.
Distribution
A reading present in witnesses that circulated in different languages and communities may reach back behind those surviving witnesses to an earlier common source. Geographic labels must be used cautiously—we often know where a manuscript was found or used more readily than where every ancestor in its textual history was copied—but broad distribution can still matter.
Genealogical relationship
Manuscripts share readings because they stand in complicated relationships to one another. Earlier scholarship often arranged these relationships into tidy “families.” The familiar Alexandrian, Western, and Byzantine labels remain useful descriptions of broad textual clusters, but they are not three sealed boxes with fixed homes and habits. Westcott and Hort themselves used a fourfold scheme, and later evidence has shown a more intricate history.
| Broad label | What it can help us notice | What it must not make us assume |
|---|---|---|
| Alexandrian | Many early papyri and major codices share a comparatively concise form of the text | Every reading in an Alexandrian witness is original |
| Western | Witnesses such as Codex Bezae and some Old Latin texts sometimes preserve freer or expanded forms | Every unusual reading arose in one western location |
| Byzantine | The large medieval Greek tradition preserves the church’s long copying history | A majority automatically settles every earlier reading—or that the tradition has no early readings |
The fair rule is neither “the oldest manuscript always wins” nor “the most manuscripts always win.” Every witness belongs to a history. The task is to reconstruct that history reading by reading.
Why the papyri matter
Early papyri moved the surviving evidence closer to the period of composition. P75, usually dated around the late second or early third century, often agrees closely with the much later Codex Vaticanus in Luke and John. That agreement shows that forms of text known from fourth-century codices can have much earlier roots.
It does not close every case automatically. It does show why early, independent agreement deserves serious weight.
4. Internal evidence, part one: what copyists tended to do
Scholars call this transcriptional probability: which reading is more likely to have produced the other during copying?
| Rule of thumb | Why it can help | Why it is not a law |
|---|---|---|
| The harder reading may be earlier | Copyists often clarified an apparent error or awkward statement | A scribe can also create nonsense by accident |
| The shorter reading may be earlier | Explanations, titles, and familiar phrases tend to grow | A line can disappear through an eye-skip |
| Harmonized wording is suspect | A copyist may draw one Gospel toward a parallel remembered from another | Two authors can independently use the same wording |
| Repeated beginnings or endings can cause loss | The eye can jump from one similar word or phrase to another | Not every omission has the necessary visual conditions |
The Latin maxims lectio difficilior potior (“the more difficult reading is stronger”) and lectio brevior potior (“the shorter reading is stronger”) are compact reminders, not mechanical rules. A badly formed reading does not win merely by being hard, and a missing paragraph does not win merely by being short.
The question remains historical: what is the likeliest path from one surviving form to the other?
5. Internal evidence, part two: what the author tended to write
This side is called intrinsic probability. Each New Testament writer has recurring vocabulary, syntax, narrative habits, and theological concerns.
- Vocabulary and style: Mark repeatedly uses words and constructions that differ from Luke’s more polished prose or John’s recurring language of life, light, truth, love, and abiding.
- Immediate context: A reading that completes an argument or carries a narrative forward may fit better than one that interrupts it.
- The book as a whole: An author can introduce a theme early and return to it later.
- Theology: A reading should be understood within the author’s thought, but theology cannot be used as a shortcut. A copyist also had theological convictions and could make a text more explicit.
Intrinsic probability requires humility. Readers can imagine that the author “must have” said whatever sounds best to them. That is why the author question and the copyist question must remain tethered to the external evidence.
6. Work the method: Mark 1:2
The opening of Mark provides a compact case in which the evidence works together.
| Reading | Familiar English form |
|---|---|
| “Isaiah the prophet” | ESV, NASB, NET, and many other modern translations |
| “the prophets” | KJV and NKJV |
Mark immediately combines language from Malachi 3:1 with Isaiah 40:3. That makes “Isaiah the prophet” look difficult: why name Isaiah when the first line is associated with Malachi?
Step one: examine the witnesses
“Isaiah the prophet” is supported by early Greek codices and by evidence from more than one stream, including Old Latin and Coptic witnesses and early Christian writers. “The prophets” is represented especially in Codex Alexandrinus and the later Byzantine tradition.
The external case favors “Isaiah the prophet”: it is earlier in the surviving Greek record and has broad support.
Step two: picture the copyist
Now ask which change is easier to explain.
If a copyist saw “Isaiah the prophet,” recognized the opening line from Malachi, and worried that Mark appeared mistaken, changing the phrase to “the prophets” would solve the difficulty neatly. It would also preserve the truth that more than one prophetic book contributes to the quotation.
The reverse is harder to explain. If Mark had written the accurate and comfortable phrase “the prophets,” why would a copyist replace it with the more vulnerable “Isaiah the prophet”?
Step three: listen to Mark
Mark 1:1–8 is not a detached quotation exercise. Isaiah’s wilderness proclamation frames John the Baptist and the coming of the Lord. The Malachi line introduces the messenger; Isaiah supplies the dominant horizon. Naming Isaiah can be a deliberate signal about how to read the Gospel’s opening.
The harder wording is not a random mistake. It makes sense within Mark’s purpose.
Step four: ask the master question
If “Isaiah the prophet” came first, the rise of “the prophets” is easy to picture: a devout copyist removed an apparent problem. If “the prophets” came first, the rise of the more difficult wording is difficult to account for.
External and internal evidence therefore point in the same direction. Modern critical editions print “Isaiah the prophet” with high confidence.
Textual criticism has not damaged Mark here. Recovering the difficult reading lets us see more clearly how Mark placed his whole Gospel under Isaiah’s announcement of the Lord coming to his people.
7. When the evidence does not converge
Mark 1:2 is an unusually teachable case. Elsewhere:
- early witnesses divide;
- a later reading has a strong internal claim;
- an apparent omission may be either deliberate abbreviation or accidental eye-skip;
- ancient translations cannot reveal the exact Greek behind them;
- patristic quotations are loose, remembered, or altered in transmission;
- two plausible histories explain the same evidence.
Editors then compare related variants, examine a writer’s habits, study the relationships among witnesses, and publish the reasons for their decision. Consensus sometimes forms. At other places, responsible disagreement remains.
The letter grades printed in UBS editions have described an editorial committee’s confidence in the text it chose. They do not measure the spiritual importance of a verse, and a translation’s footnote does not by itself mean the decision was rated C or D. Some famous disputed passages are omitted from the critical text with very high editorial confidence.
Practice
Use the Mark 1:2 comparison in Ad Fontes NT. The tool compares selected published editions; it is not a full manuscript apparatus. Answer these questions before reading the explanation:
- Which displayed editions read “Isaiah the prophet,” and which read “the prophets”?
- Which form creates the apparent difficulty?
- What change would a conscientious copyist be most tempted to make?
- How does Isaiah 40 shape Mark 1:1–8?
- Can you state the whole decision in two sentences—one from external evidence and one from internal evidence?
Discussion
- Where else do we already weigh witnesses rather than count them?
- Why might a careful, devout copyist be more likely to “fix” Mark 1:2 than a careless one?
- Why is a reading that supports our theology not automatically original?
- What would it mean to treat a critical edition as an accountable scholarly judgment rather than a new inspired text?
- Has working one case changed what you hear in the phrase textual criticism?
Working vocabulary
External evidence. Evidence supplied by manuscripts, ancient translations, and quotations in early Christian writers, considered with their dates and relationships.
Internal evidence. Evidence supplied by copying habits and by an author’s language, context, and argument.
Transcriptional probability. The question of which reading a copyist was more likely to turn into the other.
Intrinsic probability. The question of which reading the author was more likely to write.
Textual cluster. A group of witnesses that share readings and therefore show some historical relationship; a more cautious term than imagining rigid families with fixed boundaries.
Homoeoarcton. An eye-skip caused by two words or phrases beginning alike.
Homoeoteleuton. An eye-skip caused by two words or phrases ending alike.
Sigla. Compact symbols used to identify manuscripts in a critical apparatus.
Before the next session
Read: Chapter 5, “The Ending of Mark.”
Do one thing: Open every Bible available to you at Mark 16. Record exactly what happens after verse 8: uninterrupted text, brackets, a heading, a footnote, the shorter ending, or some combination.
Bring: Your Bible, your observations, and your vocabulary notes.
Think about: The copyists who changed Mark 1:2 may have loved Scripture and wanted to protect it. What is the difference between protecting the Word and listening to it?
A closing word
Thanks be to God for the apostles who wrote, the copyists who preserved what they received, and the evidence that lets later readers recognize even a well-meant correction. The work is not a contest to invent a better Bible. It is the patient work of hearing the witnesses and returning to the source.
“They were examining the Scriptures carefully every day to see if these things were so.”
— Acts 17:11 (NET)
Scripture quotations in this session are from the NET Bible, with the KJV referenced where translations are compared.