Christianity Unearthed · Research Apparatus

The Transmitted Text

Every materially significant variant in the New Testament - what changed, when it changed, and whether the timing tracks the hardening of Christian orthodoxy.

Method: earliest recoverable Greek (NA28 / UBS5, Metzger, Comfort) vs. the later Byzantine / Textus Receptus tradition.
Scope: 22 readings across the four gospels, Acts, the epistles, and Revelation.
Caveat carried throughout: collation can only see from the 3rd century onward. The century before that stays unlit.
How to read this

Two numbers, three clusters, one blind spot

The word forgery barely applies. Almost nothing here is fraud; it is scribal habit - eye-skips, harmonising one gospel to another from memory, liturgical padding, and a smaller set of theologically motivated touch-ups. Only that last set is what you'd call deliberate, and even it isn't deception so much as devotion with a pen.

So the two ratings below measure the answerable things. Materiality (1-10) asks whether the meaning actually moves if the reading changes. Certainty-secondary (1-10) asks how confident the manuscript evidence makes us that the reading is a later addition or alteration rather than original. A marks readings where "added later" versus "deliberately removed later" is genuinely contested, so the certainty figure is held loosely.

The new dimension - timing - sorts every change into one of three date-clusters. The whole point of the exercise is that they don't all line up with the doctrinal narrowing the same way. One cluster tracks it tightly. One tracks it only at second hand. One arrives a thousand years too late to have been part of it at all.

The transmission timeline

Where the changes sit against the narrowing

The events along the top are the milestones of "the narrowing" - the process by which a diverse early movement hardened into a single orthodoxy and a closed canon. The coloured bars beneath are the three clusters of textual change. Read the overlap: the Christological cluster sits squarely inside the age of the controversies; the canon-driven cluster shadows the fixing of the fourfold gospel; the late-Latin cluster floats off entirely to the right, in the world of Jerome's heirs and Erasmus's press.

◆ Schematic - horizontal spacing is indicative, not to scale
Composition → Patristic age → Nicene settlement
c.30-95Gospels & epistles written
c.144Marcion's canon rejected
c.180Irenaeus fixes the four gospels
325Nicaea - vs. Arius
381Constantinople
centuries elided
Latin West → Printed editions
c.405Vulgate (Jerome)
c.380+Comma in Latin
1516Erasmus → Textus Receptus
1611KJV
Cluster A · Christological corruption · c.150-350
Cluster B · Canon-driven harmonisation & completion · c.100-400
Cluster C · Late-Latin & printed codification · c.380-1600
A - tracks the narrowing directly B - tracks it at second hand C - post-dates it entirely
The master apparatus

All twenty-two readings

Sorted by materiality. Tap a cluster to isolate it and watch which rows light up - the highest-materiality changes at the top are mostly blue and gold, not red. That pattern is the headline finding, and it's visible before you read a word of analysis.

Isolate
Reading What changes Carried by Entered Cluster Mat. Cert.
16:9-20Mark The long endingResurrection appearances, Great Commission, baptismal salvation, snakes & poison - appended to a gospel that ended at 16:8 Both (bracketed) & TR c.100-150
in mss 3-5C
B 10 9
7:53-8:11John The woman caught in adulteryA whole floating episode; absent from P66, P75, ℵ, B; first in Greek at Codex Bezae Both (bracketed) & TR trad. 2C
in text 3-5C
B 8 9
23:34aLuke "Father, forgive them"Jesus' prayer for his executioners; omitted by P75, ℵ², B, D*, W - possibly excised on anti-Jewish grounds Both (bracketed) & TR 2C A 8 5
22:43-44Luke The bloody sweat & the angelJesus' agony in Gethsemane; double-bracketed. Anti-docetic addition or apologetic omission Both (bracketed) & TR 2C A 6 5
24:36Matthew "nor the Son"The earliest text says the Son doesn't know the hour; later scribes delete it - an Arian proof-text removed TR/Byz omits spreads 4C+ A 6 7
14:34-351 Cor Women keep silent in churchPresent in all mss but "floats" after v.40 in the Western line - a possible non-Pauline insertion Both (relocated in some) 2C? A 6 4
3:161 Tim "God" vs "who" was manifestὅς → θεός: a single stroke turns "he who" into "God." Visible correction in ℵ and A TR reads "God"; crit. "who" shift 3-4C A 5 8
20:28Acts "with his own blood""church of God" (θεοῦ) vs "of the Lord" - does God purchase the church with his own blood? Mixed; earliest holds the harder reading early A 5 5
1:18John "only-begotten God" vs "Son"Here the earliest reading is the higher Christology; the majority later softens "God" to "Son" TR "Son"; crit. "God" both 2C A 5 6
5:7-81 John The Johannine Comma"The Father, the Word, and the Holy Ghost: and these three are one." The explicit Trinity proof-text TR / KJV only Latin c.380+
Greek 14-16C
C 4 10
8:37Acts The eunuch's confession"I believe that Jesus Christ is the Son of God" - a baptismal formula added in the Western line TR adds 2C (Irenaeus) B 4 9
6:13Matthew The Lord's Prayer doxology"For thine is the kingdom…" - a liturgical close folded into the text from worship TR / KJV adds worship c.100
in text 2-4C
B 3 9
8:1Romans "…who walk not after the flesh"A clause imported from 8:4 that makes "no condemnation" conditional rather than absolute TR adds 2-4C B 3 8
11:2-4Luke The Lord's Prayer (lengthened)Luke's shorter prayer padded out toward Matthew's fuller form TR fuller 2-4C B 3 8
2:14Luke "peace, good will" vs "peace to those he favours"A one-letter difference (εὐδοκίας/εὐδοκία) in the angels' song TR "good will" early B 3 7
1:25Matthew "firstborn" sonAdded from Luke 2:7; touches the perpetual-virginity debate TR adds 2-4C B 3 7
22:19Revelation "book" vs "tree" of lifeErasmus back-translated the verse from Latin - "book of life" exists in essentially no Greek manuscript TR / KJV only 1516 C 2 9
5:3b-4John The angel at BethesdaA folk explanation of the stirred pool, absent from the earliest text TR adds 2-4C B 2 9
13:18Revelation 666 vs 616A genuine early variant - 616 in P115 and Codex Ephraemi; Irenaeus knew both c.180 Most mss 666 2C - 2 7
17:21 · 18:11 · 23:14Matthew Three harmonising additionsVerses lifted wholesale from Mark and Luke TR adds 2-4C B 2 9
1:1Ephesians "in Ephesus"Absent from the earliest witnesses - bears on whether the letter was circular Both early - 2 6
Western textActs (whole) A longer recensionCodex Bezae runs ~8-10% longer than the Alexandrian Acts - the largest text-form split in the NT Western / Bezae c.150-200 B var. 7

Mat. = materiality (does meaning move) · Cert. = confidence the reading is secondary · = add-vs-delete genuinely contested · ℵ = Sinaiticus, B = Vaticanus, D = Bezae, P## = papyri

The timing answer

Do the changes correlate with the narrowing?

Yes - but only one cluster does, and it isn't the famous one. The changes split cleanly by date, and the split is the answer to your question.

Cluster A - tracks it directly

Christological corruption

c. 150-350 CE · the age of the controversies

These are made during the live fights - against adoptionism, docetism, separationism, and early Arianism - and they lean toward the position that would win. "God was manifest" hardens the incarnation; "nor the Son" gets deleted because it armed the Arians; the bloody sweat answers the docetists who denied Jesus could suffer; "forgive them" may have been cut as the church's quarrel with Judaism sharpened. This is Bart Ehrman's Orthodox Corruption of Scripture in miniature, and it correlates with the narrowing tightly - because it is the narrowing, happening at the level of the pen.

BUT: these are the lowest-materiality changes on the sheet. The theology was already secured elsewhere; the scribes were reinforcing, not inventing.
Cluster B - tracks it at second hand

Canon-driven harmonisation & completion

c. 100-400 CE · shadows the closing canon

These don't answer doctrinal controversies; they answer a structural pressure created by the narrowing. Once Irenaeus fixed the gospels at exactly four (c.180), scribes felt the four ought to agree - so they smoothed Luke's prayer toward Matthew's, gave Mark the resurrection ending its silence lacked, and pulled verses across from one gospel to plug gaps in another. The correlation is real but indirect: harmonisation is a downstream effect of canon-closure, not of Christology. This cluster holds the heaviest changes - Mark's ending, the adulteress.

NOTE: the highest-materiality readings on the whole sheet live here, in the cluster that tracks the narrowing only at one remove.
Cluster C - post-dates it entirely

Late-Latin & printed codification

c. 380-1600 CE · centuries after the settlement

The Comma and Erasmus's back-translations are doctrinally aligned with orthodoxy - the Comma is a Trinity formula - but they arrive a millennium after Nicaea. This is not orthodoxy forming; it is orthodoxy already settled and being read back into the Greek, first through the Latin West and then through a printing press under theological pressure. It correlates doctrinally and not at all temporally.

IRONY: the change most people cite as "the church inventing the Trinity in the text" is the one made longest after the Trinity was uncontested.
The verdict

The correlation is real, displaced, and overstated all at once

Strip it to one sentence: the changes that most clearly correlate with the doctrinal narrowing are the ones that matter least, and the changes that matter most correlate with it only sideways. The Christological cluster genuinely tracks the rise of orthodoxy - but it's a set of low-stakes touch-ups to a text whose theology was already fixed by multiple attestation. The high-stakes additions are about a closing canon and a completing narrative, not about winning a Christological argument.

And the deepest point is the one that has shadowed every turn of this analysis: both the narrowing and the corruption peak in the second century - the very century that sits before our manuscript trail begins. We watch the controversies through the church fathers and we watch the text through 3rd-century-and-later copies, and by the time we can read the manuscripts directly, the proto-orthodox text is already largely in place. So the tightest correlation between text and doctrine lives precisely in the window collation cannot light. We infer it; we cannot photograph it.

Keeping it honest

Where serious scholars disagree

The "narrowing" frame is powerful but contested, and an even-handed segment should carry both voices rather than hand a listener one side as settled fact.

The Bauer-Ehrman reading

Early Christianity was plural before it was orthodox; "heresy" was sometimes the original or majority form in a region. As the Roman-centred form won, scribes during the 2nd-3rd-century controversies altered texts to support the emerging position - adoptionist, docetic, separationist, and patripassianist "corrections" are detectable across the tradition. The text bears the fingerprints of the fight.

The pushback (Köstenberger & Kruger)

The proto-orthodox text was largely stable; the theologically motivated variants are a small, often ambiguous subset, and most differences are plainly accidental. Ehrman's own conclusion concedes the decisive point - no major doctrine actually depends on a disputed reading. Orthodoxy did not need the corruptions, which is hard to square with a picture of orthodoxy manufacturing its scripture.

The honest middle: both can be true at once. Scribes demonstrably nudged the text toward orthodoxy in places and the orthodox case never rested on those nudges. The variants are evidence of a community's convictions, not the foundation of them.