The Coherence-Based Genealogical Method represents one of the most technically ambitious developments in modern New Testament textual criticism. Klaus Wachtel has occupied a central place in its development and application, while Tommy Wasserman has contributed prominently to explaining contemporary textual criteria and the method’s significance. CBGM was designed to analyze relationships among manuscript texts in a tradition marked by extensive mixture, commonly called contamination. Its computer-assisted displays calculate agreements, organize witnesses, construct networks, and allow editors to examine the consequences of their textual decisions across an entire New Testament book. These achievements are substantial, and a responsible criticism of CBGM must acknowledge them without minimizing their usefulness. The overstated promise begins when numerical coherence is presented as though it supplies an objective genealogy capable of replacing the traditional documentary evaluation of manuscripts. The method does not independently discover the original reading because its genealogical structures depend upon prior human judgments concerning which readings are earlier and which are later. Those judgments are frequently determined by internal considerations before the computer calculates the relationships that later appear to support them. CBGM therefore remains a sophisticated analytical instrument rather than an independent path to the original wording of the New Testament.

The Distinct Contributions of Wasserman and Wachtel

Wasserman and Wachtel should not be treated as though they performed identical roles in the development of CBGM. Wachtel worked directly within the Münster editorial program and participated in the method’s technical and editorial formation, including its application to the Editio Critica Maior. The computer application itself is particularly associated with Klaus Wachtel and Volker Krüger, while Gerd Mink formulated the major theoretical framework. Wasserman’s importance lies more in his careful presentation of textual criteria, his engagement with the method, and his ability to explain its value to a wider scholarly audience. His discussions help readers understand why contemporary critics seek a more exact account of manuscript relationships than the older text-type classifications provided. Wachtel’s writings advance the stronger methodological claim that genealogical coherence can redefine the nature of external evidence. That claim deserves close scrutiny because agreement statistics do not contain historical direction within themselves. Two witnesses can agree extensively because one descended from the other, because both descended from a lost exemplar, or because their texts independently preserved the same earlier readings. The distinction between advocacy, explanation, technical development, and actual proof must remain clear throughout any assessment of the method.

The Genuine Problem CBGM Was Designed to Address

The manuscript tradition of the New Testament cannot be represented adequately by a simple family tree in which every manuscript descends from one known exemplar. Scribes sometimes corrected their primary copies from another manuscript, adopted marginal readings, consulted parallel passages, or incorporated a familiar form from memory. A single manuscript can therefore contain readings associated with several streams of transmission. Older classifications such as Alexandrian, Western, Byzantine, and Caesarean remain useful descriptions of broad textual tendencies, but they do not account for every local relationship. CBGM responds to this complexity by analyzing the text within thousands of individual variation units rather than assigning every manuscript permanently to one family. It calculates which witnesses agree, evaluates the direction of readings, and identifies texts that function as potential ancestors of other texts. This procedure gives editors a clearer view of mixture and prevents the simplistic claim that every witness possesses one unmixed textual identity. It also permits a manuscript’s relationships to change from one New Testament book to another, which agrees with the documentary evidence. The problem addressed by CBGM is therefore real, but solving the problem of classification does not automatically solve the different problem of identifying the original reading.

Pre-Genealogical Coherence and Its Proper Value

Pre-genealogical coherence is the most objective component of the method because it measures the degree of agreement between two manuscript texts. The computer compares their readings across the selected variation units and produces a percentage representing their shared textual content. If two witnesses agree in a very high proportion of the places where variation occurs, they are demonstrably close in textual form. This information is valuable because it replaces vague impressions with measurable data. It can reveal relationships overlooked by earlier scholars who had access to incomplete collations or only selected test passages. It can also show that a late manuscript preserves a text closely related to an earlier witness, even though the physical manuscript itself was copied centuries later. The calculation does not establish which witness has the earlier reading in any individual variation unit. It also does not establish that one surviving manuscript was copied from the other, since both can reflect a lost common source. Pre-genealogical coherence should therefore be received as a precise description of agreement, not as proof of historical descent.

Genealogical Coherence and the Entry of Editorial Judgment

Genealogical coherence moves beyond counting agreements and attempts to determine the direction in which readings developed. Editors construct local stemmata by deciding that one reading stands prior to another because the latter can be explained as a scribal alteration of the former. This decision can involve transcriptional probability, authorial vocabulary, grammatical difficulty, contextual suitability, harmonization, expansion, omission, or another internal consideration. Once these directional judgments have been entered, the computer aggregates them and identifies which witnesses contain more prior readings relative to other witnesses. The resulting diagrams can look more objective than the judgments that produced them. An arrow on a computer-generated graph possesses no greater certainty than the editorial decision that assigned direction to the readings behind it. When an editor changes the local stemma, the genealogical relationship among witnesses also changes. Genealogical coherence is therefore not raw manuscript evidence but processed evidence shaped by prior interpretation. The method organizes editorial judgments with remarkable efficiency, but organization does not transform judgments into historical facts.

The Initial Text as a Hypothetical Reconstruction

The CBGM uses the term “initial text” for the reading placed at the head of the tradition in each variation unit. The collection of these preferred readings forms the text represented by the Editio Critica Maior for the New Testament book under examination. This initial text is not an extant manuscript, and no surviving codex contains all its readings. It is a hypothetical reconstruction composed of thousands of individual editorial decisions. The designation can easily be misunderstood because “initial” sounds like a direct identification of the apostolic autograph. In practice, its authority rises or falls with the documentary and internal evidence supporting each reading. When the initial text agrees with early papyri and Codex Vaticanus, it rests upon powerful external testimony. When it contains a reading absent from all known Greek manuscripts, its status depends almost entirely upon conjectural direction within a local stemma. The label “initial text” must never shield a reading from renewed examination of the actual witnesses.

The Risk of Circular Confirmation

A serious methodological problem arises when editorial decisions generate the genealogical structure and that structure is then invoked to confirm the editorial decisions. The editors first decide which reading is prior and which readings are posterior. Those decisions determine how the computer ranks potential ancestors and displays textual flow. The resulting coherence is then examined as evidence that the initial decision fits the broader transmission history. Such feedback can be useful because an implausible global pattern can expose a weak local judgment. It becomes circular when the coherence display is treated as independent confirmation rather than as the calculated consequence of the judgments entered into the system. A reading judged original on internal grounds can appear genealogically strong because the system was instructed to treat it as prior. This does not prove that the editors acted carelessly, since experienced scholars can make well-reasoned decisions. It proves that the result is not independent evidence in the same sense as an early papyrus, a dated codex, or a securely translated ancient version.

Why the Computer Does Not Decide the Text

Descriptions of CBGM sometimes create the impression that computers have brought mathematical certainty to textual criticism. The computer performs immense calculations that no editorial committee could carry out reliably by hand. It counts agreements, compares witnesses, searches possible relationships, and displays complicated networks in an accessible form. It does not determine which Greek expression best accounts for the origin of competing expressions. A human editor defines the variation unit, distinguishes readings from subreadings, decides whether two forms are meaningfully different, and assigns genealogical direction. A human editor also chooses the initial reading and decides when a local stemma requires revision. The program then computes the consequences of those choices. The distinction matters because computational precision can coexist with uncertainty in the premises supplied to the computation. CBGM increases the speed, scope, and consistency of textual analysis, but it does not remove the need for disciplined human judgment.

The Redefinition of External Evidence

Wachtel’s proposal to treat coherence as a redefined form of external evidence contains both insight and danger. Traditional external evidence includes the date of a manuscript, its geographical origin, its demonstrated textual quality, and the distribution of a reading among independent witnesses. Pre-genealogical coherence is external in the limited sense that it records observable agreements between texts rather than arguing directly from grammar or authorial style. Genealogical coherence differs because it depends on prior judgments about the direction of readings. The percentage of agreement can be objective while the interpretation of that agreement remains subjective. The division of one variation unit into several units can alter the total number of agreements and disagreements. The selection of witnesses and the handling of lacunae can also affect the resulting calculations. Calling genealogical coherence “external evidence” must not obscure its dependence upon internally constructed local stemmata. A new name does not give processed evidence the same independent historical value as an early manuscript.

The Reduced Role of Date and Provenance

One of CBGM’s central weaknesses is that the age and provenance of manuscripts do not directly control its genealogical calculations. A tenth-century minuscule can rank as a potential ancestor of a fourth-century codex because the term “ancestor” refers to textual content rather than to the surviving physical object. Defenders correctly explain that the late minuscule can preserve a text derived from an exemplar older than the fourth-century codex. That possibility is real and must not be dismissed merely because the surviving witness is late. Nevertheless, possibility cannot erase chronological probability. A reading preserved in P75 from 175–225 C.E. and Codex Vaticanus from 300–330 C.E. possesses an early documentary foundation that a reading appearing only in medieval manuscripts does not possess. The medieval reading can be original, but it bears the burden of explaining why the earlier witnesses uniformly lack it. A method that places chronological data outside its principal calculations risks flattening major differences in documentary value. Date does not decide every reading, but it remains indispensable evidence concerning how far back a textual form can be securely traced.

Potential Ancestors Are Not Historical Exemplars

The phrase “potential ancestor” is technical and must not be understood as identifying the actual manuscript used by a scribe. A potential ancestor is a witness whose text contains a sufficient number of readings judged prior to corresponding readings in another witness. The surviving manuscript can be younger than the supposed descendant because CBGM is comparing textual states rather than physical codices. The label therefore means that the text represented by one witness can stand earlier in the calculated flow than the text represented by another witness. It does not prove direct copying, physical contact, common location, or access to the same copying center. Numerous lost manuscripts stood between the autographs and the surviving witnesses, and most actual exemplars have disappeared. A late witness can preserve an ancient reading while also containing many medieval alterations. A fourth-century witness can contain an occasional secondary reading without losing its general documentary superiority. Potential ancestry is consequently a model of textual relationship, not a recovered chain of historical custody.

Contamination and the Limits of Reconstruction

CBGM treats contamination more realistically than methods that force every witness into one exclusive family. A scribe who consulted two exemplars created a text containing readings from more than one line of transmission. A later corrector could add another layer by changing selected words according to a different manuscript. Marginal annotations could enter the main text in a later copy, and harmonization could introduce wording from a parallel Gospel account. These documented practices explain why manuscripts frequently resist simple classification. CBGM maps the resulting mixture by identifying several potential ancestors for one witness. The difficulty is that the method normally lacks the actual lost exemplars that produced the mixture. It can identify combinations of extant texts that account for the readings, but those combinations need not correspond to the manuscripts historically consulted by the scribe. The model can reproduce a logically sufficient ancestry without recovering the real copying event.

The Instability of Optimal Substemmata

Optimal substemmata are designed to identify the smallest combination of potential ancestors capable of accounting for the readings of a witness. This feature promises to reduce the overwhelming number of possible relationships generated by a large contaminated tradition. The computational problem is severe because each additional potential ancestor greatly increases the number of combinations that must be tested. The uploaded analysis gives an instructive case involving witness 323, whose proposed optimal substemma included 1739, 617, 307, and 03. After further editorial decisions were incorporated into a revised database, 617 and 307 ceased to rank as ancestors and appeared instead as relatively distant descendants. The earlier optimal substemma therefore required reconstruction even though the physical manuscripts had not changed. What changed was the editorial classification of readings as prior, posterior, or unresolved. This volatility demonstrates that an optimal substemma is provisional rather than a recovered historical pedigree. The computer can calculate the optimal arrangement under current assumptions, but new judgments can produce a substantially different arrangement.

The Problem of Manuscript 81 and Jude

The relationships among manuscripts 81, 01, and 03 in Jude illustrate how CBGM can generate historically puzzling results. Manuscript 81 is a late witness with a mixed text, while Codex Sinaiticus, designated 01, belongs to 330–360 C.E., and Codex Vaticanus, designated 03, belongs to 300–330 C.E. In the CBGM rankings for Jude, 81 can appear as a leading potential ancestor for both 01 and 03. The method does not claim that the surviving medieval codex physically stood on a fourth-century scribe’s desk. It claims that the textual state preserved in 81 contains readings calculated as prior to readings in the earlier codices. Such a relationship requires the existence of an earlier lost manuscript whose text later survived in 81. That explanation is possible, but it remains hypothetical and depends heavily upon the prior-versus-posterior judgments assigned in the relevant variation units. A simpler historical explanation in some cases is that 81 copied or inherited many readings from a textual line represented by 03 while acquiring other readings from additional sources. When the calculated genealogy conflicts with the most natural chronological explanation, the local stemmata require renewed documentary examination.

Jude 5 and the Reading “Jesus”

Jude 5 provides a case in which CBGM analysis supported a reading that also possesses substantial traditional evidence. The principal variation concerns whether Jude wrote that “Jesus,” “the Lord,” or another designation saved a people out of Egypt and afterward destroyed those who did not believe. The reading “Jesus” is difficult because it identifies the prehuman Christ as active in the deliverance from Egypt. That theological difficulty supplies a clear reason why scribes would replace “Jesus” with the more familiar “Lord.” The reading also possesses significant manuscript support, including Codex Vaticanus, and therefore does not rest on internal reasoning alone. Wachtel explained that the coherence of the “Jesus” attestation helped finalize the editorial choice. This case shows CBGM functioning as a useful confirmation of a reading already defensible through external and transcriptional evidence. It does not show that coherence independently discovered an otherwise unsupported original. Jude 5 is strongest when the genealogical display is joined to the early documentary evidence rather than presented as its replacement.

2 Peter 3:10 and the Danger of an Unattested Initial Reading

The textual problem in 2 Peter 3:10 exposes the sharpest danger in allowing internal reconstruction to outrun documentary evidence. The familiar competing readings describe the earth and its works as being “burned up,” “found,” “disappearing,” or expressed through related forms. One CBGM-based reconstruction placed “not be found” at the head of the local stemma. No known Greek manuscript contains the required negative particle in that form. The reconstruction attempts to explain the positive “will be found” as the result of accidentally omitting the negative and then to explain the remaining readings as scribal efforts to clarify the difficult statement. That chain possesses internal ingenuity, but it lacks direct Greek manuscript attestation. A reading absent from every known Greek witness must not be granted the same confidence as a reading preserved by early and independent documents. The case demonstrates that a coherent stemma can be constructed around a conjectural starting point. Second Peter 3:10 therefore marks the boundary where CBGM’s explanatory power becomes an overstated claim to textual recovery.

The Thirty-Four Changes in the General Epistles

The application of CBGM to the General Epistles resulted in thirty-four changes between the relevant texts of Nestle-Aland 27 and Nestle-Aland 28. The number is important because it demonstrates both the method’s influence and the general stability of the Greek New Testament. Thirty-four changes across several New Testament books do not constitute a reconstruction of Christianity or a collapse of the transmitted text. Many changes involve word order, particles, spelling forms, or expressions with little effect on translation. Some readings, such as “Jesus” in Jude 5, possess strong documentary and internal support apart from CBGM. Other decisions, such as the reconstruction in 2 Peter 3:10, reveal the risks of placing an unattested reading in the initial text. The results are therefore mixed and must be assessed variation by variation. No numerical total can prove that the method is consistently superior to a documentary evaluation. The stability visible across the General Epistles also confirms that the manuscript tradition has preserved the text with a high degree of fidelity.

The Treatment of the Byzantine Tradition

CBGM has helped expose the inadequacy of treating the Byzantine tradition as one perfectly uniform text. Even manuscripts selected as strongly Byzantine do not agree in every variation unit. Some Byzantine witnesses support the initial reading of the ECM, while others preserve competing Byzantine forms. This evidence agrees with the historical fact that the Byzantine tradition developed through copying, correction, standardization, and regional transmission rather than appearing as a single complete edition. CBGM can measure these internal differences more precisely than older generalizations. It does not prove that the numerical majority of medieval manuscripts preserves the original reading in every place. Nor does it prove that every Byzantine reading is secondary merely because the surviving witnesses are late. A Byzantine reading supported by early versions, patristic citations, or an early Greek witness deserves full consideration. Documentary criticism evaluates Byzantine testimony seriously without allowing late numerical dominance to outweigh strong early evidence.

The Continuing Priority of the Alexandrian Witnesses

The early Alexandrian witnesses remain indispensable because their textual forms are documented close to the period of the autographs. P75, dated 175–225 C.E., shows a striking textual affinity with Codex Vaticanus, copied in 300–330 C.E. Their agreement demonstrates that the text represented by Vaticanus did not originate in the fourth century. P66, dated 125–150 C.E., preserves an early form of the Gospel of John that frequently supports the Alexandrian tradition while also displaying the individual work of its scribe and correctors. P46, dated 100–150 C.E., carries the documentary record of the Pauline corpus deep into the second century. These papyri establish chronological anchors that coherence calculations alone cannot provide. Their readings are not automatically original, since every manuscript contains scribal errors. Their early date, broad agreement, and demonstrable textual quality nevertheless give them superior evidentiary weight. CBGM serves textual restoration best when its graphs remain controlled by these early documentary witnesses.

Why Codex Vaticanus Remains an Anchor

Codex Vaticanus does not possess authority because of ecclesiastical tradition, denominational preference, or an assumption of miraculous preservation. Its importance rests upon its early date, disciplined scribal character, textual restraint, and repeated agreement with earlier papyri. The manuscript preserves an Alexandrian text that often contains the reading best capable of explaining later expansions, harmonizations, and grammatical improvements. CBGM itself has confirmed that Vaticanus agrees with the reconstructed initial text more frequently than any other major witness in the examined corpus. That result is significant because the initial text was not simply copied from Vaticanus. It shows that extensive internal evaluation repeatedly returned to readings already preserved in this early codex. The irony is that a method designed partly to reduce reliance on preferred manuscripts has confirmed the exceptional quality of Vaticanus. This confirmation does not make Vaticanus infallible, and individual readings must still be examined. It does show that traditional documentary confidence in Vaticanus was grounded in manuscript reality rather than nineteenth-century prejudice.

Internal Evidence as a Subordinate Aid

Internal evidence remains necessary because manuscripts sometimes divide and no external witness is error-free. Transcriptional probability asks which reading most naturally gave rise to the others through known scribal behavior. Intrinsic probability asks which reading best fits the author’s vocabulary, grammar, context, and argument. These questions can clarify a difficult unit when the documentary evidence is closely balanced. They become dangerous when an editor constructs a preferred reading that lacks meaningful manuscript support. The maxim that the harder reading is preferable is not a mechanical law because scribes also produced accidental nonsense, awkward grammar, and meaningless omissions. The shorter reading is not always original because scribes omitted material through homoeoteleuton, visual skipping, or other copying errors. Internal principles therefore require confirmation from actual transmissional patterns. CBGM gives internal judgments a systematic framework, but it does not authorize them to become masters over the documents.

The Scriptural Responsibility to Establish the Wording

New Testament textual criticism concerns the wording of writings identified as inspired Scripture. Second Timothy 3:16 states that all Scripture is inspired by God and beneficial for teaching, reproof, correction, and training in righteousness. Second Peter 1:20-21 explains that prophecy did not originate from human impulse but that men spoke from God as they were carried along by the Holy Spirit. These statements concern the words originally written under inspiration, not every copying error introduced during later transmission. Proverbs 30:5-6 warns against adding to the words of God, thereby establishing the seriousness of distinguishing the original wording from later additions. Revelation 22:18-19 likewise condemns deliberate addition to or subtraction from the prophetic message of that book. These passages do not prescribe a modern critical method or identify one manuscript tradition as doctrinally authoritative. They do establish the responsibility to handle the transmitted text carefully, honestly, and without forcing the evidence to serve a preferred theory. A method that reconstructs an unattested reading must therefore bear a greater burden than a reading firmly grounded in early documentary testimony.

Textual Certainty and the Stability of the Tradition

Criticism of CBGM must not be confused with doubt about the recoverability of the New Testament text. The existence of variants results from the abundance of manuscripts and the normal realities of hand copying. Most variants involve spelling, word order, minor grammatical forms, accidental omissions, or additions easily identified through comparison. The large manuscript tradition allows errors in one witness to be corrected by readings preserved in others. Early papyri provide direct evidence that the central Alexandrian form existed long before the great fourth-century codices. Ancient versions and patristic citations provide additional testimony where their wording can be identified with sufficient precision. CBGM adds another layer of analysis by measuring relationships across a much larger body of collated data. Its limitations concern the degree of certainty assigned to particular genealogical constructions, not the general stability of the New Testament. The text can be restored with confidence because the documentary tradition is abundant, early, diverse, and substantially coherent.

The Proper Scope of CBGM in Textual Decisions

CBGM should be used as an analytical assistant that tests decisions against patterns found across the manuscript tradition. Its pre-genealogical calculations accurately identify witnesses that share closely related texts. Its local stemmata force editors to state how they believe one reading developed from another. Its global coherence displays can reveal where a local decision produces an implausible pattern across an entire book. Its treatment of contamination corrects simplistic family trees and permits a more realistic description of mixed witnesses. Its databases also make later manuscripts available for examination rather than excluding them merely because of their date. The method exceeds its proper scope when potential ancestors are treated as proven historical exemplars or when coherence is allowed to outweigh decisive early evidence. It also exceeds its scope when an internally attractive but unattested reconstruction is printed with the confidence appropriate to a well-supported manuscript reading. The sound procedure begins with the earliest and best documentary witnesses, evaluates the actual readings they contain, uses internal evidence to explain their transmission, and employs CBGM to test rather than dictate the resulting judgment.

You May Also Enjoy

How the Loss of the Original New Testament Manuscripts Has Not Created Textual Uncertainty

About the Author

CLICK LINKED IMAGE TO VISIT ONLINE STORE

CLICK TO SCROLL THROUGH OUR BOOKS