How Particular Readings May Be Tested

Revision note: Substantially revised July 20, 2026. The argument has been reordered according to the classical rhetorical movement, compressed where necessary, and given one consolidated Notes and Further Reading section.

“The Old Testament in Hebrew … and the New Testament in Greek … being immediately inspired by God, and, by his singular care and providence, kept pure in all ages, are therefore authentic.” — Westminster Confession of Faith 1.8

If textual criticism must be ministerial rather than magisterial, how should a particular reading actually be tested? That question cannot be answered by repeating that God preserved his Word. Confessional doctrine governs textual judgment, but it does not relieve the church of textual judgment. Manuscripts must still be compared. Copying errors must still be identified. Grammar, context, versions, citations, and transmission history must still be considered. A doctrine of preservation used to excuse careless scholarship is not being honored; it is being used as a shield for incompetence.

Neither can the question be answered by adopting the academy’s method unchanged and adding providence as a devotional conclusion. If the method may declare every extant reading secondary, create canonical wording by conjecture, and treat the church’s text as indefinitely provisional, the method remains magisterial whatever pious language surrounds it.

The confessional alternative is an ordered judgment. Scripture supplies the governing truths and lawful boundaries. Historical evidence supplies real, subordinate testimony. The church receives and compares that testimony as a fallible minister under an infallible Word. No single criterion performs the whole task. “Oldest manuscript,” “most manuscripts,” “hardest reading,” “shortest reading,” “the church used it,” and “it sounds more orthodox” are all inadequate as mechanical rules. Each may point toward something true. None may be enthroned. The proper question is not, “Which slogan settles the reading?” It is, “How may all relevant evidence be weighed within what God has revealed about the identity, preservation, and authority of his Word?”

Begin with the Object Being Examined

Before weighing readings, one must identify the object of the inquiry. The critic is not searching the ruins of transmission for a revelation God once gave but may no longer have preserved. He is comparing fallible witnesses to the Hebrew and Greek Scriptures that God has kept pure in all ages. The authentic text precedes the critic’s work. It is not produced by it.

This does not mean that every manuscript is pure in every detail. It does not mean that every copyist, printer, translator, or editor was providentially protected from error. It does not mean that one surviving artifact contains no misspelling, omission, transposition, or intrusion. Preservation pertains to the Word through the totality of God’s providential transmission, not to the immaculate condition of every material witness.

It also does not mean that a critic’s conclusion about every difficult unit carries the certainty of Scripture itself. The divine text is infallible; the minister’s recognition of a reading is fallible. The church can possess an infallible rule while its members sometimes misunderstand or misidentify a detail of that rule. That distinction is ordinary Protestant epistemology, not skepticism.

Three denials therefore belong at the entrance to the discipline.

First, there is no neutral court above Scripture. The authority of the Word depends wholly upon God, its author, and not upon a scholar, church, or manuscript committee. Scripture is received because it is God’s Word, not made God’s Word by successful argument. Second, there is no private inspiration of editors. “The Spirit told me that this is the reading” is not a textual argument. The Spirit bears witness by and with the Word; he does not furnish modern critics with new revelation that bypasses public evidence.

Third, there is no need for an infallible sixteenth-century printer. Erasmus, Stephanus, Beza, the Elzevirs, the editors of the Bomberg Hebrew Bible, and later collators performed creaturely work. Their usefulness does not depend upon personal impeccability. The question concerns God’s preservation of his corporate gift, not the canonization of an editor. With the object and limits clear, particular readings may be tested under five ordered considerations.

First: The Canonical and Theological Boundary

The first consideration is not a shortcut to a preferred wording. It is the revealed boundary within which every textual conclusion must remain.

Scripture teaches its own divine origin, abiding authority, verbal stability, and covenantal preservation. Man lives by every word that proceeds from God (Matt. 4:4). Not one jot or tittle passes from the law until all is fulfilled (Matt. 5:18). Scripture cannot be broken (John 10:35). God’s words are truth (John 17:17), endure forever (1 Pet. 1:23–25), and remain the church’s sufficient equipment for doctrine and life (2 Tim. 3:15–17). Westminster gathers this biblical teaching by confessing that the inspired Hebrew and Greek Scriptures were “kept pure in all ages” and are therefore authentic.¹

These truths exclude certain conclusions in advance. The authentic reading cannot be one God allowed to vanish from the entire apographic inheritance, leaving the church without it until a modern scholar invented it. The canonical text cannot be constituted by a method that grants autonomous reason the authority to decide whether divine preservation occurred. A proposed history of transmission cannot require us to deny what God says about his care for his Word.

But theology must not be abused. The rule is not, “Choose whichever reading states doctrine most explicitly.” A fuller reading is not automatically original because it contains orthodox language. Scribes could expand as well as omit. Parallel passages could be harmonized. Liturgical formulas could enter a copy. Conversely, an authentic sentence may be difficult, abrupt, or liable to misunderstanding.

The analogy of faith means that Scripture interprets Scripture and cannot finally contradict itself. It does not authorize the critic to manufacture the most doctrinally convenient text. Theology limits impossible conclusions, tests coherence, and orders the evidence; it does not replace evidence. This distinction protects against errors on both sides. The academy game hides theological commitments beneath supposedly neutral canons. An undisciplined traditionalism may invoke theology to avoid proving anything. Ministerial criticism does neither. It confesses its governing doctrine openly and then deals honestly with instances.

Second: Apographic Availability

The next question is elementary but decisive: Is the proposed reading actually extant in the Hebrew or Greek inheritance? A manuscript need not be complete to be a witness. A reading may appear in a continuous-text codex, lectionary, quotation, or other textual artifact whose evidential character must be assessed. Ancient versions may attest the existence of an underlying reading, although retroversion from one language into another is rarely exact. Patristic citations may also provide secondary evidence, though memory, adaptation, and later copying of the father’s own works complicate their use.

The governing point is that ministerial criticism recognizes Scripture from what God preserved; it does not create Scripture where no preserved witness exists. Conjectural emendation may be useful when reconstructing what a damaged secular document probably said. It may even illustrate a possible explanation of a difficult biblical sentence. But a conjecture unattested in the apographic inheritance cannot acquire canonical authority from its cleverness.

This restriction is theological, not anti-historical. If God kept the inspired text pure in all ages, then the authentic wording remained available. The precise location and distribution of that wording may require investigation. It may be preserved against a defective copy, an erroneous printing, or a localized corruption. Yet “perhaps no surviving witness has it” cannot be treated as an innocent possibility, because that possibility contradicts the object the church has confessed.

Availability alone does not prove authenticity. Every extant variant is not thereby inspired. The point is negative and foundational: the genuine reading must belong to the preserved field of evidence. A hypothesis may explain witnesses; it may not become a new verse. John Owen’s treatment of variants reflects this distinction. He acknowledged diverse readings and errors in particular copies, yet argued that the whole Word remained preserved in the copies available to the church. Corruptions were to be corrected by the preserved rule, and allegations against the text had to be handled by particular evidence rather than sweeping conjecture.

Third: External Evidence in Providential Context

Once the actual readings are identified, their witnesses must be examined. External evidence includes the age and character of manuscripts, geographical distribution, genealogical relationship, agreement among independent lines of transmission, ancient versions, patristic citations, lectionaries, and continuity of ecclesiastical use. Age matters. A manuscript copied in the fourth century stands closer in time to the apostles than one copied in the fourteenth. But the age of an artifact is not the age of its reading. A late copy may preserve an ancient exemplar; an early copy may reproduce an early error. “Our oldest surviving manuscript” is a statement about the accidents of survival, not a synonym for “the original text.”²

Number also matters. Broad numerical support can indicate stable and extensive transmission. Yet manuscripts are not independent ballots. Fifty copies descended from a common defective ancestor do not provide fifty independent testimonies. A large surviving majority may reflect the multiplication of one textual line, while losses of manuscripts can distort the apparent proportions. Counting should therefore be joined to genealogy, distribution, and historical context.

Geographical spread can be significant because the same reading in substantially independent regions is less easily explained by a single local corruption. But geographical categories are sometimes reconstructed from limited evidence, and manuscripts can travel. A label such as “Alexandrian” or “Byzantine” does not decide the reading merely by being pronounced.

Versions and fathers likewise have real but subordinate value. A translation may show that a reading existed by a certain time and place, but features of the target language can conceal its precise source. A church father may quote loosely, combine passages, or cite from memory. The surviving manuscripts of his work may themselves be corrupt. Such witnesses must be heard, not romanticized.

Providence does not make historical analysis unnecessary. It tells us what history is: the arena of God’s covenant government, not a closed system in which the survival of Scripture is measured by ordinary literary probabilities alone. External evidence is therefore neither self-interpreting data nor raw material for apologetic decoration. It is providential testimony to be examined truthfully.

Fourth: Internal Evidence as a Servant

Internal evidence asks which reading best fits the author, context, grammar, argument, and likely history of copying. Intrinsic considerations include an author’s vocabulary, syntax, style, theology, and usage elsewhere. A reading that destroys the grammar, interrupts the argument without explanation, or contradicts a clear teaching of Scripture bears a real difficulty. Immediate context, book-level structure, and canonical coherence all matter.

Transcriptional considerations ask how one reading may have given rise to another. A scribe may skip from one similar ending to another, repeat a word, reverse word order, substitute a familiar expression, harmonize a parallel, expand an abbreviation, or place a marginal explanation into the text. Knowledge of such habits can illuminate the genealogy of variants.

Yet popular internal canons are heuristics, not laws.

“Prefer the harder reading” can be helpful because scribes sometimes smoothed difficulties. But scribes also created nonsense, grammatical roughness, and accidental difficulty. Harder cannot mean more likely simply because it is harder. “Prefer the shorter reading” can be helpful because explanations and harmonizations sometimes expanded a text. But scribes omitted words and lines by accident, and some copyists deliberately shortened. Shorter cannot mean earlier by definition.

“Prefer the reading that explains the others” is stronger, but even this requires a defensible account of transmission. A critic must show why the proposed changes were probable in the actual witnesses, not merely tell a plausible story after selecting his preferred text. Circularity is a special danger. If one first declares a narrow group of manuscripts superior, derives scribal habits from differences between that group and the received text, and then uses those inferred habits to prove the superiority of the same group, the conclusion has been built into the method. Internal evidence must be tested across the whole record and must remain corrigible by external evidence.³

Ministerial use of internal evidence is disciplined imagination. It asks responsible historical questions while refusing to let an elegant hypothesis overrule the preserved reality. Probability serves identification; it does not generate revelation.

Fifth: Ecclesiastical Reception as Providential Evidence

The church is not the author of Scripture, but it is the divinely appointed recipient, witness, and keeper of the oracles of God. The history of a reading in the worship, preaching, translation, controversy, and confession of the church is therefore relevant. This is not Rome’s claim that the church’s judgment creates certainty. Calvin argued that the church is founded upon the prophetic and apostolic Word; Scripture does not receive its authority from an ecclesiastical decree. Westminster likewise distinguishes the church’s testimony from the Spirit’s full persuasion by and with the Word.

Nevertheless, a reading broadly received across generations is not merely one archaeological datum among others. If God preserved Scripture in and for his church, stable ecclesiastical reception belongs to the history of that preservation. The ordinary text read publicly, preached, translated, disputed from, and confessed possesses a providential presumption that a recently preferred reconstruction does not possess merely because its manuscript is old. That presumption is not absolute. A printer’s mistake can enter many copies. A translation can obscure the original. A regional church can inherit a defective reading. Reception must therefore be tested alongside the Hebrew and Greek witnesses, not substituted for them.

The notorious 1631 “Wicked Bible” makes the point plain. Its omission of not from “Thou shalt not commit adultery” was a genuine reading in a printed English Bible, but it did not become Scripture through publication. Other copies, the Hebrew text, the context, the church’s reception, and the transparent cause of the error supplied overwhelming correction. The church did not need an inspired printer to know the commandment. It needed faithful comparison with the preserved rule.

More difficult variants will not be solved so easily. Yet the same authority relation holds. Long and broad reception does not make a reading untouchable; it does place a serious burden upon those who would remove it from the church’s Bible. The more extensive the reception and the more disruptive the proposed displacement, the stronger and more comprehensive the evidence must be.

No Mechanical Algorithm

These considerations do not form an equation. One cannot assign “age” five points, “majority” four, “church use” three, and “shorter reading” two. The relevance and weight of evidence vary with the unit. An obvious eyeskip may be established by direct comparison. A disputed word order may turn chiefly upon authorial usage. A longer passage may require extensive analysis of lectionary practice, patristic citation, and transmission across regions. A version may be decisive evidence that a reading existed but unable to determine its exact Greek form.

This absence of an algorithm does not imply relativism. Courts weigh different kinds of testimony without reducing truth to arithmetic. Physicians integrate symptoms, tests, and history without pretending every datum has identical force. Responsible textual judgment likewise seeks a coherent explanation of all relevant evidence.

The judgment should be public, reasoned, and corrigible. Facts must be distinguished from inferences. “Manuscript A reads X” is a fact that can be checked. “X is earlier than Y” is an inference. “A scribe changed X because he opposed a doctrine” is a claim about motive and usually requires much more evidence. Confusing these levels turns scholarship into rhetoric.

The critic should also state degrees of confidence. Acknowledging that one ministerial conclusion is less certain than another does not make the Word uncertain. It identifies the limitation of the judge. The authentic reading remains in the preserved inheritance even when a particular student has not yet weighed it correctly.⁴

A Seven-Question Protocol

The foregoing principles can be summarized as seven questions for any variation unit.

1. What are the actual readings?

Transcribe the alternatives accurately. Identify spelling, word order, addition, omission, substitution, or larger structural difference. Do not debate a paraphrase of the evidence.

2. Where is each reading extant?

List the relevant Hebrew or Greek witnesses and distinguish direct witnesses from versions, fathers, and retroversions. If a proposed wording exists nowhere, identify it honestly as conjecture rather than quietly printing it as Scripture.

3. What is the character of the external support?

Consider date, distribution, independence, genealogical relationships, known habits of witnesses, and the accidents of survival. Neither count nor age should be isolated from history.

4. How has the reading been received by the church?

Examine its presence in ecclesiastical texts, translations, lectionaries, preaching, and controversy. Ask whether its use was broad and continuous or narrow and late. Reception is testimony to providence, not an infallible vote.

5. Which reading best fits author, context, and canon?

Test grammar, vocabulary, argument, parallel usage, and the analogy of faith. Do not confuse the most explicit doctrinal expression with the authentic wording, but do not pretend theology is irrelevant.

6. What concrete transmission history explains the alternatives?

Identify plausible copying causes and test them against actual scribal behavior. Ask whether a favored canon is being used circularly or whether contrary evidence has been represented in its strongest form.

7. Does the conclusion remain ministerial?

Ask whether the decision recognizes an extant reading or creates one; whether it honors Scripture’s preservation or redefines preservation to fit the result; whether it treats the church as keeper or merely as a long deceived audience; and whether the critic states his conclusion with humility appropriate to a servant.

This protocol cannot guarantee that every critic will agree. It can expose where disagreements actually lie. Two scholars may share manuscript facts but differ over genealogy. They may agree on internal probability but differ over the weight of ecclesiastical reception. Or their apparent dispute over one word may reveal a prior dispute over whether total loss is permissible. Naming the level of disagreement is itself an advance in honesty.

The Objection of “Stacking the Deck”

The strongest objection says that this framework is circular. By confessing preservation before examining variants, the confessional critic allegedly guarantees the result. He refuses any evidence that might falsify his doctrine and therefore does not practice historical inquiry at all. The objection would be forceful if the framework dictated every reading in advance or insulated every traditional decision from correction. It does neither. Manuscripts can correct manuscripts. Strong evidence can expose a printing error, a weak editorial choice, or a mistaken account of reception. The ministerial critic is obligated to change his judgment when the preserved evidence proves it wrong.

What the framework refuses is the myth that ultimate commitments can be suspended. The neutralist also approaches the evidence with governing beliefs: that ordinary historical causation sets the range of permissible conclusions, that no divine promise may control the method, and that the possible loss of every extant reading must remain open. Those are not findings after criticism. They are conditions imposed before it.

Every system reasons from an ultimate authority. Scripture, because it is God’s Word, cannot be authenticated by a higher court without ceasing to be ultimate. Its self-attestation is not a vicious circle in which an unsupported human claim proves itself. It is the necessary epistemic form of divine authority, confirmed by the Spirit through the Word and manifested in Scripture’s incomparable excellencies. The confessional method therefore does not refuse evidence. It refuses to interpret evidence as though the God who gave it had said nothing about it. That is not stacking the deck. It is declining to surrender the court before the witnesses are heard.

Conclusion

Particular readings may be tested because God has given the church both a certain Word and public means of ministerial recognition. Scripture supplies the canonical boundary. Apographic availability excludes the creation of revelation by conjecture. External evidence locates readings in actual transmission. Internal evidence examines authorial coherence and scribal history. Ecclesiastical reception bears providential testimony. These considerations are synthesized by fallible ministers who distinguish facts from inferences, state their confidence honestly, and remain corrigible by the preserved rule.

This method neither worships evidence nor fears it. It denies that evidence is neutral, not that evidence is useful. It denies that the church is infallible, not that the church has a history. It denies that editors create Scripture, not that editors must make judgments. It denies that every copy is perfect, not that the Word has been kept pure.

The goal is not to win the academy’s approval by speaking as though God were absent from the transmission of his own speech. Nor is it to protect a favored conclusion from serious examination. The goal is obedient truthfulness: to receive, compare, defend, and publish the words the King has given and preserved. Having established how readings may be tested, the series must now address the apologetic encounter itself. Part Twenty-Three will ask: “How to Reason with Unbelievers and Heretics Without Surrendering the Court.”

Notes and Further Reading