Negative Controls: Difficulties Shared by Received and Critical Texts

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 Control before the Charge

This series is designed to test whether Reformed theology can be reconstructed from the modern critical-text ecosystem. That purpose creates a danger as serious as the danger of minimizing textual differences: every difficult passage could be blamed on the critical text merely because it is difficult.

That inference is invalid.

If the same wording, proposition, or interpretive pressure is present in both the critical and received texts, the difficulty cannot be attributed uniquely to the critical reading. It may remain a genuine exegetical problem. It may expose a careless harmonization. It may require careful distinctions of genre, chronology, agency, or theological respect. But it is not evidence that a modern editorial decision created the problem.

This is the function of a negative control. The language is borrowed from disciplined comparative inquiry, but the principle is elementary: before claiming that a difference caused an effect, ask whether the effect also appears where the alleged cause is absent. If it does, the causal charge has failed in its present form.

The control is especially necessary in a polemical field. A defender of the received text can exaggerate by assigning every biblical difficulty to modern criticism. A defender of the critical text can evade by treating every variant as though it has no greater consequence than a difficulty shared by all texts. The same control corrects both errors. It prevents false positives against the critical text, and it prevents a real textual delta from being dissolved into a generic appeal to “difficult passages.”

The governing rule is therefore:

A difficulty may be charged as a distinctive consequence of the critical text only when the relevant difficulty, degree of difficulty, or required repair differs materially from the received-text control.

The italicized qualifications matter. A problem need not be wholly absent from the received text in order for the critical reading to aggravate it. The critical reading may remove a local qualification, sharpen an apparent conflict, weaken the most direct harmonizing datum, or force a doctrinal conclusion to depend upon a passage elsewhere. Negative controls do not reduce every case to “same problem” or “different problem.” They measure the difference.

The Westminster Reason for Using Controls

Westminster theology does not promise that every passage is equally plain. The Confession expressly says that “all things in Scripture are not alike plain in themselves, nor alike clear unto all.”¹ It also directs the interpreter to search less-clear places by those that speak more clearly, because “the infallible rule of interpretation of Scripture is the Scripture itself.”²

Those claims establish two relevant principles.

First, the existence of difficulty does not by itself establish textual corruption, doctrinal failure, or canonical contradiction. A hard genealogy, condensed chronology, anthropomorphic description, complex relation between divine and human agency, or apparently divergent narrative perspective may require interpretation even when no disputed variant is involved.

Second, a proposed solution must arise from Scripture rather than from an unrecorded theological instinct. “Scripture interprets Scripture” is not a license to import any reconciliation that preserves a preferred system. The clearer passage must actually supply the relevant distinction, and the interpreter must show that the distinction fits the grammar and subject matter of the difficult passage.

Negative controls serve both principles. They stop the series from treating ordinary hermeneutical labor as proof against a textual edition. They also reveal whether a repair applied to the critical reading is a legitimate canonical inference or an accommodation made only because the inherited doctrine is already known.

What Counts as the Control?

The word received cannot be left vague. For the New Testament, this series compares the relevant critical reading with identified editions in the received Greek tradition—normally Stephanus 1550, Beza 1598, and Scrivener 1894 where each is pertinent. For the Old Testament, the control must identify the actual Reformation-era Masoretic edition or tradition under comparison. The Authorized Version and the New King James Version may clarify how received readings were translated, but neither English translation substitutes for the Greek or Hebrew control.

The control must also match the claim. If the allegation concerns a Greek main-text decision, the control is another Greek reading. If it concerns an ESV rendering, the control includes the Greek or Hebrew base, the ESV translation, and the ESV’s textual footnote. If it concerns canonical coherence, the comparison must track every passage needed for the proposed reconciliation. A comparison between the ESV and the KJV alone cannot establish which original-language text created an effect.

Four terms should remain distinct:

Without these four entries, “the critical text creates a contradiction” is not an argument. It is a conclusion awaiting its premises.

Six Classes of Comparative Result

The series will classify control results before issuing a doctrinal verdict.

1. Shared wording, shared difficulty

The relevant wording is materially the same, and the same interpretive problem appears in both corpora. No unique textual charge is available. This is the clearest negative control. If Romans and James must be related carefully on justification regardless of whether one reads the received Greek text or NA28, their relation cannot be presented as a difficulty invented by NA28. The theological work may be demanding, but the textual difference under investigation did not cause it.

2. Different wording, same operative proposition

The wording differs, but both readings yield the same proposition for the doctrinal question under review. The variant may be historically or exegetically important while remaining nonmaterial for that particular reconstruction. This classification must be proposition-specific. A change may be immaterial to the deity of Christ yet material to a claim about who performed an Old Testament act. “No doctrine is affected” is too broad unless every relevant proposition has actually been tested.

3. Shared difficulty, increased pressure in the critical reading

Both texts require interpretation, but the critical reading removes or weakens a datum that made the received-text construction more direct. The control prevents the exaggerated claim that the difficulty is entirely new, while the measured comparison still records a genuine aggravation. This category will often produce Grade C: substantial but incomplete support, rather than Grade F: contradicted or excluded. Grade C does not mean that the exact proposition has been reconstructed.

4. Shared difficulty, reduced pressure in the critical reading

A modern critical reading may alleviate an apparent tension present in the received control. Fairness requires recording that result. The reconstruction test is not a mechanism for counting only losses. If the critical reading supplies a more coherent local or canonical relation, that is evidence in its favor at the level of theological consequence. It does not by itself prove that the reading is original, because originality requires the full textual argument, but the consequence must be admitted.

5. Critical-specific difficulty

The relevant pressure arises only after the critical reading is adopted, or it becomes materially different because the critical text removes, adds, or changes a proposition. This is a candidate for a critical-text-specific finding. Even here, the conclusion must be bounded. The finding may establish loss of a local qualification, not collapse of an entire doctrine. It may create a tension, not a strict contradiction. It may require cumulative derivation, not hidden borrowing.

6. Received-specific difficulty

The received reading may possess a difficulty absent from the critical reading. The series will record it. A method that suppresses such cases cannot credibly demand candor from its opponents. The received-text construction is not strengthened by pretending that every received reading is immediately simple. Its case must rest on the actual evidence, the confessional doctrine of Scripture, and responsible exegesis—not on an artificial claim that difficulty itself is proof of inauthenticity.

Seven Tests for Every Alleged Critical-Text Difficulty

The wording-identity test

Does the same relevant wording occur in the received control? The comparison must isolate the words that generate the alleged effect, not merely note that the verses are generally similar.

The proposition-identity test

Do different wordings still assert the same proposition in the respect relevant to the doctrine? Lexical variation is not automatically propositional variation.

The interpretive-dependency test

Does the difficulty depend on a disputed interpretation rather than on the textual reading itself? If so, the series must compare interpretations within each textual base before assigning the effect to textual criticism.

The severity-delta test

Is the problem equally difficult in both texts, or does one reading materially intensify it? The result may be a difference of degree rather than a binary difference.

The repair-parity test

Can the same kind of harmonization resolve both readings? If both require the same grammatical or canonical distinction, the repair is shared. If the critical reading requires an additional conjecture, translation departure, or imported premise, the asymmetry must be recorded.

The proof-function test

Even if the general doctrine survives in both texts, does the disputed passage perform the same evidentiary function? A passage may move from principal proof to illustration, or from explicit statement to an inference supported elsewhere.

The corpus-attribution test

Is the alleged problem produced by NA28, UBS6, BHS, BHQ, the ESV, or an interpreter’s synthesis? “The critical text” is not one undifferentiated agent. Attribution must stop at the layer that actually supplies the wording.

Worked Controls

Romans 3:28 and James 2:24 have long required theological discrimination. Paul denies justification by works of the law and asserts justification by faith apart from such works. James denies that a barren claim to faith can justify and speaks of a person being justified by works and not by faith alone. The Reformed reconciliation distinguishes the ground and instrument of justification before God from the evidential vindication of living faith, attends to the polemical targets of the two writers, and refuses to treat a dead assent as saving faith.

Whatever judgment is made about that synthesis, the basic interpretive pressure is not created by a modern critical reading. The key propositions stand in both the received and critical Greek corpora. Therefore:

• the presence of the difficulty is a negative control;
• the validity of the Reformed reconciliation still requires exegesis;
• failure of a particular harmonization would be a theological or interpretive failure, not yet a critical-text-specific failure;
• a textual variant elsewhere could still affect one element of the larger doctrine, but it must be shown rather than presumed.

This example exposes a recurring confusion. “The doctrine is based on many passages” is a correct warning against single-verse reductionism. Lane Keister made that point explicitly in answering an earlier Confessional Bibliology argument.³ But the warning does not end the reconstruction test. It begins it. Once all the passages are gathered, the question becomes whether the designated corpus itself warrants the distinctions that produce the confessional proposition.

Negative Control Two: Divine Sovereignty and Human Responsibility

Acts 2:23 attributes the crucifixion of Christ both to God’s definite plan and foreknowledge and to the lawless agency of those who killed him. Similar conjunctions occur throughout the canon. Reformed theology treats divine ordination and creaturely responsibility as compatible without making God the author of sin, a formulation expressed with precision in Westminster Confession 3.1 and 5.4.⁴

The conceptual labor is present in both received and critical corpora. One must distinguish primary and secondary causation, holy decree and sinful intention, certainty and coercion. Because that labor remains even where no material variant is involved, it is a negative control against the argument that every difficult concurrence of divine and human agency is evidence of critical-text incoherence.

The control does not prove the Westminster formulation. It establishes only that the difficulty is shared. The later audit must still ask whether the relevant corpus supplies the precise exclusions—God is not the author or approver of sin; creaturely liberty and secondary causes remain real—that the Confession states.

Where the Control Does Not Eliminate the Difference: Matthew 5:22

Matthew 5:22 supplies a useful contrast. The received reading includes εἰκῇ, commonly rendered “without a cause,” while NA28 omits it from the main text and the ESV accordingly reads, “everyone who is angry with his brother.” Here the local wording differs. Yet the critical-text New Testament also contains Mark 3:5, where Christ looks around “with anger,” and Ephesians 4:26, “Be angry and do not sin.” Those passages prevent the immediate syllogism that the NA28 form of Matthew 5:22 condemns every anger without distinction and therefore makes Christ a sinner. The canonical data supply a potential distinction between sinful and righteous anger.

What, then, has the control shown?

It has not erased the variant. The received reading states a qualification locally that the NA28 reading does not. It has not proved that the omission is harmless. The critical-text construction must obtain the qualification cumulatively rather than directly from Matthew. But the control has prevented an inflated C4 verdict. The appropriate questions are whether the local explicitness is materially lost, whether the canonical repair is warranted, and whether the doctrine moves from direct to cumulative derivation. That is exactly the kind of measured result this series requires.

Where the Control Reveals a Different Proposition: 2 Peter 3:10

At the end of 2 Peter 3:10, the received Greek tradition represented by Scrivener reads κατακαήσεται, “will be burned up.” NA28 prints οὐχ εὑρεθήσεται, “will not be found.” The clauses are not merely two difficult ways of stating the same surface proposition. They present different predicates and therefore require a full textual, translational, and eschatological analysis.

The surrounding chapter supplies shared controls: fire, judgment, dissolution, and the expectation of new heavens and a new earth remain in the broader passage. Those common elements may stabilize part of the doctrine. They do not make the final clauses identical. The later dossier must test whether the changed clause alters the model of destruction, disclosure, continuity, or renewal, and whether the ESV follows or departs from its stated Greek base. Here the negative control narrows rather than dismisses the issue. It identifies which eschatological pressures are shared and which arise from the disputed clause.

Old Testament Controls Require Special Care

The New Testament comparison often places an eclectic critical edition against a received printed Greek tradition. The Old Testament comparison is not structurally identical. BHS is a diplomatic edition based principally on Codex Leningradensis, not an eclectic reconstruction that replaces its running text whenever an editor prefers another witness. Its apparatus may suggest alternatives while the main text remains the manuscript’s Masoretic reading. BHQ likewise presents a diplomatic base while expanding the evidence and evaluation in published fascicles.

Consequently, an Old Testament problem found in BHS may also be present in the Reformation-era Masoretic control. If so, it is not a difficulty caused by a BHS emendation. An English translation may nevertheless depart from the BHS main text, and an apparatus note may encourage that departure. The series must then identify the actual layer responsible:

1. the shared Masoretic wording;
2. a difference between Masoretic editions;
3. an editorial proposal in the apparatus;
4. a translation’s adoption of another witness or conjecture; or
5. an interpretive harmonization.

This is why later Old Testament numerical dossiers cannot begin with the assertion, “the Hebrew critical text contradicts itself.” They must first ask whether the same figures stand in the received Masoretic control. If they do, the contradiction charge cannot be a distinctive indictment of modern Hebrew criticism. The remaining questions—scribal transmission, inspired autographs, legitimate harmonization, and confessional preservation—are serious, but they are different questions.

How Negative Controls Affect the Verdict Scales

Negative controls interact with both series rubrics.

Derivation grades

A shared difficulty may still leave a doctrine explicitly established (A) or necessarily entailed (B). It may leave only substantial but incomplete support (C), expose external-premise dependence (D), or show compatibility or underdetermination (E) in both corpora. What it cannot do, without an additional delta, is prove that the critical text uniquely weakened or contradicted the doctrine.

A critical reading may deserve C when substantial canonical support survives but an indispensable premise is not necessarily entailed. It may deserve D when the missing premise cannot be derived from the selected corpus and is supplied from outside it. The control determines whether the loss is genuinely asymmetric; it does not convert a shared evidentiary gap into successful reconstruction.

That distinction is essential under the closure test. If both the critical and received corpora leave the same rival model open, neither corpus earns A or B merely because the weakness is shared. The proper finding is twofold: no critical-specific causal delta, and the exact proposition remains unproved in the declared corpus or corpora. Negative control governs attribution, not entailment.

Coherence codes

A shared C1 tension is not a critical-specific defect. A shared C2 repair may reveal a larger textual or translation problem, but attribution must remain shared unless the repairs differ. A C3 finding requires evidence that the selected critical ecosystem repeatedly changes repair principles in ways the control does not. A C4 finding requires strict contradiction after every legitimate distinction and control has been tested. The scales therefore record two things at once: the condition of the critical-text reconstruction and the degree to which that condition differs from the received control.

The Strongest Objections

“A problem shared by both texts is still a problem”

Correct. A negative control does not solve the problem. It prevents a false causal attribution. The series may conclude that both textual bases contain an unresolved apparent difficulty, while also concluding that the difficulty cannot be used as evidence that a modern critical decision created it.

“The received text is inspired, so it should not contain difficulties”

Inspiration excludes error, not depth, compression, diverse perspective, difficult grammar, or the need to compare passages. Westminster Confession 1.7 expressly denies that every part is equally plain. Calling every difficulty an error would concede the rationalist premise that revelation must conform to the reader’s preferred degree of immediate simplicity.

“Harmonization itself is special pleading”

Some harmonizations are. Others are ordinary interpretation. The question is whether the proposed distinction is grammatically, historically, and canonically warranted. A negative control does not immunize harmonization; it makes both sides submit their harmonizations to the same standard.

“Controls smuggle in the superiority of the received text”

They do not. A comparator is not a verdict. The control can show that the critical reading is better, worse, or materially equivalent for the proposition under review. Its purpose is to make the comparison falsifiable.

“Theology is not a laboratory science”

True. Scripture is not a chemical treatment, and doctrinal reasoning is not reducible to an experiment. The term negative control is an analogy for disciplined comparison. Its logical force does not depend upon importing a naturalistic philosophy: one may not assign an effect uniquely to X when the same effect occurs without X. Before any article assigns a textual or doctrinal defect specifically to the critical-text ecosystem, its evidence ledger must answer:

1. What exact reading is alleged to generate the difficulty?
2. Which edition prints it?
3. What is the exact received control?
4. Is the same difficulty present there?
5. If present in both, is its severity the same?
6. What repair does each text require?
7. Does the critical reading alter the proof function of the passage?
8. Is the ESV translating the critical main text or departing from it?
9. What is the narrowest conclusion the comparison warrants?

If those questions are unanswered, the allegation remains preliminary. If the effect persists unchanged in the control, the claim of distinctive critical-text causation fails. If the effect changes materially, the series proceeds to measure the delta.

Verdict

Negative controls are mandatory for a fair Westminster reconstruction test. A difficulty shared by the received and critical texts may require serious exegesis, but it cannot be charged as a distinctive product of the critical text. A different critical reading may nevertheless intensify, reduce, or relocate that shared difficulty; therefore, the comparison must test wording, proposition, severity, repair, and proof function rather than merely asking whether a “problem” exists.

This article issues no doctrinal derivation grade. Its methodological verdict is adopted: every later allegation of critical-text incoherence or doctrinal loss must include an identified received-text control and a measured account of the difference. The negative-control result must never substitute for the exact-proposition result. “The received text has the same problem” may defeat a claim of critical-specific causation, but it does not supply a missing premise, close a rival model, or earn a successful reconstruction grade for either side.

This article does not prove that every biblical difficulty is easily harmonized. It does not prove that the received and critical texts are doctrinally equivalent. It does not prove that a shared reading is original, or that a unique reading is secondary. It does not make theological consequence the criterion of textual authenticity. It does not deny that one wording can state a doctrine more directly than another even when both corpora ultimately support it.

It also does not concede that a problem found in a received printed edition belongs to the inspired autographs in the same manner. Questions of transmission, preservation, edition, and original reading remain to be argued. The negative control establishes only what may and may not be attributed uniquely to the critical-text difference under examination.

Notes and Further Reading