The Experiment: Can Reformed Theology Be Rebuilt from Modern Critical Editions?

Revision note: Substantially revised July 20, 2026. The original project brief has been recast as a classical argumentative essay, its repeated controls and appendices consolidated, and its edition record rechecked. The anticipated publication of NA29 is now correctly dated February 28, 2027.

Lane Keister states a sound caution: “The interpretation of a given passage does not equal doctrinal formulation that is usually based on many passages.”¹

The doctrine of the Trinity does not stand or fall with one disputed verse. Neither Christ’s deity nor his resurrection rests upon a solitary reading. Reformed theology gathers the whole counsel of God, compares Scripture with Scripture, and states conclusions whose premises may be distributed across the canon. Removing one familiar proof does not by itself destroy the doctrine formerly proved from it.

Yet this caution does not close the inquiry. It defines the work.

If a doctrine rests upon many passages, those passages can be named. The textual form of each can be fixed by an identified edition. Its contribution can be stated, its inferences tested, and its relation to the other passages shown. The completed construction can then be compared with the exact proposition confessed by the Westminster Standards.

“The doctrine is taught elsewhere” is therefore not a demonstration. It is a proposal to reconstruct the doctrine from the remaining textual materials. This series will perform that reconstruction.

The Question and the Confessional Target

The question is not whether theologians who use the ESV, NA28, UBS6, or BHS can affirm Reformed doctrine. Many plainly do. Nor is it whether most biblical wording remains common to received and modern critical editions. It plainly does. An honest audit must record that broad common inheritance.

The narrower question is this:

Can the doctrinal system confessed in the Westminster Standards be derived from the modern critical-text ecosystem when that ecosystem is treated as the operative textual base, rather than read through a synthesis inherited from the Reformation and post-Reformation church?

---

1. Lane G. Keister, “An Answer to Chris Thomas”, Green Baggins, November 26, 2024, p. 1 in the supplied PDF. ↩

Retention and reconstruction are different. A theologian may receive an older doctrinal conclusion, know how its parts fit together, and then find alternative support after a familiar reading has been removed. The resulting doctrine may be true and the alternative support legitimate. But its orthodoxy does not yet show how much work the new textual base performs independently.

The Westminster Confession, Larger Catechism, and Shorter Catechism provide the target because they state a mature and integrated Reformed system with unusual precision. They define subjects, relations, exclusions, qualifications, and necessary consequences across doctrine, worship, church, sacrament, law, covenant, death, resurrection, and judgment.

The Standards are not inspired. Scripture judges them. Westminster itself makes the Holy Ghost speaking in Scripture the supreme judge of religious controversy. The experiment therefore does not ask whether an infallible confession can be imposed upon a reluctant Bible. It asks whether a known doctrinal construction can be derived from a specified textual corpus.

The choice is especially fitting because Westminster 1.8 makes its own textual claim. The Hebrew Old Testament and Greek New Testament were immediately inspired, were “by his singular care and providence kept pure in all ages,” are therefore authentic, and are the church’s final appeal in controversy.² Owen likewise held that God’s whole Word as first given remained entire in the original languages, while translations possessed authority insofar as they faithfully represented those originals.³

Confessional Bibliology follows that order. The Authorized Version is neither newly inspired nor superior to Hebrew and Greek. The original-language text judges the translation.⁴

Naming the Corpus Instead of Saying “the Critical Text”

There is no single editorial object adequately named by the critical text. The Greek New Testament and Hebrew Old Testament projects differ, and an English translation may depart from the main text it usually follows.

As rechecked July 20, 2026, NA28 remains the current published Nestle-Aland edition. UBS6/GNT6 supplies the newer Greek main text, incorporating completed Editio Critica Maior work in Acts, Mark, and Revelation; that main text is intended for NA29. The German Bible Society now lists NA29 for publication on February 28, 2027.⁵

The Old Testament situation is different. BHS substantially presents Codex Leningradensis diplomatically with a critical apparatus; it does not operate like the eclectic Greek New Testament. The German Bible Society continues to describe BHS as the most recent complete scholarly Hebrew Bible. BHQ is its developing successor, but only about half of its projected twenty volumes have appeared.⁶

The experiment will therefore name each layer:

The ESV is useful because it gives readers an accessible English witness. Its preface identifies BHS as its ordinary Old Testament base and UBS5/NA28 as its New Testament base, while acknowledging departures from those editions in difficult cases.⁷ The ESV is therefore not “NA28 and BHS in English.” Translation adds lexical, syntactical, punctuation, and discourse judgments; sometimes it adopts another underlying reading. Each study must ask separately what the controlling Hebrew or Greek supports and what an ESV reader is likely to receive.

What It Means to Rebuild

Theology is not made by feeding a concordance into a syllogism. Exegesis requires grammar, context, genre, history, intertextuality, and the analogy of Scripture. All legitimate interpretive labor is allowed. What is forbidden is concealing where an indispensable premise came from.

Let the designated corpus be the named original-language main texts within their canonical context. Let the confessional proposition be the exact claim under examination. The proposition is rebuilt only when:

1. its indispensable material premises come from the designated corpus;
2. readings are selected consistently rather than borrowed ad hoc from incompatible layers;
3. grammatical and contextual inferences are warranted;
4. necessary canonical syntheses are demonstrated rather than merely inherited;
5. no excluded received reading or convenient translation supplies a hidden premise;
6. apparent conflicts are resolved by textual distinctions rather than silent repair; and
7. the result reaches the precision of the proposition being tested.

The last requirement matters. “Christ saves” is not yet Westminster’s doctrine of the Mediator. “God chooses” is not yet its doctrine of decree. “Believers obey” is not yet its complete doctrine of the moral law. The experiment must compare like with like.

Four questions must remain distinct:

• Doctrinal survival: Can some orthodox statement still be made from the corpus?
• Doctrinal derivation: Does the corpus itself warrant the proposition?
• Confessional precision: Does it establish every indispensable element of the Westminster claim?
• Canonical coherence: Can all required readings and interpretations stand together under consistent rules?

This distinction prevents opposite errors. The series will not claim that every altered proof text destroys an entire doctrine. It will not treat the continued ability to repeat an orthodox sentence as proof that the sentence has been rebuilt.

The Trinity illustrates the difference. A corpus can establish one God and the deity and personal distinction of the Father, Son, and Holy Ghost while establishing only at least three divine persons. Westminster confesses three and only three. Unless the premises close the class against a hypothetical fourth person, the exact numerical proposition has not yet been derived. Compatibility with orthodoxy is not identical to proof of every element.

Inherited Theological Capital

No reader approaches Scripture with an empty mind. Trinity, covenant of works, effectual calling, and active obedience belong to developed theological speech. The goal cannot be a fiction of traditionless interpretation. The proper control is a premise ledger.

For every necessary assertion, the study will ask whether it is directly stated, necessarily inferred, grammatically supplied, received through an ESV departure, inherited from the Reformed system, or dependent upon a reading excluded from the operative main text.

An inherited distinction is not automatically illegitimate. It may accurately summarize Scripture. But if the distinction is needed to keep a critical reading from yielding another result, the designated corpus must itself warrant it. Otherwise the reconstruction has spent theological capital without identifying its textual income.

This is the danger the experiment tests. A theologian may change the textual foundation while retaining architectural plans and load paths inherited from an earlier textual tradition. The building may look unchanged. Its appearance cannot show that the new foundation supports it independently.

The Strongest Case Must Be Heard

The experiment would be worthless if it attacked careless popular claims. Each unit will construct the strongest responsible critical-text case. It may use every unaffected passage, every legitimate grammatical distinction, the full canonical context, responsible intertextual reasoning, and the best arguments of critical-text scholars and Reformed theologians.

It may not move silently among NA28, UBS6, an ESV departure, an apparatus reading, and the Received Text whenever each produces a desired local result. Neither may this series accuse an opponent of doing so without evidence. The purpose is to discover whether a stable construction exists.

The critical-text case receives every plausible argument. When two interpretations are grammatically possible, both will be examined. When a canonical distinction genuinely resolves a problem, it will be credited. When Westminster’s proposition is fully established, the verdict will say so. A negative cumulative conclusion has force only if affirmative and negative findings are measured by one rule.

Coherence and Grading

Scripture must be compared with Scripture within each Testament and across both. Grammar, genre, covenantal setting, chronology, speaker, referent, respect, and sense must be considered. Difference is not contradiction. A general rule may coexist with a qualified application; progressive revelation may add without denying.

The doctrinal and coherence findings will be reported separately:

Only A or B with C0 or C1 counts as successful exact reconstruction. Grade C may be substantial and orthodox, but it is incomplete. Grades D and E distinguish borrowed completion from mere compatibility. Grade F requires the strongest proof.

The method also needs controls. Variants with no doctrinal effect are negative controls. Difficulties shared by both textual traditions are shared controls. Doctrines built completely from undisputed wording are positive controls. A method incapable of returning a strong affirmative verdict is not an experiment; it is a conclusion-generating machine.

A Falsifiable Thesis and Reproducible Procedure

The provisional thesis is that modern critical editions can restate much inherited Reformed doctrine because most wording is shared, but cannot independently reproduce the whole Westminster system with the same directness, precision, stability, and coherence without repeated borrowing or repair.

That thesis is not the rubric. It can be weakened or falsified. If propositions repeatedly receive A/B derivation and C0/C1 coherence, the critical corpus has passed to that extent, and the finding must be published.

Each doctrinal unit will therefore:

1. establish the historical meaning of Westminster’s language from primary sources;
2. state the exact proposition without dropping a relation, exclusion, quantifier, or necessity claim;
3. identify every indispensable premise and the function of Westminster’s proofs;
4. record the received and critical readings, literal renderings, and ESV witness;
5. build the strongest complete case from the critical corpus;
6. classify each premise as stated, entailed, probable, compatible, absent, or contradicted;
7. identify rival models left open by the premises;
8. remove borrowed received, translational, conjectural, and merely inherited premises;
9. test local exegesis and whole-canon coherence under consistent rules; and
10. state the greatest conclusion warranted before assigning its grade and code.

The same process will be applied across all thirty-three chapters of the Confession with the related catechism questions. Sources, editions, corrections, and changes of verdict will remain visible so readers can reproduce the result.

Objections and Verdict

This opening article does not prove that critical-text advocates are unorthodox, that every variant matters doctrinally, that a lost proof destroys a doctrine, that every apparent tension is contradiction, or that Westminster cannot err. It does not prove the provisional thesis. It defines how that thesis will be tested.

Nor does it permit theology to choose readings merely because they preserve a preferred doctrine. Textual evidence must be weighed honestly. But after an editorial decision is made, its theological consequences may not be hidden behind the assurance that doctrine is usually based on many passages. If many passages bear the load, show the load each bears.

No verdict on the reconstructibility of the Westminster system is warranted yet. The methodological verdict is narrower: Keister’s appeal to many passages is not an answer to the proposed audit. It is the summons to perform it.

Gather the passages. Name the editions. Separate main text, apparatus, translation, and interpretation. Construct the strongest case. Identify every premise. Test the canon for coherence. Compare the result with the precise confessional proposition. Then state no more—and no less—than the evidence proves.

Notes and Further Reading