“Through the Manuscripts”: Does an Aggregate of Competing Readings Constitute a Preserved Text?

Revision note: Substantially revised July 20, 2026. Repeated proof concerning public possession has been removed, Keister’s express claims and the inferences drawn from his method are more sharply distinguished, and the principal sources are gathered below.

Manuscripts may preserve, transmit, and witness to a text. But a collection of witnesses that gives contradictory testimony is not, merely as a collection, the determinate text to which it bears witness.

Lane Keister’s clearest affirmation of preservation contains a phrase upon which the entire controversy turns. God, he says, preserved the original text of both Testaments “through the manuscripts.”¹ The statement is important, and it should be read in its strongest reasonable sense. Keister is affirming a divine act, a preserved object, and a historical means. He is not treating the New Testament as an ordinary ancient book abandoned to chance.

Yet the phrase does not interpret itself. It may mean that God used the succession of copies as the instrument by which a determinate text remained in the possession of his church. It may also mean that the total manuscript archive preserves a field of competing readings from which scholars must reconstruct the determinate text. Those are not two ways of saying the same thing. In the first, the manuscripts transmit and witness to a text already given, preserved, and publicly received. In the second, the aggregate of manuscripts supplies the evidence from which the text must still be constituted by critical judgment.

The distinction is not an attempt to trap Keister in a preposition. It is required by the nature of a text. A thousand manuscripts can be providentially preserved without the inspired wording being formally identified. Ten competing readings can all survive while only one of them is original. An archive can contain the answer without being the answer.

The question before us is therefore exact: Does “through the manuscripts” identify the means by which God preserved a text, or does it redefine the preserved text as the aggregate of manuscript evidence?

The Grammar of Preservation

Every adequate doctrine of preservation must distinguish four things:

Element: Agent

Confessional function: God preserves by singular care and providence.

Category error to avoid: Treating survival as an unintended accident of history.

Element: Object

Confessional function: The inspired Hebrew and Greek Scriptures—the words or text—are preserved.

Category error to avoid: Substituting manuscripts, variants, or critical data for the text.

Element: Means

Confessional function: Copies, ecclesiastical use, comparison, translation, and transmission serve preservation.

Category error to avoid: Turning the means of transmission into the object transmitted.

Element: Recipient

Confessional function: The church receives, reads, preaches, translates, and appeals to the preserved Scriptures.

Category error to avoid: Making the authentic text inaccessible until modern reconstruction.

The phrase “through the manuscripts” properly names the means. God did not promise the immortality of the autographic parchment. He preserved his Word through copying, public reading, correction, translation, and the multiplied transmission of Scripture among his people. No responsible doctrine of Confessional Bibliology needs to claim that scribes were inspired, that every copy is flawless, or that one physical codex has been miraculously exempted from ordinary human error.

But means cannot replace object. We do not say that God preserved the apostolic teaching through the church and then conclude that every contradictory opinion found in church history is the apostolic teaching. We do not say that God preserved the canon through manuscripts and then identify the canon with every religious book surviving in Christian libraries. In precisely the same way, the fact that God preserved the text through manuscripts does not make every reading found in those manuscripts part of the preserved text.

The manuscripts are indispensable. The aggregate is providential. Neither proposition establishes that the aggregate as an aggregate is a text.

A Manuscript, a Reading, and a Text Are Not the Same Thing

Confusion enters this discussion because three related objects are treated as interchangeable.

A manuscript is a physical artifact: papyrus, parchment, or paper bearing written signs. It may contain the whole New Testament, one Gospel, a lectionary selection, or only a fragment. Its words may be correct in one place and defective in another.

A reading is the particular verbal form attested at a location of variation. At Matthew 1:7, for example, Asa and Asaph are competing readings. A reading is not an entire New Testament. It is one candidate wording within a textual unit.

A text is an ordered and determinate sequence of words. It says this rather than that at each point. Its identity is verbal and relational: these words, in this order, belonging to this work. The text of Matthew cannot simultaneously read both Asa and Asaph in the same place and in the same sense. Both forms may belong to the evidence. Both cannot be the one wording Matthew wrote.

B. B. Warfield himself made this elementary distinction. The words on a particular copy constitute the text of that document, but the text of the literary work is the author’s intended wording—the object the critic seeks when the copies disagree. Warfield even allowed, as a conceptual possibility, that the intended wording might not appear in any extant document and that all existing documents taken together might fail to contain it. He did not therefore identify the collection of documents with the text. The collection contains witnesses; criticism seeks the work’s wording from their testimony.²

That distinction is fatal to the claim that an aggregate of alternatives is itself a preserved text. If the critic must move from documents to readings, and from readings to the intended verbal sequence, then the documents are evidence concerning the text. They are not collectively identical with it.

This remains true even if the original reading exists somewhere in the aggregate at every place. Suppose a verse has three variants: A, B, and C. Suppose further that A is certainly original and physically survives in one manuscript. Material preservation has occurred. But a person who knows only “A, B, or C” does not yet know the text. The original wording exists among the candidates, while its identity remains unresolved.

The archive has preserved a disjunction. Scripture does not speak in disjunctions.

Contradictory Readings Cannot Collectively Be the Original

This point is logical before it is historical. Two contradictory propositions cannot both be true in the same respect at the same time. Likewise, two mutually exclusive wordings cannot both be the one original wording at the same textual location.

Mark either ended at 16:8 or continued with verses 9–20 as part of Mark’s Gospel. Matthew wrote either Asa or Asaph at Matthew 1:7, and either Amon or Amos at Matthew 1:10. First Timothy 3:16 contains either God or the relative pronoun as its original subject. The woman taken in adultery either belongs in John’s Gospel or it does not. At each point, the manuscript tradition may preserve extensive evidence for more than one answer. The existence of evidence for incompatible answers does not combine those answers into one text.

Nor does the relative smallness of the disputed portion alter the category. If ninety-nine percent of a statute is agreed upon, the remaining contradictory clauses are still not collectively the law. If two musical scores differ at one note, the bundle containing both is not thereby one determinate score at that point. If a library contains all twenty-seven New Testament books together with apocryphal gospels and epistles, the canonical books have materially survived, but the library’s complete inventory is not the canon.

The same reasoning applies to Scripture. A collection containing the authentic reading and its competitors is a collection of witnesses to the text. It becomes a determinate text only when the incompatible alternatives are excluded and one wording is received.

That exclusion need not be performed by an infallible human judge. Recognition is not inspiration. The church did not inspire the canonical books by receiving them, and a textual editor does not inspire a reading by printing it. Yet fallible recognition must still have an identifiable object. The possibility of human mistake does not turn contradictory objects into one object.

The Strongest Reading of Keister’s Statement

Fair argument requires us to begin with the strongest reading of Keister’s words. His 2024 statement by itself need not mean aggregate preservation. “Through the manuscripts” can express a thoroughly orthodox instrumental claim: God preserved the same inspired text by means of an abundant, decentralized succession of copies. Local defects arose, but the text was not universally corrupted or lost; defective witnesses could be corrected by better witnesses within the church’s transmissional possession.

If that is Keister’s meaning, the phrase poses no problem. It would distinguish the enduring text from the fallible artifacts that carry it, while affirming the providential use of those artifacts. Confessional Bibliology affirms precisely that distinction.

We must therefore move from the isolated sentence to Keister’s published account of how the text is identified. Here the relevant facts are public and do not require speculation about motive.

In his 2017 canons of New Testament textual criticism, Keister says that manuscripts are compared in order to discover the original reading. He assigns only probabilistic force to the canons, permits them to conflict, and regards the critic’s judgment and imagination as unavoidable. Continuity of attestation counts positively, but it does not exclude a reading newly recovered from a manuscript that had disappeared from the church’s known use. A rare or formerly hidden reading may displace a publicly received reading if the critic judges its evidence superior.

In his 2022 review of Why I Preach from the Received Text, Keister appeals to Warfield’s expectation that criticism can restore the inspired book word for word. He says that the original is “in the apographs,” but immediately raises the question of which apographs should be used to discover it. He also resists demands for absolute textual certainty and presents the critical-text position as holding a broader account of preservation and providence.

These facts yield a necessary inference, though not an accusation about intention. In Keister’s operative method, the manuscripts function as an evidentiary aggregate from which the original readings must be selected by probabilistic criticism. The method does not begin with a determinate ecclesiastical text whose local copying defects are corrected. It begins with competing witnesses whose claims are adjudicated in order to reconstruct the text.

That may be a coherent critical theory. It may produce correct decisions at many places. But it is not enough to answer the theological question by saying that God preserved the text “through the manuscripts.” The precise point at issue is whether the text was preserved as text, or whether only the evidence required for its progressive reconstruction was preserved.

Warfield’s Model Makes the Difference Explicit

Warfield is especially useful here because he states with unusual clarity what textual criticism seeks. The text of the work is not simply the wording found on any one document. It is the author’s intended sequence of words. The critic collects documents, classifies their testimony, rejects errors, and seeks to restore that intended wording.

Nothing in that description is objectionable merely because it involves comparison or correction. The crucial issue is the prior status of the object. Warfield’s conceptual account permits the pure text never to have existed perfectly in any surviving document. It even permits all surviving documents collectively to omit the intended reading. The text nevertheless remains a legitimate object of critical search because it is defined by authorial intention, not by ecclesiastical possession.

Warfield later expressed great optimism about the recoverability of the New Testament. He believed the autographic text lay within criticism’s reach in nearly the whole volume and hoped the book could be restored for the church word for word. We should state both sides. He was not a textual skeptic. He did not believe the New Testament had dissolved into unknowable chaos.

Yet optimism about restoration is not the same doctrine as continuous possession. The verbs expose the difference. One restores what is not presently intact as a possessed whole. One recognizes and corrects a text that remains possessed. Both activities use manuscripts, but they presuppose different accounts of what providence has already given the church.

This is why Keister’s appeal to Warfield cannot settle the meaning of “through the manuscripts.” Warfield’s model demonstrates that one may affirm an inspired original, abundant witnesses, and high recoverability while still locating textual purity formally in the lost autographic wording rather than in an extant authentic text. The doctrine may guarantee a successful or nearly successful reconstruction. It does not by itself confess that the church’s apographic Scriptures have been kept pure and are therefore authentic.

Two Meanings of “The Original Is in the Apographs”

The statement that the original is “in the apographs” sounds reassuring because historic Reformed theologians also called the extant Hebrew and Greek copies the originals. But the same sentence can carry two different doctrines.

Meaning: Apographic or ecclesiastical

What “in” signifies: The inspired wording is represented in the copies transmitted and possessed by the church; local defects are corrigible from the broader apographic witness.

Status of the text: The church possesses an authentic text, though no individual artifact or edition is inspired.

Meaning: Distributive or reconstructive

What “in” signifies: Correct readings are scattered somewhere among competing witnesses and must be selected; no manuscript, family, or received form need contain the complete text.

Status of the text: The original is the result sought by criticism and remains provisional where critical judgments remain open.

The first doctrine distinguishes text from artifact without separating text from the church. The second distinguishes the original from every extant form and locates it across an evidentiary field. Under the first, copies serve the preservation of a publicly functioning rule. Under the second, copies preserve the materials from which the rule must be critically restored.

The distinction matters because “in the apographs” can conceal the same ambiguity as “through the manuscripts.” If in means “present as the authentic text represented by the church’s apographic tradition,” the claim is confessional. If it means only “each original reading is presumed to exist in one or another witness, although the authentic sequence must still be assembled,” the claim amounts to aggregate survival plus critical recoverability.

The second proposition must not be smuggled under the historical authority of the first.

Westminster Confesses Scriptures, Not an Evidentiary Pool

Westminster Confession of Faith 1.8 identifies the Hebrew Old Testament and Greek New Testament as the writings immediately inspired by God. It then says that these Scriptures were kept pure in all ages by God’s singular care and providence, are therefore authentic, and are the final court of appeal in religious controversy.³

The grammatical and theological object is the original-language Scriptures. The confession does not say that all manuscripts collectively were kept pure, for the divines knew that individual copies contained errors. It does not say that every surviving reading is pure, since contradictory readings cannot all reproduce the same inspired wording. It says that the Scriptures were kept pure and therefore retain the public legal-theological status of authenticity.

That final clause rules out the identification of authenticity with a raw aggregate. An apparatus displaying mutually exclusive variants is immensely useful evidence, but it cannot be the final court of appeal merely as an apparatus. A court cannot render the verdict, “The authentic wording is A or B or C, according to the changing balance of probability.” A minister cannot exegete an unresolved set. A translator cannot place mutually exclusive alternatives into the same sentence in the same respect. A confession cannot bind the church to a disjunction.

The issue is not whether a critical edition prints a determinate main text. It plainly does. The issue is what gives that editorial text its status. If its determinate wording results from an eclectic series of revisable scholarly judgments, then the edition provides a current critical hypothesis about the authentic text. Its typography may be settled while its formal claims remain provisional. Westminster’s reasoning runs in the other direction: the Scriptures have been kept pure and therefore are authentic for present appeal.

Owen and Turretin Did Not Make Every Copy Perfect

Historic Reformed bibliology supplies a useful guardrail against two opposite errors. John Owen and Francis Turretin did not teach the inspiration of copyists. They acknowledged variants and local defects. They allowed the comparison of manuscripts and the correction of one copy by another. Their doctrine cannot honestly be reduced to the claim that every handwritten witness or every printed edition was immaculate.

But neither did they identify the authentic text with all variants collectively.

In his 1659 vindication of the Hebrew and Greek texts, Owen granted that the physical autographs had perished and that copyists were not infallible. Nevertheless, he spoke of the original-language copies as a treasure the church then possessed and had possessed for ages. They functioned as the rule and touchstone by which translations were tested. Public copies and the multiplicity of witnesses supplied correction against local scribal defects. His argument was not that every manuscript was perfect, but that the whole Scripture remained entire in the apographic possession of the church.⁴

Turretin likewise defined the “original texts” under controversy not as the vanished autographic sheets but as the apographs that set forth the Word in the inspired writers’ words. He expressly acknowledged variant readings arising from manuscript comparison. His question was whether the Hebrew and Greek sources had been so corrupted that they could no longer serve as the judge of controversies and the rule of versions. He denied that universal corruption.⁵

That is the relevant historic claim. Variants exist within the transmission, but the existence of variants does not demote the extant Scriptures from authentic rule to raw material for the recovery of a formally absent text. Copies can be imperfect witnesses to a preserved text. What they cannot be is contradictory constituents of one text.

Five Replies That Do Not Resolve the Problem

Several common answers leave the category mistake untouched.

1. “No two manuscripts are perfectly identical.” True, but irrelevant. Confessional Bibliology does not identify the text with one perfect physical artifact. Textual identity can persist across copies containing spelling differences, obvious slips, damaged lines, and corrigible local errors. The question is whether a determinate verbal text persists through those witnesses, not whether parchment becomes infallible.

2. “The witnesses agree overwhelmingly.” Their extensive agreement is a great providential mercy and strong evidence against skepticism. But agreement in most places does not make contradictory readings one text in the remaining places. Preservation is verbal, and the smaller set of genuine disputes still requires an account of identity.

3. “No major doctrine depends on a variant.” The redundancy of biblical doctrine protects the church from theological collapse. It does not tell us whether a particular clause is inspired. Mark 16:9–20 is either Scripture or not; its status is not answered by observing that the resurrection is taught elsewhere.

4. “The critical edition supplies a usable text.” It does, and believers reading faithful modern translations truly receive the Word of God. Yet practical usability and formal authenticity are not identical. The question is whether the edition receives a text already preserved for the church or provisionally reconstructs the text through changing critical judgments.

5. “A true reading may survive in a manuscript hidden for centuries.” Materially, yes. But if it disappeared from the church’s public textual life and remained unknown until modern recovery, its survival proves material continuity, not continuous ecclesiastical possession. To let such a reading overturn the church’s received text requires an argument that explains how this hidden survival fulfills “kept pure in all ages.” The word providence cannot perform that argument by itself.

The Verdict: Witnesses Are Not the Thing Witnessed

We can now answer the title question. No: an aggregate of competing readings does not, merely as an aggregate, constitute a preserved text. It constitutes a providential body of evidence. It may contain the original reading at every place. It may make correction and confirmation possible. It may display the extraordinary stability of transmission. But until contradictory readings are distinguished, the aggregate does not say what the text says at the points of contradiction.

This conclusion does not deny preservation through manuscripts. It defines the phrase properly. Manuscripts are the ordinary instruments by which God transmitted Scripture and the witnesses by which local defects may be exposed. The text is the determinate verbal object transmitted. The church is the historical recipient and steward. Textual criticism is a ministerial activity performed under the prior truth that God has preserved his Word; it is not the act by which a previously formless aggregate becomes Scripture.

Keister’s statement can therefore be affirmed only after its ambiguity is removed:

God preserved the inspired text through the manuscript tradition; he did not preserve the text by redefining it as the sum of mutually exclusive readings found in that tradition.

---

1. Lane Keister, “An Answer to Chris Thomas”, Green Baggins, November 26, 2024; see also his 2017 canons of New Testament textual criticism and 2022 chapter-by-chapter review. ↩
2. B. B. Warfield, An Introduction to the Textual Criticism of the New Testament (London: Hodder and Stoughton, 1886), especially 1–20. Warfield is used here as a clear historical statement of the opposing critical model, not as the controlling Reformed witness. ↩
3. Westminster Confession of Faith 1.8; see the 1647 text with Scripture proofs. ↩
4. John Owen, Of the Integrity and Purity of the Hebrew and Greek Text of the Scripture (Oxford, 1659), especially chapters 2–4. ↩
5. Francis Turretin, Institutes of Elenctic Theology, 2.10.2–3. Turretin’s question concerns whether the extant Hebrew and Greek sources remain the judge of controversies and the rule of versions, not whether each physical copy is perfect. ↩

If Keister means the first proposition, he should be able to identify the historically preserved text, explain how it functioned as the church’s authentic rule, and show how his method remains subordinate to that possession. If he means only that the correct readings survived somewhere in the aggregate and can be selected by probabilistic criticism, then his doctrine is recoverability, not the public preservation confessed in Westminster 1.8.

No judgment about his motives is necessary. The published method is enough to expose the unresolved question. God preserved through manuscripts—but what, precisely, existed through them as the church’s text?

Notes and Further Reading