Contradiction, Tension, Ambiguity, and Proof-Text Loss

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.

Propositions Before Verdicts

A contradiction does not exist merely because two strings of words look different. Nor does agreement exist merely because two passages share vocabulary. The primary objects of logical comparison are propositions: what is asserted or denied of a subject under specified conditions.

Aristotle distinguished an isolated term from an affirmation or denial capable of being true or false. He also observed that contradictory propositions require the same subject and predicate without equivocation.¹ That ancient discipline remains elementary but indispensable. Before a dossier can classify the relation between passages, it must state what each passage actually commits the interpreter to affirm.

For this series, a local proposition record will identify at least:

• the subject under discussion;
• the predicate affirmed or denied;
• the referent of pronouns, titles, persons, places, and events;
• the scope of negations, quantifiers, exceptions, and conditions;
• the time or temporal relation involved;
• the modality of the statement—actual, necessary, possible, commanded, promised, permitted, or predicted;
• the respect or aspect in which the predicate applies;
• the agency by which an action is performed;
• the sense in which material terms are being used; and
• the textual wording and interpretation from which the proposition was drawn.

The record matters because many alleged contradictions dissolve before the propositions become identical enough to contradict.

“Christ died” and “Christ lives” are not contradictory when the first concerns one time and the second another. “The Son is equal to the Father” and “the Father is greater than I” do not formally contradict unless both statements concern the same nature, office, state, relation, and respect. “A person is justified apart from works” and “a person is justified by works and not by faith alone” cannot be adjudicated by isolating the verb justify while ignoring the kind of works, the polemical target, the evidentiary relation, and the sense in which vindication or declarative righteousness is at issue.

The same rule applies to textual criticism. A received reading and a critical reading may produce different sentences without producing contradictory propositions. Conversely, very similar Greek or Hebrew expressions can assert incompatible propositions if a negation, subject, number, referent, or relation changes.

The dossier must therefore translate wording into propositions before it translates difficulty into a verdict.

Two Axes, Not One Ladder

The central distinction of this article is easiest to see as two independent axes.

The axes interact, but neither can replace the other.

Suppose a longer reading explicitly states a doctrinal relation and the shorter reading does not. The shorter reading may remain perfectly coherent with the doctrine. The result is proof-text loss without contradiction.

Suppose two passages appear to pull in different directions but a warranted distinction preserves both. The result is resolvable tension without proof loss.

Suppose a sentence admits two viable meanings, each consistent with the wider canon, but only one proves a Westminster clause. The result may be semantic ambiguity and doctrinal underdetermination even though neither interpretation contradicts the canon.

Suppose a selected reading plainly denies an exact confessional proposition, while no other biblical passage addresses the same question. The result can be textual incompatibility with that proposition without an internal biblical contradiction.

Finally, suppose two operative readings yield P and not-P after every legitimate distinction is exhausted. That result is a strict contradiction within the selected reconstruction. It does not follow that God inspired contradiction. It follows that at least one element of the reconstruction—reading, translation, interpretation, or inference—must be wrong.

This two-axis model prevents two pervasive non sequiturs:

Proof lost, therefore doctrine contradicted.

---

1. Aristotle, On Interpretation 1, 5–6, trans. E. M. Edghill, especially the distinction between significant terms and propositions and the definition of contradictory affirmation and denial. See the Internet Classics Archive text. ↩

and

Doctrine not contradicted, therefore no proof was lost.

---

1. Aristotle, On Interpretation 1, 5–6, trans. E. M. Edghill, especially the distinction between significant terms and propositions and the definition of contradictory affirmation and denial. See the Internet Classics Archive text. ↩

Both are invalid.

Strict Contradiction: The C4 Threshold

Aristotle’s classical formulation says that the same attribute cannot both belong and not belong to the same subject “at the same time” and “in the same respect.”² The qualifications are not loopholes invented to protect Scripture. They are part of what contradiction means.

In its simplest propositional form, strict contradiction is:

A: P
B: not-P

---

1. Aristotle, On Interpretation 1, 5–6, trans. E. M. Edghill, especially the distinction between significant terms and propositions and the definition of contradictory affirmation and denial. See the Internet Classics Archive text. ↩
2. Aristotle, Metaphysics IV.3, 1005b19–24, trans. W. D. Ross. See Metaphysics, Book IV. ↩

The pair cannot both be true. Yet biblical claims are rarely exhausted by a bare letter. A useful expanded form is:

A: Subject S has predicate P under respect R, at time T, in sense σ, with modality M, under conditions K.
B: The same subject S does not have predicate P under the same R, T, σ, M, and K.

---

1. Aristotle, On Interpretation 1, 5–6, trans. E. M. Edghill, especially the distinction between significant terms and propositions and the definition of contradictory affirmation and denial. See the Internet Classics Archive text. ↩
2. Aristotle, Metaphysics IV.3, 1005b19–24, trans. W. D. Ross. See Metaphysics, Book IV. ↩

A C4 verdict requires identity across every material coordinate.

Statements about God, Christ according to his divine nature, Christ according to his human nature, the covenant head, the visible church, the invisible church, Israel as a nation, the elect within Israel, and humanity in Adam cannot be interchanged without proof. A shared name does not always establish a shared subject in the relevant respect.

Near-synonyms are not necessarily identical predicates. To be “destroyed,” “dissolved,” “burned,” “exposed,” “not found,” “renewed,” and “passed away” may overlap in an eschatological account, but their logical relations must be established rather than assumed.

The same person may be greater in office and equal in nature, condemned according to law and justified in Christ, physically dead and spiritually alive, or unchangeable in being while entering new relations with creatures. A distinction of respect resolves a conflict only when the texts warrant it.

Promise and fulfillment, humiliation and exaltation, death and resurrection, the present age and the age to come, and successive covenantal administrations introduce temporal differences. A later state can differ from an earlier state without contradiction.

Equivocation prevents a genuine contradiction from being demonstrated. If faith, works, world, flesh, law, justify, or death changes sense between propositions, the apparent P / not-P pair has not yet been established. The change of sense must itself be contextually warranted; it cannot be asserted merely because contradiction would otherwise result.

“All,” “some,” “none,” “not all,” “whoever,” “unless,” “if,” “only,” and “except” can determine the logical force of a passage. The contradictory of “every instance of anger is sinful” is “not every instance of anger is sinful,” not “no instance of anger is sinful.” Contraries and contradictories are not the same. Contraries cannot both be true, but they may both be false; a contradictory pair divides the relevant possibilities so that one must be false if the other is true.³

What God commands, what ordinarily occurs, what may occur, what must occur, and what will occur are different. “You may eat” does not contradict “you need not eat.” “This can happen” does not contradict “this will not happen.” A promise, prediction, permission, and description cannot be collapsed into one modal category.

The contradiction must belong to one declared corpus under interpretations that the corpus can own. A proposition drawn from NA28 cannot be combined silently with a proposition drawn from a different apparatus reading and then described as a contradiction “in NA28.” An ESV rendering cannot be attributed to the Greek main text when it reflects another option. Article 8 will distinguish these layers in detail; Article 7 fixes the logical requirement that the compared propositions have an identified source.

Only after these identities have been established and every warranted distinction has failed may the series assign C4 — Strict contradiction.

Tension: Material Pressure without P and Not-P

A textual or canonical tension exists when two operative readings, emphases, implications, or natural interpretations pull a doctrinal construction in materially different directions without yet yielding a formal contradiction.

Tension is stronger than surprise and weaker than contradiction.

Examples of tension can include:

• a broad prohibition beside an apparently approved instance of the same general action;
• destruction language beside renewal language where the mode of continuity is not specified;
• an absolute-sounding statement beside an exception established elsewhere;
• two narratives whose details resist a single chronology but do not explicitly deny one another;
• an Old Testament wording and a New Testament citation that make different lexical assertions about the same source;
• a selected reading whose most natural local sense pressures, but does not logically negate, an established doctrine; or
• two passages that can be reconciled only after a distinction is supplied from a third passage.

Tension may be resolvable, residual, or repair-dependent.

The text itself supplies a distinction strong enough to preserve both claims. The result ordinarily fits C1. The tension was real at the level of first appearance, but the final construction is coherent.

More than one reconciliation remains viable, or a proposed distinction reduces but does not eliminate the pressure. Residual tension may contribute to a grade of E — Underdetermined, but it does not automatically do so. The unresolved detail may be immaterial to the doctrine being tested.

The readings cohere only after an apparatus reading, conjecture, translation departure, or other textual substitution is introduced. The resulting theology may be true, but coherence has not been demonstrated from the declared main text. That condition fits C2.

Tension is therefore a necessary category of intellectual honesty. If every difficulty is called contradiction, the argument becomes sensational. If every nonformal conflict is called harmless, material theological pressure disappears from the record.

Ambiguity: More Than One Viable Sense

An expression is ambiguous for the purposes of this series when its operative wording or evidence supports two or more materially distinct interpretations and the relevant context does not yet select among them.

Ambiguity is epistemic. Westminster’s claim that the true and full sense of Scripture is one does not mean that every reader immediately knows that sense or that every transmitted wording is equally clear. A passage can have one intended sense while the evidence available in a selected edition leaves interpreters unable to establish which sense it is.

Several kinds of ambiguity must be distinguished.

Article 8 will assign textual, translational, and interpretive responsibility more fully. Here the important point is logical: an ambiguous sentence does not yet supply one determinate premise.

This does not mean that every imaginable reading counts. A viable interpretation must satisfy grammar, lexical usage, discourse, historical setting, and canonical context. Remote possibility is not ambiguity. If a sentence can bear sense A naturally and sense B only after ordinary syntax is violated, the existence of B does not make the passage evenly underdetermined.

Nor does ambiguity entail contradiction. If a word can mean X or Y, the text has not asserted both X and not-X. The reader lacks sufficient grounds to choose. Treating that uncertainty as contradiction overstates the evidence; selecting the doctrinally preferred option as though it were necessary understates it.

Proof-Text Loss: An Evidentiary Category

Proof-text loss occurs when a passage in the operative corpus no longer establishes the proposition, premise, boundary, relation, denial, or attribution for which it was used.

The definition is indexed. It must name:

1. the passage and operative reading;
2. the exact proposition being tested;
3. the proof function previously assigned to the passage;
4. the comparison text or confessional use relative to which loss is alleged; and
5. the evidence, if any, that remains elsewhere.

Without those indices, “the proof text is gone” is too vague. The verse number may remain while the material clause is absent. The clause may remain while a limiting word changes. The doctrinal truth may remain elsewhere while the literary attribution disappears. Or the passage may never have been a sound proof of the proposition in the first place.

Proof texts can perform different functions.

This functional analysis prevents equal treatment of unequal changes. Losing an illustration is not the same as losing a necessary premise. Losing one redundant proof is not the same as losing the only passage that states an exclusion. Losing an explicit formula is not identical to losing the doctrine’s conceptual content, though the formula may possess significant confessional and ecclesiastical value.

Proof-text loss can be total or partial. A total loss occurs when the operative passage supplies none of the assigned proposition. A partial loss occurs when the subject remains but a modifier, denial, relation, scope marker, or attribution disappears. Partial loss is often the more important category in this series because Reformed doctrine frequently depends upon exact qualifications rather than isolated nouns.

Proof-Text Loss Is Not Textual Incompatibility

The difference between proof loss and grade F — Textually incompatible is the difference between failing to establish and actively excluding.

If a passage is silent about proposition D, it does not prove D. It also does not, by silence alone, prove not-D.

If a shorter reading removes an explicit predicate, the predicate may still be true. If an ambiguous reading permits D and a rival D2, D remains possible but underdetermined. If a selected reading asserts not-D under the same respect in which the confession asserts D, the relation is incompatibility.

The progression is therefore:

Grade F therefore requires more than omission, uncertainty, reduced explicitness, or loss of a familiar proof. It requires the best-supported natural reading of the operative corpus to contradict or exclude the exact confessional proposition.

The converse is also important: grade F does not require a C4 contradiction within the biblical corpus. A single selected reading can exclude a Westminster-specific claim without contradicting another biblical passage. The derivation scale measures relation to a doctrinal target; the coherence scale measures relation among operative propositions.

Strong Objections

“These philosophical categories impose an alien system on Scripture”

The categories do not dictate what Scripture may say. They state what affirmation, denial, ambiguity, inference, and contradiction mean. Scripture itself reasons, distinguishes, denies consequences, answers objections, and draws conclusions. Westminster’s “good and necessary consequence” assumes that some conclusions follow and others do not.

The greater danger is an undefined use of contradiction that changes meaning whenever the argument requires it.

“Because Scripture is coherent, a C4 finding is impossible”

A C4 finding is impossible of the inspired truth as God gave it. It is not impossible for a humanly reconstructed corpus, translation, or interpretation. If selected readings yield P and not-P, the confessional conclusion is that the reconstruction must be corrected—not that contradiction has become true.

“If the doctrine is taught elsewhere, proof-text loss is meaningless”

The doctrine may survive while the passage’s function changes. Directness, precision, biblical formula, exclusion of error, literary attribution, and cumulative load are all real evidentiary properties. Westminster did not confess only a set of conclusions; it claimed that those conclusions are expressed or necessarily deduced from Scripture.

“If an orthodox interpretation is possible, ambiguity is harmless”

Possibility may be enough to refute a contradiction allegation. It is not enough to establish a doctrine by necessary consequence. An ambiguous passage can protect a system from grade F while still contributing to grade E.

“A harmonization need only show that contradiction is avoidable”

That is sufficient only for the narrow claim that C4 has not been demonstrated. It is not sufficient for C1, which requires a warranted resolution, or for grades A and B, which require positive derivation. A merely possible harmony can defeat an overconfident critic without completing the reconstruction.

“The received text also contains difficulties”

Certainly. The same definitions apply to received readings. Article 10 will use shared difficulties as negative controls. A defect cannot be assigned uniquely to the critical corpus when both corpora possess the same operative propositions. Conversely, a problem shared by another text does not cease to be a problem.

“Formal logic cannot capture poetry, narrative, typology, or progressive revelation”

Formalization must respect genre. A metaphor is not converted mechanically into a physical assertion; a narrative report is not automatically divine approval; a type and antitype are not identical in every feature. But once an interpreter says what a poem, narrative, or type teaches, he has made propositional commitments. Genre complicates the work of stating them; it does not permit them to be both affirmed and denied in the same respect.

Verdict

Contradiction, tension, ambiguity, and proof-text loss name different conditions and must never be used as interchangeable measures of textual damage.

A strict contradiction requires P and not-P of the same subject, predicate, referent, time, respect, sense, scope, modality, and conditions. It receives C4 only after every warranted grammatical, historical, canonical, and theological distinction has been exhausted. An inconsistent doctrinal set must display the complete chain by which its several premises entail both conclusions.

A tension records material pressure without claiming formal contradiction. It may be resolved by an evidenced distinction, remain residual, or require textual or translational repair. The category permits the series to acknowledge a real difficulty without sensationalizing it.

An ambiguity exists when two or more materially viable senses remain available. It does not assert both senses as true. If the wider corpus fails to adjudicate them and the difference bears upon the doctrine, the result is underdetermination. An orthodox possibility is not yet a good and necessary consequence.

A proof-text loss occurs when an operative passage no longer performs a specified evidentiary function. The doctrine may still be directly established elsewhere, cumulatively reconstructed, materially weakened, dependent upon inherited synthesis, underdetermined, or excluded. Proof loss therefore requires a local finding and a separate whole-corpus verdict.

These distinctions clarify the founding examples. The omission of εἰκῇ from Matthew 5:22 removes an explicit local qualifier and creates a serious question about the relation between culpable anger, righteous anger, and Christ’s sinlessness. It does not, by omission alone, prove that the critical text makes Christ a sinner. The shorter reading must first be shown to condemn the same moral species of anger displayed in Mark 3:5.

The shorter 1 John 5:7–8 loses an explicit heavenly triad that numbers and names the witnesses and declares their unity. That loss does not contradict monotheism or erase the deity and distinction of the Father, Son, and Holy Spirit elsewhere; it does require a separate closure test before the exact proposition “three and only three persons” may be credited to the reduced corpus. The shorter Matthew 6:13 loses the main-text basis for Westminster’s claim about the conclusion of the Lord’s Prayer without denying God’s kingdom, power, or glory. Second Peter 3:10 presents different Greek and English predicates whose textual, translational, and eschatological relations must be separated before contradiction or doctrinal loss can be measured.

The series’ verdict system is consequently two-dimensional. Grades A–F measure derivation of the exact confessional proposition. Codes C0–C4 measure coherence among the operative readings and interpretations. Grade F does not require C4, and C4 does not follow from proof loss. A doctrine can be coherent but underdetermined, preserved but weakened, excluded without internal contradiction, or cumulatively derived after a warranted resolution.

This framework raises the standard of the critique. It refuses to protect a critical reconstruction by vague appeals to doctrine elsewhere, and it refuses to condemn that reconstruction by an invalid leap from variant to contradiction. Every later allegation must name the proposition, fix the corpus, trace the premises, classify the relation, identify the proof function, admit the strongest response, and issue only the verdict the evidence can bear.

Notes and Further Reading