Exegetical Method

How Scripture is worked on this site — the conviction that governs the work, the discipline that serves the conviction, and the standard by which the work may be tested and, where wrong, corrected.

The Governing Conviction

This site is governed by one conviction, and every page on it stands downstream of that conviction: the text must win, and the system must yield.

Scripture is the apostolic deposit — the authoritative witness of Christ and the apostles He commissioned, preserved in the canonical writings, with the Old Testament as the covenantal, prophetic, and typological world in which that witness speaks. What the apostles taught, commanded, exemplified, and warned against is the controlling material from which Christian doctrine is drawn. It is not raw data to be organized by a system standing over it; it is the standard to which every system answers.

Doctrinal systems are downstream of exegesis, never the reverse. A system earns its keep only as long as it remains accountable to the text. The moment it decides in advance what a passage is permitted to mean — before the passage has been read on its own terms — it has stopped serving Scripture and started replacing it. Here, systematic conclusions are suspended until the textual case has been built.

And doctrine must be drawn from what Scripture actually says, exemplifies, commands, or necessarily implies. Inference is legitimate; Scripture reasons, and so must its readers. But inference is governed by what stands in the text, not by what a system needs the text to support. Where Scripture is silent, that silence is not an empty space to be filled from elsewhere and stamped with apostolic authority.

None of this claims a reading scrubbed of presupposition. No one reads from nowhere, and the pretense that one does is itself a way of letting the system work unseen. The claim is narrower and harder: that whatever I bring to the text is not permitted to decide in advance what the text may say. The presuppositions are admitted so they can be checked, not denied so they can operate in the dark.

The Lineage of the Method

The conviction is mine; the discipline that serves it is not invented from nothing. The method employed here is grammatical-historical-canonical exegesis, which draws on four streams of careful interpretive labor. I name them plainly, as both a debt and a disclosure — so the work can be checked against the same standards as its sources.

From Gordon Fee and Douglas Stuart, I take the conviction that exegesis is a disciplined, step-by-step recovery of the author’s intended meaning, owed to every text before a word is preached from it.1 Fee’s account of the New Testament and Stuart’s of the Old share one instinct: the meaning is in the text — in its grammar, syntax, and historical setting — and the interpreter’s task is to draw it out, not to bring it in.

From Thomas Schreiner, I adopt the practice of tracing the argument: following an apostle’s reasoning clause by clause, mapping how each proposition supports, qualifies, or advances the one before it, so that the unit of interpretation is the argument rather than the isolated verse.2 A verse is a unit of citation; the paragraph and the discourse are the units of meaning. Schreiner judged this the single most important step in exegesis, and I work the same way for the same reason: it is the discipline that most reliably exposes a proof-text wrenched from the flow that gave it sense.

From Peter Gentry and Stephen Wellum, I adopt the practice of reading every passage across three horizons — the textual (the passage in its own grammar, genre, and immediate context), the epochal (its place in the unfolding history of redemption), and the canonical (its place in the whole of Scripture as it comes to rest in Christ).3 A text read only at the first horizon is read atomistically; a text rushed to the third is systematized prematurely. Keeping the three in tension keeps both errors at bay.

One thing must be said plainly about the latter two streams, because it is the governing conviction applied to my own sources. Schreiner, Gentry, and Wellum work within a Reformed and covenantal systematic tradition whose conclusions I do not share — regarding the structure of the covenants, the ordering of salvation, baptism, and more. I adopt their tools, not their system. The whole purpose of those tools is to let the text speak on its own terms; when the same authors’ systematic conclusions are not in fact established by the text those tools uncover, those conclusions yield, exactly as my own must. To borrow a method is not to inherit a system. A method that could only ever confirm the system that produced it would be no method at all — only the system wearing a method’s clothes.4 And separating the tool from the conclusion is no hostile act toward these men: each of them holds that his method serves the priority of the text over the system. I am holding them to that claim — and holding myself to it in the same breath.

The Method Turned on Its Own Sources

I do not say this as a suspicion about motives. I say it as a finding. When I use their own tools — the tracing of the argument, the three horizons — without the covenantal system steadying the result in advance, the system’s fingerprints come into view: at decisive points, the conclusion is carried by the framework rather than by the text the framework was meant to be reading. Their stated order — text first, system second — is one I share. What I have found is that this order is not always kept, and the tools themselves reveal the lapse.

A claim like this is worth nothing left unstated, so take one example. Gentry and Wellum trace the covenant signs with real density — the rainbow, circumcision, the priesthood, the temple — each followed across its textual, epochal, and canonical horizons and allowed the full weight Scripture gives it. Then they reach baptism, and the density thins: baptism is rendered “the sign of this faith,”5 and the tracing largely stops there. But the apostolic texts will not sit still at “sign of faith.” Baptism is joined to repentance and the forgiveness of sins (Acts 2:38), to the washing away of sins (Acts 22:16), to burial and resurrection with Christ (Romans 6:3–4), to the circumcision made without hands (Colossians 2:11–12), to the washing of regeneration (Titus 3:5), and to salvation as an appeal to God (1 Peter 3:21). Run the three horizons on baptism with the patience spent on circumcision, and it will not reduce to the sign of a faith already complete without it.

The point is not that they neglect the method. It is that they apply it richly to the signs their system can absorb and sparingly to those it needs to remain a bare marker. The tool was sound. It was set down early — at precisely the point where finishing the work would have cost the conclusion. A fuller treatment of this asymmetry, with the parallel cases, is developed in a separate study; here, it stands as one demonstration that the method on this site is not a slogan but a discipline I am willing to apply to my own sources.

How a Passage Is Worked: The Ten Stages

What follows is the procedure itself, in the order I actually use it. The stages are not a ritual; they are a way of refusing to arrive before the evidence does. For a short question, several stages are compressed into a sentence; for a contested doctrine, each is worked through in full. None is skipped.

1. Set down the inherited reading. Before asking what the passage means, I set aside what I have been told it means. This is not a claim to read without presuppositions — no one does — but a refusal to let a systematic conclusion decide in advance what the text may say. I begin with the words before me: what is said, in what words, in what order, to whom, in what circumstances.

2. Grammar and syntax. I work through the lexis (the word’s range and the author’s usage), the syntax (clause relations, case function, agreement), verb aspect, tense, voice, and mood, and the connective tissue that carries the logic — γάρ, οὖν, διά, ἵνα, ὥστε, καί, δέ, and their Hebrew analogs — along with the discourse markers that signal emphasis, contrast, climax, or a structural turn.6 Translations are compared here. Where a rendering resolves a difficult reading too smoothly, I treat the smoothness as a flag, not an answer, and produce a working translation with the grammar noted.

3. Immediate context and the flow of the argument. A verse is a unit of citation; the paragraph, the discourse, and the book are the units of argument. I trace what precedes and follows, the passage’s structural function, and any literary shape — chiasm, parallelism, inclusio — to build a brief map of how the argument actually develops.7 This is the discipline that most reliably exposes a verse pulled loose from the reasoning that gives it sense.

4. The canonical pattern. I move outward to the rest of Scripture, asking whether the language, sequence, command, or warning recurs — with particular attention to the apostolic deposit in Acts, the Epistles, Hebrews, and Revelation. A canonical pattern is a feature of the apostolic witness that recurs, is structurally embedded, and is treated as normative within Scripture itself. The apostles are authoritative witnesses to Christ and His fulfillment of the Law, the Prophets, and the Writings; their use of the Old Testament governs Christian synthesis without erasing the Old Testament’s own historical and literary meaning.

5. The three horizons. I read the passage at three levels: the textual horizon (its grammar, genre, and immediate setting), the epochal horizon (its place in the unfolding history of redemption), and the canonical horizon (its place in the whole canon as God’s purposes come to rest in Christ).8 Held together, the three guard against two opposite errors — reading a text in atomistic isolation and rushing it into a system before it has spoken.

6. Typology, tested. Where the text invites a type, I follow it — but typology must arise from Scripture’s own pattern of fulfillment, not from interpretive imagination. I ask whether the New Testament itself draws the connection, whether the pattern recurs, and whether the type’s structure actually matches the claimed fulfillment. Canonical restraint is preferred over typological enthusiasm.

7. Command, example, and necessary inference. I distinguish what Scripture commands, what the apostles and the earliest church did, and what follows by necessary inference from textual premises. Each carries different weight; the distinctions and the objections they most often draw are set out in the next section.

8. Test the inherited claim — now. Only after the textual case is built do I bring the candidate doctrine to it and ask directly: What words must mean what the system says they mean? What grammatical features support or resist the reading? What canonical patterns does it honor or violate? What apostolic commands does it neutralize or relocate? What warnings does it explain away? What sequences in the apostolic deposit does it reorder? And what words is the system using that are not Scripture’s own — is the substitution doing interpretive work the text does not do? A claim that requires the text to say what it does not say has failed the test, whatever its pedigree, elegance, or the prestige of its proponents.

9. The witness of the early church. I consult the earliest post-apostolic readers as a confirming or flagging witness — never as a court of final appeal. This is developed below.

10. Articulation and obedience. I state the doctrine in Scripture’s own vocabulary and emphases, especially where they diverge from later technical language, and then ask what it requires — of me, of the church, and of habit and practice. A doctrine is not finished when it is articulated. It is finished when it is obeyed.

Command, Example, and Necessary Inference

Not everything Scripture contains binds the conscience in the same way, and a method that cannot distinguish among them will either bind too much or too little. I work with three categories — command, example, and necessary inference. Because they stem from a particular tradition with a particular reputation, I define them in their disciplined form and then address the objection they most often raise.

A command is what Scripture explicitly instructs the church to do, believe, refuse, or endure. It is normative unless Scripture itself revokes or limits it. The burden of proof rests on any claim that a plain command has been suspended, not on the command.

An example is what the apostles and the earliest church actually did. An example becomes binding when Scripture itself treats it as normative — when it recurs, when it is held up for imitation, when departure from it is rebuked, or when it is woven into a command. An example that meets none of these tests is instructive but not, on its own, binding, and I mark it as such. Bare precedent — “it happened, therefore we must” — does not bind. The default runs the other way: Scripture narrates only description until its normative intent is shown.

A necessary inference is a conclusion that follows validly from premises that are themselves in the text. It is binding only when both conditions hold: the premises are textual, and the logic is sound. If either fails, the inference is merely possible, and possible inferences do not carry doctrinal weight. In particular, the argument from silence is not a necessary inference, and it is not used here as if it were. The line between necessary and possible inference is where much system-driven reading hides — in every tradition, including this one — and I try to keep it honest.

The standard objection is answered. A common challenge claims that this method leans on command, example, and inference — the framework often labeled CENI — while Gordon Fee, whose exegetical procedure I credit, warned against exactly that. The challenge misfires, and the reason matters. Fee did not address command, example, and necessary inference as a system; what he challenged was the habit of treating whatever a narrative records as something the church must repeat — the move from “it happened” to “do it.”9 That move I reject as fully as he does. But Fee does not stop there. He holds that precedent carries normative weight when it is anchored in the author’s intent, and he names the conditions: a consistent pattern repeated within the New Testament, harmony with what Scripture teaches plainly elsewhere, and due account of what is merely cultural. Those are, nearly word for word, the conditions under which an example becomes binding here. So the conflict the objection imagines is not between Fee and this method; it is between Fee and a cruder precedent-based reasoning that this method also rejects. Where I treat an example as binding, I am applying Fee’s test, not evading it.

This obliges me to disown the harder forms of the tradition by name. The “law of silence” that would forbid whatever Scripture nowhere addresses, treating any apostolic example as binding by default, and inferring to plug a gap in the record — these run afoul of Fee’s caution and are not the method practiced here. I place the burden of proof in Fee’s direction and cite him for his principle of precedent, not as a general endorser of conclusions he would not himself share — the same separation of tool from system that governs every source I use.

How Conclusions Are Weighed

A claim is only as strong as the evidence supporting it, and honesty requires stating how strong that evidence is. Conclusions on this site are marked, explicitly or in effect, at one of five levels:

  • Certain — an explicit apostolic command or a repeated, unambiguous teaching.
  • Strong — a repeated canonical pattern reinforced by grammar and context.
  • Probable — strong grammar and context in a single passage, with limited, consistent reinforcement elsewhere.
  • Possible — a plausible inference the text permits but does not compel.
  • Speculative — not sufficiently grounded to bear doctrinal weight.

I do not upgrade a claim beyond what the evidence supports. Treating a probable inference as certain doctrine is no less an error than ignoring a plain command; both substitute the interpreter’s confidence for the text’s. When the case is thin, it is marked as thin. And the level itself is a judgment open to challenge: if a reader can show that a conclusion here has been graded higher than its evidence warrants, the regrading is a correction the method is bound to accept.

On the Text and Its Translations

I quote the New King James Version throughout this site for its readability and fidelity to the traditional text. But quotation and exegesis are not the same act. When I work a passage, I read it against the ESV, NIV (1984), NASB (1995), ASV, and NET — and I do not stop at any English version. The argument is finally settled at the level of the Greek and Hebrew, assessed against the manuscript evidence as a whole rather than against a single printed edition or text-type. No translation, and no editor’s reconstructed text, is permitted to stand in for the witnesses themselves. The text must win; the system must yield.

The Witness of the Early Church

After the textual case is made, I ask whether the earliest Christians — before the major doctrinal developments of later centuries — read the passage roughly as the exegesis suggests. This is corroboration, not authority; and because a corroborating witness is easy to summon when it agrees and to dismiss when it does not, I state the rule before I use it and hold to it even when it cuts against me.

The rule is this. The Fathers are witnesses, not Scripture and not a court of final appeal; witnesses are weighed, not counted. Their agreement with a reading drawn from the text strengthens it. Their division is reported as a division, not as a consensus that never existed. Their silence is evidence, but rarely conclusive, since the surviving record is uneven and some apostolic teaching was preserved only indirectly. Where a later, developed consensus has plainly traveled some distance from the apostolic witness, the witness governs the consensus, not the other way around. And where the earliest readers tell against a conclusion I would otherwise reach, I say so and let it press on the conclusion, rather than quietly setting the Fathers aside when they become inconvenient. This stage confirms or flags; it does not decide.

What This Method Refuses

The discipline is as much about where I refuse to begin as about how I proceed. The following moves are not starting points here. Each may be brought to the text as a candidate claim for testing; none may be applied to the text as a settled lens. If a passage genuinely requires one of them, that requirement is the burden of proof to be discharged, not an assumption to be granted.

I do not begin with a decretal or monergistic scheme imposed as the lens through which conversion, election, and perseverance texts must be read. I do not begin with a fixed ordo salutis that settles the sequence and relation of faith, repentance, regeneration, baptism, and the reception of the Spirit before the texts themselves have established it. I do not read the apostolic warnings — in Hebrews, Romans 11, Galatians, John 15, the Gospels — as hypothetical or aimed only at the merely professing, when the text addresses real participants and warns of real loss. I do not relocate a concrete apostolic command into the categories of “ceremonial,” “cultural,” or “dispensational” without explicit textual warrant for the move, nor do I flatten such commands into abstract “means of grace” language that obscures what was actually commanded. And I do not treat any confessional, creedal, or denominational document — of any tradition — as an exegetical authority. Such documents may be tested against the text; they may not be applied to it.

The refusal cuts toward my own house as well, and it must, or it is not a method but partisanship. The convictions this site is known for — that baptism is the divinely appointed point at which the New Testament locates the forgiveness of sins and union with Christ, and that the apostolic warnings address real covenant participants who can fall away — are presented here as conclusions drawn from texts, not as lenses brought to them. They sit on the table to be tested like any other claim, and the day a reader shows that the grammar, the pattern, and the apostolic argument will not bear them is the day they go. A confessional document of my own tradition carries no more exegetical authority than anyone else’s. To exempt my own distinctives from the test would be to commit the very fault I charge against others: to let the system secure the conclusion before the text has finished speaking.

Behind all of this stands a single refusal, and Paul names it: I am not “as so many, peddling the word of God” (2 Corinthians 2:17). When a cherished conclusion collides with the text, the conclusion goes.

An Honest Method Invites Testing

A method that hides its steps cannot be checked, and a conclusion that cannot be checked is asserted rather than demonstrated. Everything above is laid out for the opposite purpose. The stages are visible so the reasoning can be followed; the certainty levels are stated so the weight of each claim is known; the sources are named so borrowing can be audited; the refusals are declared so the starting points are not smuggled in; and the standing objections are answered openly so they can be pressed harder, not buried.

That is why this site maintains an open challenge. If a reading here rests on grammar the text does not support, on a pattern claimed from too few passages, on a typology stretched beyond Scripture’s own, on an example treated as a command without warrant, or on a conclusion held with a certainty higher than the evidence allows — bring it to the text and show it. The method is accountable to one thing: the apostolic deposit, in its grammar, its patterns, its commands, its examples, its warnings, and its earliest reception. That is the only correction the work is bound to accept. It is also the only correction worth giving.

Notes

  1. Gordon D. Fee, New Testament Exegesis: A Handbook for Students and Pastors, 3rd ed. (Louisville: Westminster John Knox, 2002); Douglas Stuart, Old Testament Exegesis: A Handbook for Students and Pastors, 5th ed. (Louisville: Westminster John Knox, 2022). Fee frames the task as recovering the divine-human intention locked within the text through an ordered sequence of steps: Stuart does the same work for the Hebrew Scriptures. ↩︎
  2. Thomas R. Schreiner, Interpreting the Pauline Epistles, 2nd ed. (Grand Rapids: Baker Academic, 2011), ch. 6, “Tracing the Argument,” at 97. The tracing, or “arcing,” discipline that Schreiner employs derives from Daniel Fuller. ↩︎
  3. Peter J. Gentry and Stephen J. Wellum, Kingdom through Covenant: A Biblical-Theological Understanding of the Covenants, 2nd ed. (Wheaton: Crossway, 2018), esp. 113, 119–25. Wellum credits the three-horizon framework to Richard Lints, The Fabric of Theology: A Prolegomenon to Evangelical Theology (Grand Rapids: Eerdmans, 1993), 259–311, esp. 293, who in turn drew it from Edmund Clowney and the Geerhardus Vos line of biblical theology. ↩︎
  4. The governing conviction applied reflexively to the method’s own sources: adopting a grammatical-historical-canonical procedure entails no commitment to the covenantal or Reformed conclusions some of its developers reach with it. Those conclusions remain candidate claims, tested against the text like any other. ↩︎
  5. Gentry and Wellum, Kingdom through Covenant, in the course of arguing that new-covenant membership is constituted by regeneration and faith, and applying that conclusion against infant baptism, 818-19. ↩︎
  6. On grammatical and syntactical analysis as the foundation of exegesis, see Fee, New Testament Exegesis, and Stuart, Old Testament Exegesis (n. 1). ↩︎
  7. On tracing the structure of the argument across a discourse, see Schreiner, Interpreting the Pauline Epistles, ch. 6 (n. 2). ↩︎
  8. On the textual, epochal, and canonical horizons, see Gentry and Wellum, Kingdom through Covenant (n. 3), with the prior caution that the reading discipline is adopted here apart from the covenantal system its authors build with it. ↩︎
  9. Gordon D. Fee and Douglas Stuart, How to Read the Bible for All Its Worth, 4th ed. (Grand Rapids: Zondervan, 2014), ch. 6, “Acts: The Question of Historical Precedent.” Fee and Stuart hold that what is only narrated or described does not function normatively unless Scripture commands it or the author’s intent shows it to be normative, and that biblical precedent may be a repeatable pattern under stated conditions — a consistent pattern repeated within the New Testament, in harmony with what is taught plainly elsewhere, and with allowance for what is culturally conditioned. The categories used here (“CENI” — command, example, necessary inference) descend from the American Restoration Movement; this method employs them in a form constrained by these conditions and disavows the harder applications (the argument from silence; apostolic example presumed binding by default), 128-131. ↩︎