How This Study Was Done

This project makes a large claim — that every place the New Testament speaks of someone being saved has been examined in context — so the method behind it should be visible and checkable. This page sets out what was included, what was left out, which texts were used, and how the difficult cases are handled.

The Governing Question

Every study asks the same thing:

What does the New Testament mean here, in this passage, when it says someone or something was saved?

That question implies a rule, and the rule governs everything else: the meaning of saved must be determined in each context. The full lexical range of a word must not be poured into every occurrence of it.

This is not scepticism about the doctrine of salvation. Salvation is comprehensive because Scripture describes a comprehensive human ruin and a comprehensive divine rescue — not because every instance of one Greek verb contains the whole of it. A word is not a doctrine. Ignoring that produces a particular kind of error, where a passage about a woman’s healing is read as though it were about her conversion, and the actual thing Jesus did for her disappears.

There is a matching editorial rule: salvation is a category name, not usually an adequate explanation. Saying “this verse is about salvation” tells a reader almost nothing. Each study is required to name the peril, the deliverance, the agent, the means, the result, and the time horizon actually present in the text.

What Is in the Corpus

Phase One includes the union of four things:

Every occurrence of σῴζω (sōzō) in either principal Greek text. Every occurrence of διασῴζω (diasōzō), the compound that describes bringing someone safely through danger. Every KJV use of save, saved, savest, saveth, or saving that denotes rescue, preservation, healing, deliverance, or the securing of life — including the handful produced by a different Greek word entirely. And every material textual difference affecting one of those occurrences.

The search runs in both directions, which matters more than it may sound. Working only from the Greek would miss the places where the KJV says made whole or escaped and the reader never suspects the saving verb is there. Working only from the English would miss the occurrences the KJV renders some other way. Both directions were run, and the results reconciled.

That yields 123 occurrences, grouped into 92 passage studies.

What Is Logged and Excluded

The King James Version also uses save and saving as a preposition meaning except — “no man knoweth the Son, but the Father,” “save that the Holy Ghost witnesseth.” There are 26 of these.

They are not saving vocabulary and they are not part of the corpus. They are nevertheless listed on the occurrence index, so that anyone checking the English-side search can see they were found, examined, and set aside for a stated reason rather than silently dropped.

The Deferred Word-Family

Three related nouns and adjectives are not included in this phase: σωτηρία (salvation), σωτήρ (savior), and σωτήριος / σωτήριον (bringing salvation).

They are inventoried and will eventually receive the same treatment, but including them now would roughly double the corpus and delay the verb study indefinitely. Where one of them appears inside a passage under study — Luke 19:9, for instance, where salvation interprets the verb in the next verse — it is explained in that study, because the passage requires it.

Outside the scope of this phase entirely: exhaustive Old Testament and Septuagint treatment, the Hebrew saving roots, the history of sōzō in extrabiblical Greek, and complete commentary on every surrounding passage. Those are real subjects. They are not this one.

Translation and Greek Texts

Displayed Scripture is the King James Version, in keeping with the established practice of Fireproof Studies. Where the KJV’s rendering obscures something — made whole concealing the saving verb, health concealing the noun — the study says so plainly.

The Textus Receptus is not treated as the controlling Greek text. The principal study texts are the Tyndale House edition of the Greek New Testament and The Greek New Testament According to the Majority Text, second edition, edited by Zane C. Hodges and Arthur L. Farstad.

The Textus Receptus is consulted where it is needed to explain the KJV — the Greek form lying behind a KJV rendering, an occurrence present in the KJV tradition but absent elsewhere, or a construction the KJV specifically reflects.

Each occurrence is tagged for its presence in each text: in both principal editions, Tyndale-only, Hodges-only, TR/KJV-specific, or an English saving word produced by a different Greek lexeme.

How Textual Differences Are Handled

Textual variants are discussed only when they affect the occurrence, the construction, the referent, or the interpretation. This is not a general textual commentary, and a study that turned every difference into a manuscript debate would fail the reader it was written for.

Where a difference does matter, the study states it in a fixed order: what the difference is; which principal texts include or omit the wording; the directly verified Hodges–Farstad evidence where it exists; whether the difference changes the passage’s meaning; and why the occurrence is included in or excluded from the project.

Two cases show the range. At Matthew 18:11, the whole sentence — “For the Son of man is come to save that which was lost” — is present in the KJV, the Textus Receptus, Robinson–Pierpont, and Hodges–Farstad, and absent from the Tyndale House edition. The surrounding parable teaches the same thing either way, but the verse states the Christological ground explicitly. At Revelation 21:24, the phrase rendered “of them which are saved” has narrow Greek support and is absent from Hodges–Farstad; the passage nevertheless describes exactly what the phrase claims, since verse 27 admits only those written in the Lamb’s book of life. That occurrence is labelled honestly as a KJV/TR reading.

Verification against the printed Hodges–Farstad second edition is in progress and is recorded occurrence by occurrence. Where the project relies on a published apparatus rather than a direct check of the printed page, the study says so. Automated Byzantine data used elsewhere in the project comes from Robinson–Pierpont 2005 and is never represented as Hodges–Farstad.

How Passages Are Grouped

A study covers one coherent passage, not one verse. Several occurrences share a page when they belong to the same scene or argument, when one interprets another, or when separating them would produce thin and repetitive pages.

Separate pages are made when the same expression appears in genuinely different literary settings, when a parallel Gospel writer gives a saying a distinct function, when a textual variant needs its own treatment, or when the interpretive question has its own search identity and deserves a complete answer.

Parallel Gospel passages normally receive separate pages, because literary context can change what a saying is doing. They are cross-linked rather than duplicated.

Every occurrence remains individually addressable on the occurrence index even when its explanation is grouped with others.

How Each Occurrence Is Classified

Classification came before exegesis, and it is deliberately faceted rather than a single list of senses. A flat list mixes several different questions and produces categories that overlap and argue with each other.

Each occurrence therefore records its deliverance domain — the ruin overcome and the condition that results — along with its temporal horizon, its agency, any named means or response, and any interpretive flags for irony, typology, warning, disputed construction, textual difficulty, or pastoral sensitivity.

Seven domains emerged from the complete corpus and were tested against twelve pilot studies spanning all of them: bodily healing and restoration; temporal rescue and preservation; historical, corporate, and covenantal deliverance; salvation into life in Christ; continuing preservation and restoration; final salvation, resurrection, and consummated life; and rhetorical, ironic, or false self-salvation.

No eighth domain proved necessary. Difficult passages are explained by combining domain, horizon, agency, and means rather than by inventing ever-narrower senses of the word.

What Every Study Must Do

Each study is required to answer the same governing questions, though not with identical headings — a shipwreck narrative and an epistle argument do not want the same shape.

It must state the answer plainly before any technical discussion. It must name the peril and the resulting condition. It must distinguish the ultimate saving source, the immediate agent, and any human or created instrument. It must use Greek only where the Greek clarifies something, never as display. It must disclose textual differences in proportion to their significance. It must state uncertainty honestly. And it must answer, explicitly, whether the passage refers to comprehensive salvation in Christ — yes, no, or indirectly.

That last question is the principal guard in both directions: against treating every rescue as regeneration, and against treating narrower rescues as though they revealed nothing about the saving work of God.

What This Project Does Not Claim

It does not claim to be a systematic theology of salvation. It does not claim that its classifications are beyond revision — individual assignments may be refined as studies are written. It does not claim that the textual verification is finished; that work is ongoing and its state is recorded rather than hidden.

And until every study is written, it does not claim to be exhaustive. The occurrence index shows exactly how many are complete and how many remain, and that figure is generated from the data rather than typed by hand.

This study is part of What Does It Mean to Be Saved?, an examination of every New Testament use of saving language in context, within Resource Cluster 3: Life in Christ. 24 of 92 passage studies are complete — see every occurrence.