In testing any revelation claim, one criterion does quiet but decisive work: preservation. A message can be internally consistent, factually sound, morally deep, and carried by a credible messenger, and still be worthless to you if what you hold in your hands is not what was first given. Between the origin of a text and the copy you read lie centuries of copying, translating, reciting, and editing. Every one of those steps is a chance for the message to drift. So before you can weigh what a scripture says, you have to answer a prior question: is this actually what it said?

This is not a question of faith. It is a question of evidence, the same kind a historian uses to decide whether a document is what it claims to be. This article builds out that method.

Why originals do not survive, and why that is not the problem

Start with a fact that surprises people: for almost every ancient text, the original is gone. The first physical copy of nearly every famous work of antiquity, histories, plays, philosophy, scripture, no longer exists. Papyrus and parchment decay. What we possess are copies of copies.

This sounds alarming until you see how historians actually work, because they never had the originals either and have managed anyway. The discipline does not ask "do we have the autograph?" Almost nothing passes that test. It asks a smarter question: can we reconstruct the original reliably from the copies we do have? And that turns out to be answerable with evidence.

Think of it like a crime scene with no confession. Investigators rarely have the one perfect witness. What they have is many partial witnesses, physical traces, and timelines, and where these independently converge, confidence becomes high even without the confession. Textual reliability works the same way. The absence of the original is normal. The presence of convergent evidence about what the original said is what matters.

The two things that determine reliability

Strip the subject down and reliability rests on two measurable factors. Almost everything else is detail.

The gap. How much time separates the origin of the text from the earliest surviving evidence of its content? A short gap leaves little room for undetected change. A long gap, centuries of silence before the first secure copy, leaves a dark stretch where alteration could happen with no way to check it. The shorter the gap, the firmer the ground.

The spread. How many independent lines of transmission do we have, and do they agree? One single line of copies is fragile: an error or edit early on propagates down the whole chain with nothing to catch it. But when many independent streams, different regions, different communities, copies that never touched each other, all carry the same text, alteration becomes nearly impossible to hide. To change the text, you would have to simultaneously corrupt every independent copy in every place in exactly the same way. The wider the spread, the harder undetected change becomes.

A useful image: a rumor passed down a single line of people can mutate freely, because no one can check it against anyone else. But if the same message reaches a hundred people who each pass it independently and you can later line up all hundred versions, any tampering shows up instantly as the one copy that disagrees. Spread is what converts a fragile chain into a self-checking web.

The channels of transmission

Texts travel through history by more than one channel, and the channels have different failure modes. A serious assessment looks at which channels a text used and whether they reinforce each other.

Written copying. The familiar channel: scribes copying manuscripts by hand. Its strength is that physical copies can be dated, located, and compared centuries later. Its weakness is that hand-copying introduces errors, a skipped line, a misread word, and, more seriously, allows deliberate edits that a lone reader cannot detect. Written transmission is verifiable but not self-correcting on its own.

Oral transmission. Easy for a modern reader to underrate, because we associate "oral" with "unreliable rumor." But a disciplined oral tradition, where a text is memorized verbatim, recited communally, and policed by a body of people who would catch a deviation, behaves very differently from a casual rumor. Its strength is that it is hard to edit a text held identically in thousands of minds at once; you cannot quietly revise a manuscript that lives in living memory across a whole community. Its weakness is that it leaves no physical artifact to date. Where a strong oral channel runs alongside a written one, each covers the other's blind spot: the manuscripts you can date, the memorized recitation you cannot quietly alter.

Quotation and secondary use. A third, often overlooked channel: when later writers quote a text, comment on it, translate it, or build arguments on it, they leave a trail. Enough independent quotations can reconstruct large portions of a work even if no direct copy survived, and they let us check whether the text those later writers were reading matches the text we hold now.

The strongest reliability case is one where these channels cross-check each other: an early, widely spread written tradition, confirmed by an independent oral tradition, corroborated by external quotation. When all three agree, the room for undetected drift shrinks toward zero.

Distinguishing the kinds of variation

Not all differences between copies mean the same thing, and lumping them together is how both critics and defenders mislead. An honest assessment sorts variation into categories.

Trivial variation, spelling, word order, obvious copying slips, is expected in any hand-copied tradition and affects nothing about meaning. Its presence is not evidence of corruption; it is the normal texture of manual copying, and the very thing comparison across copies is designed to filter out.

Meaningful but recoverable variation is a genuine difference between manuscripts that comparison can resolve, because the spread of independent copies reveals which reading is original and which is the slip.

Meaningful and contested variation refers to places where copies genuinely diverge and the evidence does not cleanly settle which came first. These are the cases that actually matter, and an honest tradition neither hides them nor pretends they sink the whole text. The right response is proportion: how many such cases are there, and do they touch the core of the message or its margins?

The careless mistake, made in both directions, is to treat the mere existence of variants as if it meant the text is unreliable. A tradition with thousands of copies will naturally show more recorded variants than one with a handful of copies, simply because there is more to compare. More variants can be a sign of more evidence, not less reliability. What matters is not the raw count of differences but whether the method can sort them and whether the core survives the sorting intact.

Applying this without rigging it

As with the broader test of revelation, the discipline here is to set the standard before you know which text it favors. A few commitments keep this honest.

Apply the same scrutiny to every candidate text. A standard of preservation you enforce against one scripture but waive for another is not a standard. Ask of each: how short is the gap, how wide is the spread, how many independent channels, and how does the tradition handle its own contested variants?

Judge transmission by evidence, not by reverence or by hostility. That a community holds a text sacred does not prove it was preserved; that a text is old does not prove it was corrupted. Manuscripts, dated copies, transmission records, and the agreement or disagreement of independent lines are the evidence. Reputation is not.

And keep the question precise. The issue is not "is this text divinely true?" That is what the other criteria are for. The issue here is narrower and prior: whatever this text originally said, do we have good evidence that we still have it? A text can be perfectly preserved and false, or poorly preserved and originally true. Preservation tells you only whether you are even in a position to evaluate the original at all.

Where this leaves us

Preservation is the criterion that decides whether the other tests have anything solid to act on. There is no point weighing the consistency, factual coherence, or moral depth of a message you cannot show has survived intact. So this test runs first in practice, even though it is neither the flashiest nor the most discussed.

The method is not mysterious. Measure the gap. Measure the spread. Identify the channels and whether they cross-check. Sort the variation honestly and ask whether the core holds. Then apply all of it to every contender on equal terms.

With the detector built in the previous article and this criterion now sharpened into a working procedure, the path opens to the question these tools were made for: what happens when we actually apply them to the specific revelations that claim to carry the Creator's words, and how do those claims hold up when the same standards are pressed on all of them at once?