There are no originals. Everything we call the New Testament is reconstructed from copies of copies - an embarrassment of riches with one stubborn shadow at its source.
Not one autograph - no page in Paul's hand, no first copy of Mark - survives. What survives instead is volume: more than 5,800 Greek manuscripts, tens of thousands in Latin and other languages and a dense web of quotations in the early church fathers. The task of textual criticism is to work backward from all of it toward the wording that stood at the head of the chain.
That abundance is the good news and the catch at once. So many copies means errors can be triangulated and caught - but it also means the text exists only as a tradition of copies, never as an artefact. Here is the chain those copies descend through.
Copies made in the same region inherited the same habits and the same errors, so the tradition sorts into recognisable families. Knowing which family a reading belongs to is half of weighing it.
The oldest and generally most disciplined stream. Shorter, harder readings - the backbone of the modern critical text.
Free, expansive, paraphrasing. Most dramatic in Acts, which it runs ~8-10% longer. Early but loosely copied.
The smoothed, fuller, harmonised text that came to dominate by sheer numbers - and the stream the Textus Receptus drew from.
The reason a King James Bible carries the long ending of Mark, the Trinity formula in 1 John and "book of life" in Revelation - all as plain scripture - traces to one rushed printing. Erasmus worked from a few late Byzantine manuscripts, filled a gap in Revelation by translating the Latin back into Greek and added the Trinity formula only under pressure. That text hardened into the Textus Receptus and the Textus Receptus became the KJV. Four centuries of English Bibles inherited the readings of a single 1516 edition - which is why the modern translations can look like they're "removing" verses when they're really just declining to follow that printed text.
For all the abundance, the chain has one irreducible weakness and it sits right at the top. The books were written in the first century; our first substantial copies date from around 200 and the great codices from the 300s. The gap between the autograph and the earliest copies we can actually read - roughly a century and a half - is precisely the window in which the text was copied most freely and edited most actively. By the time the manuscript trail begins, the text is already largely settled.
The argument that stands is both narrow and solid: from the third century forward, the text stabilises and the variants become traceable. Collation fails to illuminate the distance between the original authors and the earliest surviving copies. That unlit stretch is not a disconnection in the narrative; it is the narrative's defining edge - and more manuscripts, no matter how many or how early, cannot illuminate it.