Traditionally, the aim of OT and NT textual criticism has been to determine the original text (or autograph) of the canonical books. However, some contemporary critics have challenged the operating assumption. Take Bart Ehrman:
Assume, for a second, just for the sake of the argument, that chapter 21 and 1:1 — 18 were not original components of the Gospel. What does that do for the textual critic who wants to reconstruct the "original" text? Which original is being constructed? All our Greek manuscripts contain the passages in question. So does the textual critic reconstruct as the original text the form of the Gospel that originally contained them? But shouldn't we consider the "original" form to be the earlier version, which lacked them ? And if one wants to reconstruct that earlier form, is it fair to stop there, with reconstructing, say, the first edition of John's Gospel? Why not go even further and try to reconstruct the sources that lie behind the Gospel, such as the signs sources and the discourse sources, or even the oral traditions that lie behind them? B. Ehrman, Misquoting Jesus (HarperCollins, 2005), 61-62.
There are many problems with that example:
i) Textual criticism is scarcely unique to the Bible. Producing critical editions of Shakespeare's plays involves the same notion–an original text.
ii) It's true that words like "original" or "autograph" are vague without further definition. These are terms of art in textual criticism.
iii) Ehrman's statement is silly and confused. It willfully confounds textual criticism with source criticism. But there's no good reason to conflate the two. A preliminary, unpublished draft is not a different edition of the same book, but an earlier compositional stage. The question at issue is not the stages of composition, or sources (if any) which may have fed into the composition, but the final product. Not a rough draft, but a final draft. The text that the Bible writer issued and intended for popular consumption.
iv) Here's how one textual critic defines it:
When we speak of the original text, we are referring to the "published" text–that is, the text as it was in its final edited form and released for circulation in the Christian community. For some books of the NT, there is little difference between the original composition and the published text. After the author wrote or dictated his word, he (or an associate) made the final editorial corrections and then released it for distribution. Philip Comfort, "Texts and Manuscripts of the New Testament," P. Comfort, ed. The Origin of the Bible (Tyndale, 1992), 183.
If that involves collaboration between the author and a scribe, we could dub the original or autograph the authorized text. Say a scribe takes dictation in shorthand. After he writes it out in longhand, the author reviews the transcript, edits it, the scribe then produces a final draft, incorporating the corrections. If the revision meets with the satisfaction of the author, he signs off on that. If we postulate that the scribe is inspired, then that simplifies the process.
v) This is not an arbitrary definition. The books of Scripture aren't diaries. The author isn't writing to and for himself. Rather, he produces a text for wider circulation. The autograph doesn't consist of his notes, but the edition he intended for popular consumption and issued for general distribution. That's how it was meant to function.
vi) But suppose, for the sake of argument, that there was more than one authentic edition in circulation, or more than one authorized copy in circulation. Why does Ehrman think that would be a problem? They'd all be authentic literary products of the writer. All of them would meet with his approval. They'd all be authoritative. It's not as if the prologue and epilogue of John contradiction what's in-between.
vii) However, we can approach the identification from the other end: Instead of asking, What is the original?, we can ask, What is not the original?
The traditional objective of textual criticism is to strip away changes in the text that were introduced by subsequent scribes. What is unoriginal, what is not the autograph, are changes or variations added to the text by someone other than the author or the author in collaboration with his amanuensis. That's a clear-cut distinction. A principled distinction, not an ad hoc distinction.
To take a secular illustration, consider an editor who tampers with The Adventures of Huckleberry Finn to eliminate the racist slurs, without alerting the reader to his changes.
viii) I'd mention a further qualification. Some changes don't affect the sense. Suppose a 4C scribe spelled a few words differently than the author or his amanuensis. Recovering the autograph hardly requires us to recover the original spelling. It's the same word. I don't think there was standardized spelling back then. It would be mindlessly pedantic to insist that restoring the autographic text requires us to identify and "correct" changes that have no bearing on the semantic content of the text.