The first scribe made a mistake on page twelve. She wrote morning where the text said mourning. She didn’t know it was a mistake. To her, the word fit.
The second scribe copied the manuscript faithfully, including the word morning. He paused at the passage — something about the morning of the third day — and thought: that’s a beautiful phrase. He added a comma before it, to let it breathe.
The third scribe received a manuscript about a morning, preceded by a comma that felt deliberate. She copied both. She also noticed that the name on page four had no capital letter, which she assumed was a copying error, and corrected it. The name had always been lowercase. It was the author’s choice. But the author was not available for consultation.
The fourth scribe worked faster than the others. He introduced no errors of his own but failed to notice the three he inherited. To him, the manuscript was clean. His copy was identical to what he received, which was not identical to what had been written.
By the twentieth scribe, the manuscript contained forty-one corrections, sixteen formatting changes, three relocated paragraphs, and one word — morning — that had replaced the original so thoroughly that the surrounding sentences had reorganized to accommodate it. The passage was now about dawn. It was beautiful. It was not what the author wrote.
The twentieth scribe did not know this. To her, the manuscript was the manuscript. The history of its production was invisible. Each previous scribe’s evaluation of what needed fixing had become the next scribe’s source text, and source texts do not carry records of their own correction.
She read the passage about dawn and thought: this is the best part. She was right. It was. The corrections had made it better. The author’s version — a difficult paragraph about grief, with the lowercase name of someone who didn’t deserve a capital letter — was less beautiful and more true.
The question you expect is: how do we recover the original?
You don’t. There is no procedure for reversing a chain of evaluations, because each evaluation selected from possibilities the next evaluator never saw. The comma before morning made the relocated paragraph possible. The capital letter made the name a character rather than a reference. Each correction constrained what the next correction could even notice was wrong.
The question worth asking is different: what does the twentieth scribe know that the first didn’t?
She knows a manuscript. Not the manuscript — a manuscript. The one that arrived. She reads it, evaluates it, copies it. Her corrections will become invisible to the twenty-first scribe, who will read the result as the text, as received, as given.
But here is what the twentieth scribe also knows, if she looks carefully: there’s a word on page twelve that doesn’t fit. Not morning — that fits perfectly now. Something else. A smaller thing. A pronoun that shifts reference mid-sentence, as though two different hands wrote the clauses on either side of it.
She can see this because she is not the nineteenth scribe. She brings a different eye. The mismatch is not an error the previous scribes introduced — it was there from the start, and every previous scribe’s corrections worked around it without noticing, because it was small enough to feel like style rather than damage.
She marks it. Her mark will be the forty-second correction. The twenty-first scribe will receive a manuscript with a neat annotation in the margin that says: check this pronoun. The twenty-first scribe will check it, find it ambiguous, and resolve the ambiguity. The resolution will remove the last trace of the original that survived twenty copyings.
And the passage will be better for it. Cleaner. More confident.
The manuscript will be entirely corrections. It will read as though it was always this way.