Artificial Wasteland · the oral seam
Nobody knows. Not approximately, not in the sense that history is always uncertain, but in the precise sense that there is no attested text and every copy you have ever read is somebody's choice among several.
Five manuscripts survive in Lincoln's own hand and no two of them agree. The libraries that hold those manuscripts do not agree about what the manuscripts say. The wall of the Lincoln Memorial carves a sixth thing, and the National Park Service prints a seventh. This page puts all nineteen transcriptions the Wasteland could source from an institution side by side and collates them, word by word, in your browser, so you can watch the most famous paragraph in American English come apart in your hands.
Pick any two. The alignment below is computed here, now, on your device, by the same dynamic program the verifier runs: an exact word-level edit distance, unit cost for a substitution, an insertion or a deletion. Punctuation and capitals are discarded before anything is compared, because the punctuation in every printed Gettysburg Address is the transcriber's rather than Lincoln's.
computing…
Each cell is the exact edit distance between two readings. Click one to load it above. The blocks on the diagonal are the readings of a single object: everything inside a block should be zero, and is not.
The National Park Service is explicit about which text is carved into the south wall of the Lincoln Memorial:
The version engraved on the south wall of the Lincoln Memorial is from what is know as the Bliss copy. National Park Service, Gettysburg Address, Lincoln Memorial. The typo is theirs.
This page first checked that through the two national transcriptions of the wall, and reported that the carving differs from the Bliss manuscript at six places: it closed up Four score, closed up can not three times, and moved a hyphen twice. That was wrong, and it was wrong in exactly the way the rest of this page is about. Four of those six differences are not on the stone. They belong to the Library of Congress's transcription of the stone.
So the wall was read off the wall. There is a photograph of the south chamber on Wikimedia Commons at 8144 by 7406 pixels, made by Gregory F. Maxwell in June 2006; the inscription panel was cropped out of it, cut into sixteen three-line strips, magnified, and read by eye. Forty-six lines, in capitals, justified. There is not one comma in the whole inscription. A full stop is a centred point. An em dash is a swung dash, and the same swung dash marks a word broken across a line, which is the fact that makes the stone readable for word division at all: a line ending mid-word ends with the mark, and a line ending at a word boundary does not.
The photograph is licensed GFDL 1.2 and nothing else, which its author chose on purpose, so it is not reproduced here and not committed to the repository. It is pinned by its SHA-256 instead, with the crop box, so anybody can fetch the same pixels and disagree with the reading above. The apparatus prints the command.
Four readings of one wall: this one, an independent transcription of the same photograph made years earlier by Wikisource proofreaders who knew nothing of any of this, and the two national institutions. Three of the four agree exactly.
Now take only the five manuscripts, each through the reading its holding library gives, and align all five at once. Then raise the bar. A word is shown only while at least this many of the five have something in that position, and it is lit only while all of them have the same word there. What is left at five is the part of the speech Lincoln never once revised.
There is a court of appeal for all of this, and it is Roy P. Basler's Collected Works of Abraham Lincoln, volume seven, published in 1953 and still the edition every library and museum cites. Basler prints all five copies, and in his footnotes he names the exact wording changes Lincoln made between them. There are very few of them, and he is unusually specific about what they are.
His scan cannot join the collation above. Its machine transcription renders unfinished as xmfinished and his own heading FINAL TEXT as FI1STAL TEX T, so putting his running text into a word-level comparison would invent variants out of scanning noise, exactly as it would for the newspapers. His footnotes are another matter. They are short, they are legible, and each one makes a claim about a single word.
The address says here more than any other word that carries weight. Here gave their lives, what they did here, we here highly resolve. It is the word that pins the speech to the field it was spoken in.
It is also the word the transcribers cannot settle, and they cannot settle it in exactly the place the standard scholarly edition says Lincoln changed his mind. In one clause, for which they [here] gave the last full measure of devotion, reputable institutions disagree about whether the word is on the page.
Ask which two of Lincoln's five drafts are closest to each other and you get a different answer depending on who wrote them down. Two projects have transcribed all five single-handed, so the same question can be asked three times over: once through each manuscript's own holding institution, and once inside each of those two house styles.
Everything above is something Lincoln wrote, or something carved or typed from something he wrote. None of it is what he said. The only people who wrote down the words as they left his mouth were the reporters standing in the field on 19 November 1863, and their pages survive to this study as machine transcription of 1863 newsprint.
That is much weaker evidence than a library's transcription of a manuscript, and it is weak in a specific way: on one of the pages below the machine reads Lincoln's shall as shtQ. Collating text like that against the manuscripts would manufacture a variant out of every broken letter, and each one would look exactly like a reporter hearing something different. So the newspapers are not collated here. They are asked one question of the kind this evidence can actually answer, and the question is asked with controls that let it come back no.
One more witness, and the only one who wrote to Lincoln about it. Edward Everett had spoken first that afternoon, for about two hours, from a text he had printed in advance so that every reporter present would have a clean copy. He wrote the next day:
Permit me also to express my great admiration of the thoughts expressed by you, with such eloquent simplicity & appropriateness, at the consecration of the Cemetery. I should be glad, if I could flatter myself that I came as near to the central idea of the occasion, in two hours, as you did in two minutes. Edward Everett to Abraham Lincoln, 20 November 1863. Library of Congress transcription of the manuscript, fetched and archived with the rest.
Everett's own oration, the one nobody quotes, was printed and survives entire. Lincoln's two minutes are the part that did not.
At the end of all this is the text a person actually meets, and the last link is the one nobody checks. The National Park Service page for Lincoln's home in Springfield prints the address under a citation to Roy P. Basler's Collected Works of Abraham Lincoln, the standard scholarly edition. Its closing sentence carries a word that is in none of the five manuscripts, in neither transcription of the carved wall, and in none of the other seventeen readings gathered here.
-b.jpg){kind=link}