6 points on quotes by era misquotes that matter

Ask which edition and most quotation arguments dissolve: two people quote the same work accurately yet disagree about wording, because they hold different books.

Authors revise, publishers reset type and add errors, editors modernize spelling, and translators are replaced between printings.

None of that is unusual, and all of it means a work does not have one text.

The discipline that handles this is old and well developed, and quotation checking can borrow most of it without becoming scholarship.

The six points

Why one work has several texts

Author revision. Writers change their minds and their sentences. A line quoted from a late edition may not exist in the first, and the reverse. Both are the author's words, at different times, which is a real difference rather than an error.

Resetting. Every time type was reset for a new printing, new errors entered. Before photographic reproduction this happened constantly, and the errors are usually small and occasionally significant.

Editorial modernization. Spelling, punctuation, and capitalization get updated for later readers. The sense survives; the wording you are quoting is now partly an editor's.

Expurgation. Passages removed for a market or a period, sometimes silently. Quoting the shortened edition can misrepresent an argument by omission.

Retranslation. For a translated work, a new translator produces an entirely new set of English words.

What the scholarly tradition offers

Textual criticism exists to work out the relationships among surviving versions of a text and to justify a reading rather than assert it. Two of its habits are worth stealing outright.

The first is the critical apparatus: the notes showing which witnesses support which reading, so a reader can disagree with the editor. A quotation entry can do a miniature version in three lines.

The second is the variorum approach, which prints the variants alongside each other rather than choosing between them. For a contested quotation, listing the versions with their sources is more honest and more useful than presenting one as correct.

The practical procedure

Both the Chicago Manual of Style and the MLA Handbook expect a book citation to name the edition and the year, so this habit costs nothing extra.

A recognized first printed edition is a fixed point everyone can refer to. Scholarly citation leans on it. It is a fixed point, not an authority: the author may have preferred a later reading, and a first edition can contain errors the author later removed.

A quotes by era source check can bound a wording even when the original printing is lost.

Deciding which version to quote

The second row is often overlooked and is frequently the right answer. If you are writing about how an argument was received, the text that matters is the one people actually read, even where a better text exists now.

Where this shows up as a misquotation

A great many alleged misquotations are edition differences that nobody checked. Someone quotes accurately from one printing; someone else checks against another and finds a discrepancy; the line is declared corrupted. Both are quoting real text.

The tell is that the variants are small and sensible: a changed conjunction, a modernized spelling, a tightened clause. Genuine transmission damage looks different, and our catalog of how lines get damaged in circulation sets out the patterns. Edition variants sit inside the work. Transmission damage happens outside it, in retelling.

Translated material is the messiest case, because edition differences and translation differences compound. Our material on why translated wording belongs to the translator covers the second half of that problem.

Misquotes that actually matter

Not every disputed line is an edition difference. Some famous lines were never said in the form we quote them. Casablanca (1942) contains no "Play it again, Sam"; Ilsa's line is "Play it, Sam." Arthur Conan Doyle never wrote "Elementary, my dear Watson." Marie Antoinette has no contemporary source for "let them eat cake." The King James Bible reads "the love of money is the root of all evil" in 1 Timothy 6:10, not "money is the root of all evil." The Empire Strikes Back (1980) gives "No, I am your father," which audiences remember as "Luke, I am your father."

The shelves to check a line like these are Bartlett's Familiar Quotations, first published in 1855 and revised through many editions since, the Oxford Dictionary of Quotations, and The Yale Book of Quotations. Each is a record of wording and attribution as it stood when that edition went to press.

What to publish

State the wording you are using, the edition it comes from, and any variant you know of, with its source. Three lines. That converts an assertion into something checkable, and it makes you immune to the most common form of correction, which is somebody holding a different printing.

Our labels for describing what evidence actually supports give a category for a line with a real source and unstable wording.

Our account of how collecting practice changed across periods explains why older anthologies so often print a wording matching no edition of anything.

Common questions

Which edition should I buy, if I only buy one?

A scholarly edition with an apparatus and notes, where one exists. In practice that means an Oxford World's Classics or Norton Critical Edition volume, a Library of America volume for American writers such as Whitman, or the Loeb Classical Library, which prints the original Greek or Latin facing an English translation. It costs more, and it answers questions the cheap reprint cannot, including which readings are contested.

Is a modernized spelling edition acceptable to quote?

Yes, with the edition named. It becomes a problem only when the modernization is presented as the author's own orthography, which matters in some kinds of argument and not in others.

What if my source does not say which edition it used?

Treat the wording as unverified and go to a text yourself. An unsourced wording from an unnamed edition is exactly the input that produces the disputes described here.

How can I tell whether a difference is a variant or an error?

Look at how many witnesses carry it and how early they are. A reading present in early printings across several places is a variant. One that appears in a single later printing is probably an error introduced when type was reset.