He never wrote it. The sentence is younger than the poster paint it usually arrives in: it was born in 1906, in English, in a biography — and its own grammar says it is not a quotation.

Evelyn Beatrice Hall, writing as S. G. Tallentyre in The Friends of Voltaire, is describing how Voltaire took the persecution of a friend’s book. Here is her sentence, with the ending the posters leave off:

“‘I disapprove of what you say, but I will defend to the death your right to say it,’ was his attitude now.”

The occasion deserves better than a fake attribution, because the true story is stronger. In 1759 Helvétius’s De l’esprit — On the Mind, in Hall’s rendering — was “burnt by the public hangman.” Voltaire had “soundly hated (and roundly abused)” the book. He defended it anyway: in Hall’s account he “fought for its right to live, tooth and nail, up hill and down dale.” She reports him exclaiming, when he heard of the burning, “What a fuss about an omelette!” A man defending a book he hated, for the principle alone — that is the only kind of defence that proves anything, and it needs no invented sentence to dress it.

Hall watched it happen and said so, in print. In May 1935, in a letter to the Saturday Review: “I did not mean to imply that Voltaire used these words verbatim and should be surprised if they are found in any of his works.” The author corrected the record. The record declined the correction.

So credit both, because both earned it. The sentence is hers; the stand was his. Hall wrote the best one-line summary of intellectual tolerance in the language, and Voltaire lived the thing it summarizes. Quoting him with her words shortchanges the pair of them — the writer loses her sentence, and the man loses the harder, truer story of defending what he despised.

A quotation is a claim with a receipt attached; when the receipt names a different author, so does the quotation. Which famous quotation should we put on trial next?

The word travels the internet in two costumes. In the first it means “beautiful thinking” — eu, well, plus nous, mind, read as a compliment to the mind’s owner. In the second it is trivia: the shortest English word, the listicles say, to use all five vowels. Both costumes are charming. Neither is Aristotle’s.

Aristotle is doing something less flattering and more useful with it. Opening Book II of the Rhetoric, he lists the reasons we believe a speaker “apart from any proof of it”: “There are three things which inspire confidence in the orator’s own character” — and Roberts’s translation names them plainly: “good sense, good moral character, and goodwill.” The Greek behind the plain English is phronesis, arete, eunoia. Then he shows each one failing, and the third failure is the interesting one:

“…or finally, they are both sensible and upright, but not well disposed to their hearers, and may fail in consequence to recommend what they know to be the best course.”

The word at work, today: a pitch that opens with the buyer’s risk instead of the seller’s slides is eunoia. A lecture that starts from what the room already knows is eunoia. An apology that names the other person’s cost, not the speaker’s embarrassment, is eunoia. None of these is beautiful thinking. All of them are goodwill made visible — which is why the audience can tell, and why it works.

Keep the vowels; they are free. But when you use the word, point it the way Aristotle did — outward. When did you last believe someone because they meant you well?

A printing house in Boston. Benjamin Franklin — sixteen, and four years into an apprenticeship he signed at twelve — sets type for his brother by day and teaches himself to write by night. “About this time I met with an odd volume of the Spectator,” he remembers. Met with, then bought. “I bought it, read it over and over, and was much delighted with it.” He thought the writing excellent and “wished, if possible, to imitate it.”

What he did next is the reason this page exists. His method was mechanical to the point of comedy: reduce each sentence of an essay to a hint, put the hints in a drawer, wait for the prose to fade, then rebuild the essay from the hints alone and grade the rebuild against the original. A hint is not a copy. It is the sentence’s idea in a few flat words of your own — close to what a reader now would call a one-line summary — and that distinction carries the whole machine, because a copy preserves the wording and a hint deliberately loses it.

Then the grading. “Then I compared my Spectator with the original, discovered some of my faults, and corrected them.” Slow down over that sentence, because everything self-taught writers usually skip is inside it. He did not compare himself to the Spectator in the aggregate and resolve to do better. He set his page beside the page, line against line, and let the difference name the exact move he did not yet own — this verb, this transition, this order of clauses. The gaps were not failures. They were a syllabus written in his own hand.

The rebuilds kept exposing the same shortage. “I found I wanted a stock of words” — wanted, in the old sense: lacked. So he made the drill harder. He took some of the tales — the Spectator’s stories — turned them into verse, waited until the prose had faded, and turned the verse back into prose. Meter is a synonym engine. It keeps demanding “words of the same import, but of different length” “to suit the measure,” or of a different sound for the rhyme — and the constant hunt fixes the variety in the mind, where it is ready the next time a sentence needs it.

The grading stayed honest to the end. Comparing his work with the original, he “discovered many faults and amended them.” But sometimes he “had the pleasure of fancying” that “in certain particulars of small import” he had been “lucky enough to improve the method or the language.” Fancying — believing, not proving; even the daydream is filed as a daydream, and the original keeps the last word. Yet it was enough. The fancy “encouraged me to think I might possibly in time come to be a tolerable English writer, of which I was extremely ambitious.” A tolerable writer. The ambition is exactly as modest as the method is ruthless, and the pairing is the whole lesson: hold the standard next to your page, not above it.

None of this happened in school hours; his schooling had ended years before the indenture. The drill ran in the margins of an apprentice’s day — at night after work, in the morning before it, and on Sundays, when he contrived to be alone in the printing-house with the borrowed quiet.

He engineered those margins as deliberately as the drill itself. At about sixteen, converted to a vegetable diet by a book — of course by a book — he proposed to feed himself for half of what his brother paid for his board, and “presently found that I could save half what he paid me. This was an additional fund for buying books.” The meal hour became the other classroom: while the rest went off to eat, he stayed in the printing-house alone, dispatched a light repast — bread, a handful of raisins, a glass of water — and had the room and the hour to himself. He even logged the return on the arrangement, crediting his progress to the “greater clearness of head and quicker apprehension which usually attend temperance in eating and drinking.” It is the same signature as the drawer: find a fixed cost, redesign it, collect the difference in pages.

Notice, too, what the method refuses to protect: the ego. Most writing practice is arranged — by instinct if not by intention — to avoid exactly the moment Franklin scheduled, the one where your page and a better page sit side by side with the differences legible. General advice is comfortable because it is general. The comparison is uncomfortable because it is specific: not write more vividly but here, in this sentence, the original chose a verb and you chose three adjectives. A loop that never surfaces an error cannot correct one. Franklin’s loop surfaces one per sentence, and makes him fix it before it lets him feel finished.

The method survives untranslated, and it costs nothing but the wait. It needs one writer you would steal from, one piece, and a drawer. Mark the piece. Copy out its hints — one line for each sentence’s idea, none of its wording. Wait until the prose fades; days, not minutes; the forgetting is the ingredient, not the obstacle. Rebuild the piece from the hints, at full length, in your own words. Then set your page beside the original and read the gap slowly. The gap does the teaching.

If the wait feels like the step worth skipping, notice that it is the only step that cannot be faked. The drawer is doing what willpower cannot: it deletes your access to the wording while leaving the idea intact, so that what comes back is composition, not recall wearing composition’s clothes.

Two and a half centuries of writing advice have not improved on the drawer. Most of it exists to spare you the comparison — the one step Franklin refused to skip. What would you rebuild first?