How this book checks its quotations
Famous quotations are the ones most often wrong. The more often a line is repeated, the less often anyone goes back to the page. It gets shorter, or smoother, or moves to a more famous name, and after a while everyone simply knows it.
A book about the Lord’s Prayer quotes a great many people: the Gospels, the church fathers, monks and reformers, saints, and philosophers. If even one of those quotations is invented, the reader has no reason to trust the others. The book follows one rule, and I’d like to explain it here.
Nothing in Words Given is quoted from memory—not a sentence, not a date, not a name. Every quotation is checked against the text itself, and the result is written down: which edition, which page or section, and the exact words.
Each source has its own file in the book’s research, and each quotation in it carries a status. Checked against the original text is the best. Checked in a reliable secondary work is allowed when the original can’t be had, and then the book says ‘quoted in’ and names that work. The third status is unverified, and unverified material never reaches the book, however famous it is.
Old texts in the public domain are downloaded and kept with the research, so a quotation can be checked against the stored text and its web address. Books still under copyright may be quoted only once a passage has been checked against a copy that’s been bought. Then every factual claim in the readings gets its own row in a ledger, with its source and its status. Today that ledger has nearly a thousand rows.
The book was drafted with an AI model, working under my direction. I say that on this site and in the book, and it’s also the main reason for the rule.
A model writes fluent sentences, and a fluent sentence can carry a quotation that sounds exactly right and isn’t. People do the same, of course; that’s where false quotations come from. But I didn’t want the reader to have to trust either of us. The model isn’t allowed to remember a quotation. It has to find the text, and the text has to be stored where anyone can look at it again.
The rule caught things, and the corrections stay in the ledger as warnings.
The first is Kierkegaard. Two lines that circulate under his name proved real, though one is usually quoted without its setting. A third, that prayer can only be learned by praying, isn’t his at all. I wrote a separate essay about him: What Kierkegaard said about prayer, and what he didn’t.
The second is the hour. The claim that Teresa of Ávila prescribed an hour for a single Our Father is wrong as a rule. In The Way of Perfection, written for her sisters in the 1560s, she describes people in a deep, quiet prayer who find speaking hard: ‘they spend an hour in saying one Pater Noster’ (Teresa, Way of Perfection 31.2). A few chapters earlier she tells her sisters they ‘need not repeat it more than once in a whole hour’ (29.5), where ‘it’ is the Our Father. One describes, the other permits. Neither sets the clock. The rule belongs to Ignatius of Loyola, in his Spiritual Exercises: the one praying ‘will be an hour on the whole Our Father’ (Ignatius, Spiritual Exercises 253, the first rule of his second method of prayer). Two saints of the same century, two true sentences, and the rule went to the wrong one.
The third case is the one I learn most from, because the checking corrected the research itself. People often say that ‘trespasses’ in the Lord’s Prayer comes from William Tyndale, who translated the New Testament into English in the 1520s. An early note in the book’s own research said the opposite: that Tyndale’s text of the prayer didn’t have the word. When the reading on ‘trespasses’ was drafted, his 1526 text was checked again. It does have it: ‘And forgeve vs oure treaspases/ even as we forgeve them which treaspas vs’ (Tyndale, Matt 6:12; the slash is his own punctuation). The note was wrong, and it was corrected.
What isn’t Tyndale’s is the line English-speaking churches still pray, ‘as we forgive them that trespass against us’. That wording appears in a primer, a book of prayers in English, printed in 1535, and in the official books of the following years, up to King Henry’s Primer of 1545 (Burton; Lloyd). From there it went into the first Book of Common Prayer in 1549. The popular story is half right, and the first note was wrong. The checking found both.
I don’t want to claim more than is true.
The rule means that some things couldn’t go in. Where a sentence depended on a book that isn’t yet bought, it was cut, and the old wording is kept aside in the research until the book is there to check it against. The translations from Kierkegaard’s Danish in the book are my own, partly because the pages of the standard English edition haven’t yet been checked in a bought copy. A check is also only as good as the text it was made against. Tyndale’s 1526 text, for example, was checked in a reprint of 1836 that calls itself verbatim, not in one of the surviving copies of 1526.
Much of the checking was done by the same AI model, working from the stored texts, and every source file records what was checked, against what, and when. That record is the point. In the book, every note names the edition and the page, so you don’t have to take my word, or the model’s. You can look it up.
Sources
The Book of Common Prayer. 1549. Matins, transcribed by Charles Wohlers.
Burton, Edward, ed. Three Primers put forth in the reign of Henry VIII. Oxford: University Press, 1834. Spelling modernized. The primer of 1535 and King Henry’s Primer of 1545.
Ignatius of Loyola. Spiritual Exercises. Translated by Elder Mullan. New York: P. J. Kenedy & Sons, 1914.
Lloyd, Charles, ed. Formularies of Faith put forth by Authority during the Reign of Henry VIII. Oxford: University Press, 1856. Spelling modernized. The Institution of a Christian Man (1537) and A Necessary Doctrine and Erudition for any Christian Man (1543), with the same wording.
Teresa of Ávila. The Way of Perfection. Translated by the Benedictines of Stanbrook, revised by Benedict Zimmerman. London: Thomas Baker, 1911. Her first version was finished before April 1567.
Tyndale, William, trans. The New Testament … published in 1526 … Reprinted verbatim. Edited by George Offor. London: Samuel Bagster, 1836.