The new revised international KJV-ASV-DRB-DBT-ERV-WBT-WEB-YET-AKJV-WNT version
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


The new revised international KJV-ASV-DRB-DBT-ERV-WBT-WEB-YET-AKJV-WNT version of Ecclesiastes.

Each sentence in this translation of Ecclesiastes is drawn uniformly at random from one of ten translations of the Bible on every refresh. Study the text. Learn it. Love it. Live it. Refresh the page and you'll never see it again.

Put slightly differently: this book is sampled from the distributions of possible understandings of Ecclesiastes, like a more systematic Septuagint. You haven't really understood the text until you've studied the nuances of each of its variations.

Made for National Novel Generation Month 2017, Ecclesiastes (KJV-ASV-DRB-DBT-ERV-WBT-WEB-YET-AKJV-WNT) runs to 10k words per translation, but as it draws its strength by consuming ten (slightly) different translations of the Bible, the hyperobject weighs in at well above the threshold of 50k.

Much thanks to the King James Bible, American Standard Version, Douay-Rheims Bible, Darby Bible Translation, English Revised Version, Webster Bible Translation, World English Bible, Young's Literal Translation, American KJV, and Weymouth New Testament. (Someone on Twitter noted that not only are there many different translations, there are also many different source texts, which is an important complication to keep in mind while reading!)

Different translations are optimized for different qualities, such as "thought-for-thought" or "word-for-word." Here are three comparisons (1, 2, 3) and a page on the Douay-Rheims Bible.

I'd be curious to see a quantitative study on the qualities of the translations, such as where the mean words lie in an embedding space.


bible.hs is a Haskell file that generates the HTML for the page. You'll need to manually copy the Javascript at the top and bottom of the current bible.html into the generated bible.html, and edit the generated verses again.

bible.html contains the js to randomly sample verses from the different translations.

This repository contains the original corpus for the book, ecc.txt, as well as ecc-nl.txt if you want to just read all the translations at once. Credit for the full corpus—a truly amazing Excel spreadsheet of ten Bibles—goes to the site, specifically to Biblehub. (N.B. There is a parse error in the corpus in Exodus.)

This book is made out of a spirit of respect and love, not sacrilege.