Three ways in to this site. The Pacioli work on this site was created with 100% AI under the author’s direction and control: several models read the pages, their work was compared, and disagreements were judged, by machine, against the page images. They can still be wrong, so the “How it was made” pages say what has been checked and what has not. This was a five day project.
Not sure where to start? Read the Project Status page first. It is one page and says what exists, what has been checked and what is still open.
Understand debits and credits
The Immutable Ledger (Draft 20, 2026)An article by me, the human in all of this: the 15th century story and explanation are my own while the passages that quote Pacioli are kept in line with the project’s current AI translation. My little story explains double entry using common sense.
The Immutable Ledger (1914 Historical Edition, Draft 15)An earlier draft of the same article. Its quotations have not all been checked against the books. There was a major change of method between drafts 15 and 16, so this draft 15 marks that spot.
Read what Pacioli wrote
English version (machine translation)A literal, line-by-line English of the bookkeeping treatise, with doubts marked. Not proofread by a person, which is my purpose in all of this, to discover what AI can do at this time. So this is multiple major AIs working together, from Anthropic, OpenAI, Google plus Openclaw and Kraken a specialised ancient text neural net engine. That last is a balance with different tech, but still machines only. So the machines debate among themselves, because they never agree while always confidently presenting themselves as right.
The 1494 text, as read from the 300dpi Library of Congress scan of the printed original.The original 15th-century Italian with its abbreviation signs, line by line. For specialists and for checking the other versions. So this work does not stand on the shoulders of another, it starts from scratch or as close as we can get to it as the handwritten original has never yet been found. This means we are at the mercy of the 15th century printers. I say plural because I have used the later 1523 version where the printer expanded the abbreviations as a control check likewise reverse engineered Geijsbeek’s abbreviation expansions back to their abbreviations to see if they match, so we have various double checks while not using the other sources as our source - as shown where we disagree.
Key to the little signs in the 1494 printPhotographs of the printer’s abbreviation marks, and what each one means. This really is the key to this translation work, to untangle what amounts to 15th century shorthand.
How this was made, and how far to trust it
Project StatusWhat exists, how far it has been checked, what is still open and what has not been done.
Editorial methodHow the reading and the translations were made, which books were compared, and what the AIs got wrong.
Doubtful readings (textual apparatus)Places where the reading is uncertain (a list from the early review, not complete), and the questions of meaning that scholars have not settled.
Gitti and the readingHow the 1878 Italian edition was used, and where it was not, with the changes that cite it.
The 1523 reprintHow this second printing of the same book was used as a cross-check of the abbreviations.