How World texts work

Where these texts come from, why these works are here, and what the number beside a parallel means.

What a parallel is

A parallel is a passage of another work that a computer found near in meaning to a paragraph of the Papers.

A language model, OpenAI's text-embedding-3-large, reads each paragraph of the Papers and each passage of each work, and gives each one a place in a space of meaning. Texts that say similar things get places near each other. For each paragraph, the Hub shows the passages whose places are nearest.

Each work is cut into passages of about the length of one paragraph of the Papers. A passage never runs past the end of a chapter, a hymn, or another titled part. The Tao Te Ching is compared by whole chapters.

The model compares meaning only, and it knows nothing of history. A parallel says that two passages say similar things. It does not say that one came from the other, or that they agree.

What the number means

The number is the score of that comparison, as a percent: how near the two places are. It is not a share of words in common.

• Scores are lower than most people expect. Two paragraphs of the Papers that are nearest to each other usually score from 58 to 73.
• A passage of another work below 40 does not show. In a read of the results for 35 paragraphs, each passage from 46 up was near in meaning, most from 40 to 46 were, and most below 40 shared a word and no meaning.
• The Book of Mormon and the Rig Veda start at 45. From 40 to 45, their passages shared a word, a subject, or the style of the Bible, not a meaning.
• When no passage of another work reaches its floor, the list offers the five nearest below it, marked as likely not related.

Where the texts come from

Each text is an English translation in the public domain in the United States, from a copy that people proofread: Project Gutenberg, Wikisource, or eBible.org. The Bible is the World English Bible.

The Hub shows the words as the source has them. The one change is a typo of the online copy that the printed book does not have. The page of each work names its translator, its year, and its source under “This edition”.

Why these works

The first works are texts of the religions that Paper 131, “The World's Religions”, describes: the Cynics, Buddhism, Hinduism, Shinto, Taoism, and Confucianism, with the Bible for Judaism. Where a work belongs to a section of Paper 131, its page links to that section.

The Koran and the Japji of the Sikhs came next. For studies that set passages of the Papers beside earlier books, see the work of Matthew Block.

The rest widen the view to other traditions: Egypt, Greece, the Norse, the Aztecs, Japan, China, and India, and the Book of Mormon. Scriptures are texts that their traditions read as sacred. Classics are stories, histories, and teachings of a tradition that are not read as scripture.

Paper 131 also describes Zoroastrianism and Jainism. For their texts, the Gathas and the Uttaradhyayana, we found no proofread copy online of an English translation in the public domain.