Library: Hebrew Bible Now Reads in Nine Languages

What shipped

As of June 7, 2026, the Wheel of Heaven library hosts the full Hebrew Bible — Genesis through Zephaniah, thirty-six books — in nine languages. Every chapter renders the English ASV 1901 baseline alongside imported public-domain editions in German, Spanish, French, Japanese, Korean, Russian, Simplified Chinese, and Traditional Chinese. Readers consulting the library in any of the nine site languages now see the text in their reading language without having to route through English.

The imported editions are named, dated, and openly licensed. The catalog also reports per-language coverage: where versification shifts between the Masoretic and Vulgate or Synodal traditions leave a verse unaligned, the English ASV remains as a fallback.

Per-book provenance — translation name, year, license, source URL, retrieval date — is recorded in data-library/catalog.json for every entry, under each book's translations.{lang} block.

Why literal translations

Every imported edition above is, by its translation theory, a formal-equivalence translation — sometimes loosely called "literal." The chosen editions sit on the concordant side of their respective traditions: Elberfelder is famously the most word-for-word German Bible of the modern era; Reina-Valera, Crampon, the Synodal, the KRV, and the Chinese Union Version all belong to the same formal-equivalence lineage in their own languages. 口語訳 is the colloquial-Japanese modernization of the classical 文語訳 — still formal-equivalent by the standards of the tradition, deliberately positioned against the dynamic-equivalence 新共同訳 (1987) that later became the Japanese Catholic-Protestant standard.

The choice is deliberate, not nostalgic. A formal-equivalence translation tries to preserve the lexical and syntactic shape of the source — one Hebrew word, one target word, where the languages will permit it; one Hebrew clause, one target clause; the same word in the same source verse gets the same translation across the corpus where it can. A dynamic-equivalence translation does the opposite — it tries to preserve the effect on the reader, which means it routinely collapses, paraphrases, and smooths the source's choices into idiomatic target language.

For a library whose entire reason to host eight translations is comparative — what does each tradition do with the divine-name pericopes, with Genesis 1:1, with Psalm 82, with the Elohim plural? — formal equivalence is the only register that preserves the question. Once a translation has decided to render יהוה as "the LORD" because that reads better in English, or as 主 because that reads better in Japanese, the lexical choice the canon wants to track is already gone. The reader sees the translator's tidying, not the translator's reading.

The same reasoning applies to the Elohim referent itself. Elberfelder keeps "Jehova" / "Elohim" visible where the Hebrew distinguishes them; Crampon writes "Yahweh" and "Élohim" with the same discipline; the Synodal preserves "Господь" / "Бог" pairings that map back to the underlying source. A dynamic-equivalence translation would, in the same verse, give the reader one undifferentiated God — easier to read, but the comparative argument no longer has anything to compare.

Why this is the right shape for the library

The library has always treated the Hebrew Bible as the operative source-text the project reads, not as a destination translation. The English ASV is the reading surface for the project's primary audience; the underlying Hebrew is what the canon actually engages with. What was missing, until today, was a clean way for non-English readers to consult the library without making English a mandatory intermediate stop.

The choice of edition per language follows the canon's source-tier discipline. Each imported translation is a critical or near-critical edition from the nineteenth or twentieth century — old enough to be public domain, recent enough to be philologically serious. Where the canon's preferred reading would require a modern edition still under copyright, the project either uses the public-domain subset of that edition or notes the gap explicitly. The clearest case is French: the Dhorme Pléiade Bible — the edition Raël himself worked from during the 1973 contact — sits under French copyright until roughly 2071. The library uses Crampon 1923 as the baseline, with Dhorme's earlier public-domain Job (1929) and Samuel (1910) replacing Crampon for those three books where they exist. The full reasoning for every language — why Elberfelder rather than Luther, why Crampon rather than Darby, why 口語訳 rather than 新共同訳, why the 1961 KRV rather than the strict 1938 — lives on the docs site at docs.wheelofheaven.world/contributing/content/imported-translations.

The corpus reads the Hebrew Bible as a layered text whose Elohim referent is operative in the original language. Modern translations are downstream interpretive moves, and which interpretive move a translation makes is visible at specific verses — Genesis 1:1, Psalm 82, Exodus 3:14, the divine-name pericopes throughout. Hosting eight imported translations side-by-side with the ASV makes that visibility ordinary: the reader who wants to compare what Elberfelder, Crampon, the Synodal, and the CUV do with a contested verse can now do so without leaving the library page.

Where this sits in the longer plan

The imported-translations pipeline is one half of the library's mandate. The other half is the Wheel of Heaven Translation, currently advancing through Genesis chapter by chapter — a fresh English reading produced from the Hebrew under the project's own glossary discipline, with a Translator-Editor-Reviewer pipeline behind every chapter. The imported translations cover the breadth (every book, eight languages, public-domain editions, shipped today); the WoH Translation provides the depth (one tradition, one book at a time, philologically defended, still in progress).

Both pipelines feed the same paragraphs[].i18n[lang] field on each chapter file, which means the library template renders them through the same interface. When the WoH Translation of Genesis reaches its first stable release, it will sit alongside these imported readings as the project's own contribution to the comparative landscape it has been building.

Where to see it

• Library index: /library/ — every Hebrew Bible book now lists nine languages in its availability matrix
• A sample chapter: /library/genesis/ (or in your reading language: /de/library/genesis/, /ja/library/genesis/, …)
• Catalog and provenance per book: data-library/catalog.json on GitHub
• Methodology and per-language source-program reasoning: docs.wheelofheaven.world/contributing/content/imported-translations

Known follow-ups

Three. The French Dhorme placeholders for 1-Samuel, 2-Samuel, and Job are still pending a clean OCR pass — the underlying djvu scans on the Internet Archive interleave Hebrew apparatus and commentary too aggressively to extract programmatically. The Obadiah chapter file has a scaffolding gap (only one of twenty-one verses is split out at the paragraph level) that the catalog currently misreports as complete coverage. And the 1-Chronicles German Elberfelder source-of-record (bibel-online.net) has a known duplicate-content quirk in chapters 13, 17, 18, 22, and 28 that the catalog notes in _meta.notes but the merge pipeline has not yet filtered downstream. None of these block readability of the imported text; all three are tracked for the next pass.

— Filed June 7, 2026, Wheel of Heaven editorial desk.