README.md

---
license: cc-by-4.0
language:
  - en
  - grc
  - he
pretty_name: Bible Reference Corpus — Lexicons, Concordance, Topical and Geographic Data
size_categories:
  - 100K<n<1M
task_categories:
  - text-retrieval
  - question-answering
  - token-classification
tags:
  - bible
  - lexicon
  - greek
  - hebrew
  - christianity
  - knowledge-base
configs:
  - config_name: strongs
    data_files: strongs.jsonl
  - config_name: dictionary
    data_files: dictionary.jsonl
  - config_name: meanings
    data_files: meanings.jsonl
  - config_name: renderings
    data_files: renderings.jsonl
  - config_name: sblgnt
    data_files: sblgnt.jsonl
  - config_name: hebrew_verses
    data_files: hebrew_verses.jsonl
  - config_name: word_editions
    data_files: word_editions.jsonl
  - config_name: map
    data_files: map.jsonl
  - config_name: tsk
    data_files: tsk.jsonl
  - config_name: nave
    data_files: nave.jsonl
  - config_name: places
    data_files: places.jsonl
  - config_name: papyri
    data_files: papyri.jsonl
  - config_name: parallels
    data_files: parallels.jsonl
---

# Bible Reference Corpus

Thirteen aligned reference datasets for study of the biblical text: Greek and
Hebrew lexicons keyed to Strong's numbers, an interlinear word map, the critical
apparatus of eight Greek editions, cross-reference and topical indexes, and
geolocated places.

Published by [SermonIndex](https://www.sermonindex.net).

Everything in this repository is **public domain or CC BY 4.0**. Sources with
share-alike terms are kept in a separate repository,
[`sermonindex/bible-reference-sa`](https://huggingface.co/datasets/sermonindex/bible-reference-sa),
so that a share-alike obligation is never inherited by accident.

## Contents

| Config | Rows | Keyed by | What it is |
| --- | --- | --- | --- |
| `strongs` | 14,513 | Strong's no. | Greek and Hebrew lexicon entries from 12 lexicons |
| `dictionary` | 14,117 | headword | Bible dictionary and encyclopedia articles |
| `meanings` | 13,401 | Strong's no. | Plain-English gloss per word (**machine-generated**, see caveats) |
| `renderings` | 5,410 | Strong's no. | How translators rendered each word, with shares |
| `sblgnt` | 7,939 | verse | SBL Greek New Testament text |
| `hebrew_verses` | 23,261 | verse | Hebrew OT, word by word with morphology |
| `word_editions` | 7,948 | verse | Where NA28, NA27, SBL, WH, Tregelles, Tyndale, TR and Byzantine differ |
| `map` | 31,064 | verse | Original-language to English word alignment |
| `tsk` | 29,055 | verse | Treasury of Scripture Knowledge cross-references |
| `nave` | 31,787 | verse | Nave's Topical Bible topics |
| `places` | 5,616 | verse | Geolocated places with coordinates and confidence |
| `papyri` | 3,216 | Strong's no. | Documentary papyri attestations by century |
| `parallels` | 2,526 | verse | Verses with shared vocabulary, scored |

Verse-keyed configs share `reference`, `book`, `book_order`, `testament`,
`chapter`, `verse`. Word-keyed configs share `strongs` and `lemma`. So any of
these can be joined to each other, and to
[`sermonindex/bible`](https://huggingface.co/datasets/sermonindex/bible), on
those fields.

## Lexicons in `strongs` and `dictionary`

| Entries | Licence | Module |
| --- | --- | --- |
| 9,258 | Public domain | International Standard Bible Encyclopedia |
| 8,723 | CC BY 4.0 | Translators Brief Hebrew Lexicon (STEPBible) |
| 8,674 | Public domain | Brown-Driver-Briggs glosses |
| 8,674 | Public domain | Strong's Hebrew Dictionary |
| 8,594 | Public domain | Brown-Driver-Briggs Hebrew Lexicon, full |
| 5,741 | Public domain | Strong's Greek Dictionary |
| 5,523 | CC BY 4.0 | Liddell-Scott-Jones, full edition (STEPBible) |
| 5,488 | Public domain | Abbott-Smith, Manual Greek Lexicon of the NT |
| 5,408 | Public domain | Dodson Greek-English Lexicon |
| 5,298 | Public domain | Nave's Topical Bible |
| 4,771 | Public domain | Thayer's Greek-English Lexicon |
| 4,537 | Public domain | Smith's Bible Dictionary |
| 4,032 | Public domain | Moulton & Milligan, Vocabulary of the Greek Testament |
| 3,921 | Public domain | Easton's Bible Dictionary |
| 2,610 | Public domain | Hitchcock's Bible Names Dictionary |
| 2,277 | Public domain | American Tract Society Bible Dictionary |
| 1,723 | Public domain | Thompson Chain Reference topics |
| 626 | Public domain | Torrey's New Topical Textbook |
| 220 | Public domain | Trench, Synonyms of the New Testament |

Each entry carries its own `module`, `name` and `licence` fields, so
attribution can be reconstructed per row rather than taken on trust from this
table.

**Deliberately excluded.** Two modules that appear on sermonindex.net are *not*
in this release: the Mounce Concise Greek-English Dictionary (© 1993 teknia.com,
non-commercial use only) and the Intermediate Liddell-Scott-Jones (CC BY-SA).
Do not expect this dataset to reproduce the site's lexicon panel exactly.

## Other attributions

- Hebrew text and morphology: TAHOT, [STEPBible](https://github.com/STEPBible/STEPBible-Data), CC BY 4.0
- Greek word data and renderings: Translators Amalgamated Greek NT, STEPBible, CC BY 4.0
- SBL Greek New Testament, CC BY 4.0
- Place geocoding: [OpenBible.info](https://www.openbible.info/geo/) Bible Geocoding, CC BY 4.0; geometry partly OpenStreetMap
- Papyri attestations: Duke Databank of Documentary Papyri, CC BY 3.0
- Treasury of Scripture Knowledge and Nave's Topical Bible: public domain

## Example

```python
from datasets import load_dataset

s = load_dataset("sermonindex/bible-reference", "strongs", split="train")
agape = [r for r in s if r["key"] == "G25"][0]
for e in agape["entries"]:
    print(e["name"], "—", e["licence"])
```

## Caveats

- **`meanings` is machine-generated.** The `simple` field is a plain-English
  paraphrase produced by a language model from the public-domain lexicon
  entries, then reviewed against them. It is a study aid, not a lexicographic
  source, and it should not be treated as ground truth or used to train a model
  that will be evaluated as a lexicon. Everything else in the config
  (`kjv_def`, `tyndale`, `brief`, `lexicon_note`) is from the public-domain
  sources.
- `places` carries a `confidence` and `score` per identification. Low-confidence
  identifications are disputed among geographers; do not read them as settled.
- Strong's numbering is itself an editorial scheme from 1890 and does not map
  one-to-one onto modern lexical analysis. Some keys carry letter suffixes
  (`G1003a`) where later editors split an entry.
- OCR and transcription of the older lexicons is good but not perfect,
  particularly for pointed Hebrew and polytonic Greek diacritics.

## Citation

```bibtex
@misc{sermonindex_bible_reference,
  title  = {Bible Reference Corpus},
  author = {SermonIndex},
  year   = {2026},
  url    = {https://huggingface.co/datasets/sermonindex/bible-reference}
}
```