README.md

---
license: cc-by-4.0
language:
  - la
task_categories:
  - token-classification
tags:
  - nuberea
  - biblical-studies
  - latin
  - vulgate
  - dependency-parsing
  - stanza
  - morphology
  - gated-spec
size_categories:
  - 1M<n<10M
status: active
configs:
  - config_name: token
    data_files:
      - split: train
        path: data/token/train-*.parquet
  - config_name: sentence
    data_files:
      - split: train
        path: data/sentence/train-*.parquet
  - config_name: dep_edge
    data_files:
      - split: train
        path: data/dep_edge/train-*.parquet
  - config_name: proper_noun_registry
    data_files:
      - split: train
        path: data/proper_noun_registry/train-*.parquet
  - config_name: stanza_corrections
    data_files:
      - split: train
        path: data/stanza_corrections/train-*.parquet
  - config_name: nt_token
    data_files:
      - split: train
        path: data/nt_token/train-*.parquet
  - config_name: nt_sentence
    data_files:
      - split: train
        path: data/nt_sentence/train-*.parquet
  - config_name: nt_dep_edge
    data_files:
      - split: train
        path: data/nt_dep_edge/train-*.parquet
  - config_name: latin_strongs_dict
    data_files:
      - split: train
        path: data/latin_strongs_dict/train-*.parquet
---

# NuBerea Vulgate Latin Dependency Parse Primitives

Morphological and syntactic annotation layer for the Latin Vulgate Bible (Clementine edition), covering both Old and New Testaments. The dataset provides sentence segmentation, word-level morphology, and Universal Dependencies-style dependency parses of the Vulgate text, produced with the Stanza NLP toolkit and refined with hand-curated corrections (proper-noun lemma resolution and lemmatization fixes). A companion lexical mapping relates Latin lemmas to Strong's concordance numbers.

These primitives support computational study of the Vulgate as a translation witness — Latin syntax and lexicon in relation to the Hebrew and Greek source traditions — within the broader NuBerea corpus of biblical-studies datasets. Full schemas and build documentation are available in the gated `SPEC.md`.

## License

CC BY 4.0.

## Attribution

- **Clementine Vulgate** — public domain source text.
- **Stanza** (Stanford NLP Group) — https://stanfordnlp.github.io/stanza/ — Apache License 2.0 (toolkit used to generate the parses).
- **MACULA Hebrew** (Clear Bible) — https://github.com/Clear-Bible/macula-hebrew — CC BY 4.0 (Strong's anchoring for the proper-noun registry).

Parse layer and manual corrections curated by the NuBerea project (2026).