A Thesaurus for Biblical Hebrew

What it is about

The authors build a graph thesaurus of Biblical Hebrew roots and lexemes, with three kinds of edges. Semantic edges link headwords in Klein's etymological dictionary, and in Brown-Driver-Briggs, whose English glosses have a tf-idf cosine above 0.33. Phonetic edges come from Clark's phonemic classes, which group roots after Hirsch's theory that similar-sounding roots are related, plus gradational variants (hollow, geminate) and single-letter substitutions. Distributional edges come from PPMI vectors over a lexeme window and from Word2Vec at a cosine of 0.9, both on the ETCBC lexeme text. Human review gives precision of about 0.91 for Klein's semantic edges (6,793 of 7,472), 0.76 for Brown-Driver-Briggs, 0.82 for PPMI and 0.98 for Word2Vec at the high threshold. The phonetic edges are weak: sibilant replacement reaches 0.24, gradational variants 0.11, and membership of a Clark class only 0.03. The graph is served from a Neo4j database.

Ideas

• I1. Test the gradational rules of hollow and geminate roots (yod, vav, he and the repeated middle letter) against a root table as a deterministic generator of candidate variants. The check belongs in the weak-verb chapter, where those rules are stated.
• I2. Use the paper's per-type precision as a negative control: any "similar-sounding roots share meaning" feature has to beat these figures, and most phonetic classes do not. Write such a warning into the Hebrew test set before a phonetic feature is used.
• I3. A review protocol: sample 200 neighbour pairs from a distributional graph, have a human judge them, and report precision by source, the way the paper does. Use it for any lexicon-neighbour suggestion before it is shown to a reader.

For Mova

I1 is a test for The weak verbs; I2 is a measurement in tools/hbo and in The lexicon; I3 is a test protocol for The lexicon. The data is non-commercial: it may be used for tests and measurements only, and nothing taken from it goes into what is built.

Cite

The paper itself, not this card. BibTeX from the library's record of the work.

@inproceedings{azar2020thesaurus,
  title = {{A Thesaurus for Biblical Hebrew}},
  author = {Azar, Miriam and Pahmer, Aliza and Waxman, Joshua},
  booktitle = {Proceedings of LT4HALA 2020 - 1st Workshop on Language Technologies for Historical and Ancient Languages},
  year = {2020},
  url = {https://aclanthology.org/2020.lt4hala-1.10/}
}

References

10 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Brown, F., Driver, S.M., Briggs, C.A. (1906). A Hebrew and English Lexicon of the Old Testament, England. find
2. Clark, M. (1999). Etymological Dictionary of Biblical Hebrew: Based on the Commentaries of Samson Raphael Hirsch. Feldheim, Nanuet, NY. find
3. Hirsch, S. (1867 – 1878). Uebersetzung und Erklärung des Pentateuchs, 5 vols. Frankfurt-on-Main, Germany. find
4. Klein, E. (1987). A Comprehensive Etymological Dictionary of the Hebrew Language for Readers of English. Carta, Jerusalem, Israel. find
5. Mandelkern, S. (1896). Veteris Testamenti Concordantiæ Hebraicae Atque Chaldaicae. find
6. Veit et Comp., New York Ordan, N., Wintner, S. (2007). Hebrew WordNet: a test case of aligning lexical databases across languages. International Journal of Translation 19(1):39-58, 2007. find
7. Roorda, D. (2015). The Hebrew Bible as Data: Laboratory - Sharing – Experiences, https://arxiv.org/pdf/1501.01866.pdf ↗ our card
8. Sefaria, 2018. New Releases: Jastrow and Klein Dictionaries https://blog.sefaria.org/blog/2018/11/12/jastrow-and-klein-dictionaries ↗
9. Strong, J. (1894). The Exhaustive Concordance of the Bible. Hunt and Eaton, New York. 8. Language Resource References Eep Talstra Centre for Bible and Computer. (2015). find
10. ETCBC Dataset Sefaria (2018). The Sefaria MongoDB. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2020-azar-thesaurus-biblical-hebrew/