Detecting Narrative Patterns in Biblical Hebrew and Greek

What it is about

The authors look for recurring narrative patterns ("type-scenes", such as a man meeting a woman by a well, or a hero being tempted) as an information-retrieval task. A Siamese BERT-style encoder is fine-tuned with a triplet loss so that verses linked by an openbible.info cross-reference (vote threshold 5; about 17,900 Hebrew and 45,300 Greek pairs) get similar vectors. Each query is a verse with two verses of context on each side, and the encoder ranks all candidate verses by dot product. Hand-written queries, rephrased from scholarship, reach relevant verses in the top 10 in many cases, but recall is low in absolute terms. Keeping the query verse unmasked nearly doubles the score over masking, and a batch size of 16 and a window of 4 verses work best. Failures are concentrated in vague queries ("a hero is tempted", "the words of a dying hero"), and Job dominates the results for a prophet query.

Ideas

• I1. Use the cross-reference pairs as a gold set for event-level similarity: two events are "linked" when a cross-reference joins them. Measure whether world links the same events, reporting precision at 10 against the 5-vote pairs.
• I2. A test set of type-scene queries with hand-checked answers (the paper's Table 4 gives ten queries and returned verses with labels). Use it as a tools/hbo retrieval test, reporting top-10 hits, and treat the failures as the list of cases a rule must handle.
• I3. The finding that concrete nouns (man, woman, well, son) retrieve better than abstract ones suggests a rule-based similarity that weights concrete lexemes higher than function words or abstract predicates; test it against I2 before any neural model.
• I4. The paper finds that "hero" has no single Hebrew equivalent (gibbor or nabi). Record this as a lexicon entry in The lexicon: one English gloss with two Hebrew candidates, and the rule that a query gloss may map to several lemmas.

For Mova

I1 belongs to world (the event chain built from the trees) and is measured with tools/hbo. I2 is a test set for tools/hbo retrieval. I3 is a rule for the similarity step, to be tried before any neural retrieval. I4 goes into The lexicon.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/2024.ml4al-1.26

@inproceedings{mcgovern2024detecting,
  title = {{Detecting Narrative Patterns in Biblical Hebrew and Greek}},
  author = {McGovern, Hope and Sirin, Hale and Lippincott, Tom and Caines, Andrew},
  booktitle = {Proceedings of the 1st Workshop on Machine Learning for Ancient Languages (ML4AL 2024)},
  year = {2024},
  doi = {10.18653/v1/2024.ml4al-1.26},
  url = {https://doi.org/10.18653/v1/2024.ml4al-1.26}
}

References

20 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Andrew Piper, Richard Jean So, and David Bamman. 2021. Narrative theory for computational narrative understanding. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 298–311, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. find
2. Robert Alter. 1978. Biblical type-scenes and the uses of convention. Critical Inquiry, 5(2):355368. find
3. Robert Alter. 1983. How convention helps us read: The case of the bible’s annunciation type-scene. Prooftexts, 3(2):115–130. find
4. Walter Arend. 1933. Die typischen Scenen bei Homer. Weidmann. find
5. Erich Auerbach. 1959. Scenes from the Drama of European Literature: six essays. Meridian Books. find
6. Vassileios Balntas, Edgar Riba, Daniel Ponsa, and Krystian Mikolajczyk. 2016. Learning local feature descriptors with triplets and shallow convolutional neural networks. In British Machine Vision Conference. find
7. Brian Britt. 2002. Prophetic concealment in a biblical type scene. The Catholic Biblical Quarterly, 64(1):37–58. find
8. Eryl W. Davies. 1995. The Journal of Theological Studies, 46(2):572–575. find
9. Mark Finlayson. 2009. Deriving narrative morphologies via analogical story merging. find
10. Niv Fono, Harel Moshayof, Eldar Karol, Itai Assraf, and Mark Last. 2024. Embible: Reconstruction of Ancient Hebrew and Aramaic texts using transformers. In Findings of the Association for Computational Linguistics: EACL 2024, pages 846–852, St. Julian’s, Malta. Association for Computational Linguistics. find our card
11. Cheng Jiayang, Lin Qiu, Tsz Chan, Tianqing Fang, Weiqi Wang, Chunkit Chan, Dongyu Ru, Qipeng Guo, Hongming Zhang, Yangqiu Song, Yue Zhang, and Zheng Zhang. 2023. StoryAnalogy: Deriving story-level analogies from large language models to unlock analogical understanding. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 11518–11537, Singapore. Association for Computational Linguistics. find
12. Malte Ostendorff, Terry Ruas, Till Blume, Bela Gipp, and Georg Rehm. 2020. Aspect-based document similarity for research papers. find
13. Vladimir Propp. 1968. Morphology of the folktale. Austin: University of Texas Press. find
14. Shaltiel Shmidman, Avi Shmidman, and Moshe Koppel. 2023. Dictabert: A state-of-the-art bert suite for modern hebrew. find our card
15. Pranaydeep Singh, Gorik Rutten, and Els Lefever. 2021. A pilot study for bert language modelling and morphological analysis for ancient and medieval greek. In The 5th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2021). find
16. Katherine Thai, Yapei Chang, Kalpesh Krishna, and Mohit Iyyer. 2022. Relic: Retrieving evidence for literary claims. find
17. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lhoest, and Alexander Rush. 2020. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 38–45, Online. Association for Computational Linguistics. find
18. Ivan P. Yamshchikov, Alexey Tikhonov, Yorgos Pantis, Charlotte Schubert, and Jürgen Jost. 2022. Bert in plutarchs shadows. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. find
19. Diederik P. Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In Proceedings of the 3rd International Conference for Learning Representations (ICLR). find
20. John Koutsikakis, Ilias Chalkidis, Prodromos Malakasiotis, and Ion Androutsopoulos. 2020. Greek-bert: The greeks visiting sesame street. In 11th Hellenic Conference on Artificial Intelligence, SETN 2020, page 110117, New York, NY, USA. Association for Computing Machinery. English Query Translation (hbo) find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2024-mcgovern-detecting-narrative-patterns-biblical-hebrew/