T'OMIM: A Morphologically Annotated Dataset of Parallel Passages in the Hebrew Bible
What it is about
The full text is not in the library; this card rests on the published abstract and catalogue record, and its ideas are provisional. The dataset fills a gap: scholars have long listed the Bible's inner parallels, but no machine-readable, morphologically annotated set existed. T'OMIM pairs 554 narrative parallels from the Chronicles synoptic tradition with 256 poetic parallels collected from major studies of biblical parallelism. Every pair is aligned to the ETCBC's BHSA text at verse and word level, so each word carries BHSA's morphology. It is released as four Parquet files on Zenodo, and it is the training and test material of the MiqraBERT paper.
Ideas
- I1. Use T'OMIM's pairs as the gold for Mova's own parallel finder (lemma overlap plus tree alignment): CC BY 4.0 data may be used and cited. Map BHSA word ids to MACULA ids (both rest on the Leningrad Codex) once, and keep the mapping as a small file.
- I2. The word-level alignments of the 256 poetic pairs are a ready test of "which word answers which" in parallel cola; use them to score the role-alignment idea in the MiqraBERT card (I2) without hand-labelling.
- I3. From the narrative pairs, count the kinds of change between the earlier and later text (a word replaced, a word added, a construction changed) as a small, dated table of Early vs Late Biblical Hebrew substitutions.
For Mova
I1 and I2 need a BHSA→MACULA word mapping (a step in tools/macula), then the parallel tool proposed in hbo-2025-smiley-intertextual-parallel-detection-biblical-hebrew. I3 is a table for 12. Biblical, Mishnaic and Modern Hebrew compared (the dating debate), with counts dated. Most pairs lie outside the Torah, so our trees cover only the Torah side of the few Torah poetic pairs; the rest needs MACULA's other books.
Cite
@article{smiley2026t,
title = {{T'OMIM: A Morphologically Annotated Dataset of Parallel Passages in the Hebrew Bible}},
author = {Smiley, David M.},
journal = {Journal of Open Humanities Data},
year = {2026},
doi = {10.5334/johd.608},
url = {https://doi.org/10.5334/johd.608}
}This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/hbo-2026-smiley-tomim-morphologically-annotated-dataset-parallel/