Citation apparatus

How heavily each dictionary cites sources, how broad its source apparatus is, and what it cites most (UC-RD-06).

Deep comparison uses validated feature adapters only. Broad coverage/overlap covers eligible local Sanskrit/BHS headwords; missing deep markup is not counted as zero evidence.

Trust Block

Evidence: src/data/dicts/citation-apparatus.json, tagged <ls> citations, prose iti proxies, and dictionary source links.
Limitations: <ls> density and iti proxies are not directly equivalent, and citation overlap does not prove source dependence by itself.
Validation: generated by npm run build-citation-apparatus; checked by npm test and npm run build.
Owner repo: csl-atlas.
Next use: inspect highlighted rows, then open exact dictionary source records before citing the pattern.

Validated <ls> adapters feed source overlap; density is citations per entry. WIL is essentially untagged for citations, and VCP/SKD/KRM cite in prose or source hints — for those, density uses an iti proxy and is not directly comparable to the <ls> counts.

PWG stands out: a citation-dense, source-rich bilingual dictionary. MW is moderately cited; AP and PWK are leaner; the Petersburg compact (PWK) compresses PWG's apparatus.

Distinct source sigla (locus detail stripped).

Now aligned by a canonical siglum (diacritic/case fold + a reviewed alias table), so MW MBh, PWG MBH and ṚV/RV count together. Counts are citations of each source per dictionary.

Alignment is fold + a small reviewed alias table; sigla that still differ by abbreviation scheme (e.g. MW Pāṇ vs PWG P) are surfaced for review on the source-siglum review queue and merged by adding an alias.

Generated by npm run build-citation-apparatus. See docs/DICTIONARY_COMPARISON_PLAN.md. CC-BY-SA-4.0.