Data

The dataset is rebuilt from the platform on every sync. Files are stable in shape; the zip carries the harvest date in its name. Transcriptions synced from みんなで翻刻 on 5 Sept 2026.

TEI XML

TEI P5 diplomatic transcription, one file per witness volume, with page breaks tied to the IIIF canvases and contributor credit in the header.

tei/moshiogusa/ninjal-1.xml 蝦夷方言藻汐草　乾巻（国立国語研究所）
tei/moshiogusa/ninjal-2.xml 蝦夷方言藻汐草　坤巻（国立国語研究所）
tei/ezo-hogen/leiden.xml 蝦夷方言（Wereldmuseum Leiden）
tei/hokkai-zuihitsu/tsukuba.xml 北海随筆（筑波大学附属図書館）
tei/ezo-yabubanashi/hokudai.xml 蝦夷薮話（北海道大学附属図書館）
tei/ezo-kiko/ryukoku.xml 蝦夷紀行（龍谷大学図書館）
tei/ezoto-kikan/hokudai.xml 蝦夷島奇観（北海道大学附属図書館）
tei/ezo-soshi/nijl.xml 蝦夷草紙（国文学研究資料館）
tei/ezo-soshi/ryukoku.xml 蝦夷草紙（龍谷大学図書館）

File

File	Contents
erdal-2026-09-05.zip	Whole dataset (zip)
entries.jsonl	Wordlist items: gloss, historical form, alternatives, remark, editorial modern form with certainty and citation, and the page permalink.
entries.csv	Wordlist items: gloss, historical form, alternatives, remark, editorial modern form with certainty and citation, and the page permalink.
pages.jsonl	One record per transcribed page with the platform's raw text and the page's status and editor.
contributors.csv	Transcription credit per contributor and page.
sources.json	Source and witness metadata.
sources.yaml	Source and witness metadata.

API

The same JSON the site reads is served under /x/ with open CORS: index.json, entries.json, contributors.json and units/<source>--<unit>.json.

Licence

Transcriptions and derived data are CC BY-SA 4.0, attributed to the contributors of the platform project. Facsimile images belong to the holding institutions and are shown from their IIIF servers under their own terms, linked from each witness.