Sources, method & validation

1 · The source hierarchy

1. Arabic critical edition — authoritative. الداء والدواء, ed. Muḥammad Ajmal al-Iṣlāḥī (takhrīj: زائد بن أحمد النشيري, supervised by Bakr ibn ʿAbd Allāh Abū Zayd); دار عطاءات العلم / دار ابن حزم. PDF as distributed by Alukah (alukah.net) library, Dār ʿAṭāʾāt al-ʿIlm collection.
2. English translation — reference only. The Disease and The Cure, tr. Osman Hamid, Hikmah Publications, First Edition: 1441H / Feb 2020CE, ISBN 978-1-5323-6941-4. © 2020 Hikmah Publications. All rights reserved.
3. Structured metadata — topics, references, relationships and alignment, each derived from the text with a recorded method and confidence.
4. The interface — navigation and short summaries, always labelled as The Cure’s.

The English never silently overrides the Arabic. Where the two could not be matched, the page says so.

2 · Reading the scanned Arabic

The Arabic PDF contains page images only, so every page was read by optical character recognition (Tesseract, Arabic model), with the main text and the editor’s footnotes separated at the printed rule. Printed page numbers were verified against the PDF by matching digit shapes. The Arabic you read here has not been proof-read: vocalisation and some letters can be wrong. Every passage gives its page in the printed edition.

3 · Topics and relationships

Candidate concepts and their Arabic and English vocabulary were listed by hand; a topic is published only if passages supporting it were found in the Arabic text, and its Arabic name is wording that occurs in the book. Samples of each topic’s passages were read to remove false matches (a word used in another sense); topics that failed this check were withdrawn. Minor themes are marked as such. Two topics are shown as related only when at least three passages mention both; each such link was then read and kept only if the passages really discuss the two together. The page for each link lists those passages.

4 · Arabic ↔ English alignment

Passages were matched using references printed in both editions (Qur’an verses, hadith numbers), shared names and passage length. Of 2,533 links: 1088 aligned, 1271 probable, 174 uncertain. A hand-checked sample found aligned links right in 7 of 8 cases and probable in 5 of 8, so English quotations are shown only for aligned passages — and, where one English paragraph spans several Arabic passages, only after a person checked that the quoted words belong to the passage beside them. The translators note that a few segments were omitted in their edition.

5 · Qur’an and hadith

Only references that Ibn al-Qayyim cites are shown. Qur’an references in the Arabic were identified from the surah name and the quoted words, and confirmed where the English edition prints the same reference; unconfirmed ones are marked. Hadith numbers come from the English edition’s footnotes and are mapped to the Arabic by alignment; a number is shown beside an Arabic passage only when that passage is aligned with the footnoted English one, or the Arabic editor’s own takhrīj gives the same number.

6 · Who is speaking

Each passage is labelled automatically as Ibn al-Qayyim’s text or as something he cites (Qur’an, hadith, a Companion, poetry), from cues such as «قال رسول الله». The labels are usually right but not always, and a passage of his text can contain a short quotation. Errors found in review have been corrected by hand. The long speech he writes in the voice of Iblīs (Arabic ed. pp. 229–242) is labelled as a literary device on every passage.

7 · Copyright

• Ibn al-Qayyim’s Arabic text (d. 751 AH) is shown in full. The edition’s critical apparatus (footnotes, takhrīj, indexes) is not reproduced — only counts and facts derived from it. The scanned page images are not published.
• The English translation is © Hikmah Publications. It is used internally for indexing and alignment only. Publicly, The Cure shows headings and at most 3 attributed quotations of up to 40 words per topic — 3,830 words in total. The architecture allows a licensed translation to be added later.
• Canonical Qur’an text: quran-json (QuranEnc Uthmani text), CC BY-SA 4.0.

8 · Open validation issues

Checks from the latest ingestion run that need human attention: