WIKETYM: An Automated Arabic Etymological Dictionary — Dataset

WIKETYM is an automated Arabic etymological dictionary derived from English Wiktionary through a three-layer evidence architecture. The dataset covers 21,815 Arabic lemmas, of which 6,763 (31.0%) carry an etymological assignment and 4,244 (19.5%) are supported by direct documentary evidence — structured Wiktionary etymology templates or explicit derivational linkage — independent of root-index inference. Evidence is resolved through three ordered layers: template-based documentary evidence from structured etymology sections, explicit proximate derivation linking a lemma to a documented parent form, and root-index inference connecting lemmas via shared Semitic roots. Each entry carries a four-grade reliability score (A–D) and an epistemic-status field distinguishing documented assignments from inferential defaults. Template and category evidence co-occur at a rate of 0.880, without requiring identical source-language granularity. The dataset further records 21,834 scholarly references, 29,523 descendant-language connections, 1,344 recovered textual attestations, and 2,946 full etymology chains. Outputs are distributed as structured JSONL files with bilingual Arabic–English field labels, generated by an open-source Python pipeline adaptable to any language with structured Wiktionary etymology coverage.