Select a root, or search for a word.
Every star is a root family; its size is the number of words that grew from it. Click a root to open its tree.
Pick any two words — Persian or English — and walk the road between them.
«فارسی شکر است»
Rishe (ریشه, pronounced ree-sheh) is the Persian word for root. I made this because I love Farsi — the language I think and feel in — and I wanted to see how deep its roots go. Every Persian word carries centuries of travel in it: down from the steppe through Avestan and Old Persian, in and out of Arabic, Greek, and Turkish, sometimes all the way into English and back again. A dictionary tells you this one entry at a time; I wanted a place where you could wander it — follow برق to emerald, watch پردیس become paradise, and see the whole ancestry of a language laid out like a night sky.
The backbone of everything here is Ali Nourai’s An Etymological Dictionary of Persian, English and other Indo-European Languages — a remarkable labor of love charting over 1,600 shared roots of some 4,700 Persian and 3,300 English words. This site is an interactive rendition of his hand-drawn charts: the roots, the words, the arrows, the verse citations and the cross-references are all his, and every chart links back to a scan of the printed page it came from. The book is freely available at the Internet Archive.
Nourai built his charts on a bibliography of more than a hundred works, and this site now stands on the same shelf. To verify his derivations I assembled that library — the Indo-European dictionaries of Pokorny, Walde-Pokorny, Mann and Buck and Watkins’ root appendix to the American Heritage Dictionary; Bartholomae’s Avestan dictionary, Kent’s Old Persian lexicon, MacKenzie and Nyberg for Pahlavi, Gharib for Sogdian, Bailey for Khotanese; Horn and Hübschmann on New Persian etymology; Klein, Skeat, Ernout-Meillet and Walde-Hofmann on the European side; Jeffery, Addi Shir, Asbaghi, Fraenkel and Zimmern on the Semitic exchanges; Borhān-e Qāṭeʿ with Moʿin’s etymological footnotes, Moʿin’s own Farhang-e Fārsi, Aryanpur, Abolghasemi, Noushin’s Shāhnāmeh glossary, Asadi’s Loghat-e Fors, and many more — plus sources he could not use, above all Cheung’s Etymological Dictionary of the Iranian Verb, the Encyclopaedia Iranica, Wiktionary with its citations, and Etymonline. Hover any citation code on a chart (KLN:164, BQT:2053…) to see the full record of the book it points to and where it can be read.
The book exists only as a scanned image — no digital text — and ordinary OCR destroys its hand-drawn charts. So the extraction was done by a fleet of Claude (Anthropic) vision-model agents working page by page: each of the 541 chart pages was rendered at 200 DPI, every box was read with the Perso-Arabic script transcribed character by character from magnified crops (diacritics included), the drawn arrows were traced to reconstruct each derivation tree, and the results were checked structurally — every parent link, page number, and script field validated — with disputed glyphs re-read against classical verse where the charts quote Ferdowsî, Hâfez, or Sa‘dî.
Extraction only copies the book; verification asks whether the book is right. Chart by chart, every derivation arrow is re-examined against modern scholarship and against the very pages Nourai cites: each box gets a verdict — confirmed, plausible, disputed, unverified, or misprint in book — and an expandable “How this derives” panel with the phonetic and historical explanation, the archived web sources, what each of his references actually says at the cited page, and every other book from the library that was consulted. Where the sources disagree with him, the panel says so and says why; where a strange form turns out to be printed that way in the book itself, it says that too. Nourai remains the author of every claim on this site — the library is here to test his work, not to replace it — and the tally under each root card tells you how far the checking has reached.
Even so, small transcription slips are possible — when something matters, press “View scanned page” under any chart and check the original. Cross-references marked ☞ are the book’s own margin notes (usually one element of a compound), not claims of shared ancestry.
Built by Sajad F. Maghrebi. I did the extraction, the verification runs, the site and the research section; the etymologies are Ali Nourai’s, and the judgments on them rest on the reference library listed above — Pokorny, Bartholomae, Horn, Klein, Moʿin, Cheung and the rest. This is a labor of love, not a work of professional philology, so if you know these sources better than I do, corrections are welcome. The data and code live on GitHub.
Choose a word and watch its whole family fan out across lands and centuries — your word’s own road in bright gold.