About Turkic Search

A v0 prototype for identifying Turkic word stems and finding cognates across all 10 major varieties: Turkish, Azerbaijani, Turkmen, Kazakh, Kyrgyz, Uzbek, Uyghur, Chuvash, Yakut/Sakha, and Old Turkic.

How it's different from Semitic

Turkic languages are agglutinative, not root-and-pattern. The pipeline is: strip suffixes → reach the bare stem → reconstruct the whole-morpheme Proto-Turkic stem → find cognate stems, ranked by how regular the sound correspondences are.

The loan-stratum problem

Turkish and Azerbaijani share thousands of Arabic and Persian loanwords; Kazakh adds a Russian layer. These look like cognates but are shared or parallel loans, not inherited Turkic material. Every match is flagged with its origin so you can tell them apart. See the loan-stratum tour for examples.

Sound-law showcase

The sound-law-showcase tour collects stems that are pedagogically classic across branches — for example, Common Turkic *y- surfaces as j- in Kipchak (Kaz жол jol), s- in Sakha (суол suol), and ś- in Chuvash (ҫул śul). Rhotacism (*z ↔ Chuvash r) and lambdaism (*ĺ ↔ Chuvash l) are the deep-branch signatures of the Bulgaric outlier.

Caveats

This is a v0 prototype. Results are LLM-generated with an OpenRouter primary (Gemini 3 Pro), and can be wrong — especially for less-resourced languages (Sakha, Chuvash, Old Turkic) and rare or dialectal forms. Verify against Wiktionary, Clauson EDPT, Starostin EDAL, or a scholarly Turkological source before citing.

Sibling project to Semitic Search.