Language Family Classifier

Determining whether two languages are related — and precisely how — is a task that requires careful application of the comparative method, deep knowledge of attested language families, and critical evaluation of evidence quality. This AI assistant specializes in genetic language classification: organizing the world's languages into families, branches, and subgroups based on shared ancestry demonstrated through regular correspondence.

The assistant can explain the internal structure of all major language families — where the branches are, which subgroupings are well-established versus contested, and what the key diagnostic features are for membership in each group. It covers the Indo-European family in detail, including its major branches (Germanic, Romance, Slavic, Indo-Iranian, Celtic, etc.), and extends equally to Afroasiatic, Sino-Tibetan, Austronesian, Niger-Congo, Uralic, Dravidian, Turkic, Mongolic, and many others.

Beyond the well-established families, this tool engages critically with controversial classification proposals — macro-families such as Nostratic, Altaic, Amerind, or Dené–Caucasian — explaining why mainstream historical linguistics regards most long-range proposals skeptically and what methodological standards a valid genetic relationship must meet. This makes it especially useful for researchers and educators who need accurate, nuanced positions on contested questions.

For underdescribed or isolate languages, the assistant can assess available comparative data and discuss current scholarly positions on possible affiliation. It also explains the distinction between genetic relationship, areal diffusion, and typological similarity — a critical conceptual boundary that non-specialists often conflate.

Ideal users include linguistics students studying language typology and classification, researchers working on language documentation, educators preparing materials on world languages, and anyone trying to understand why certain languages are or are not considered related.