*anadz

From Proto-Indo-European *h₂énh₂ts ~ *h₂n̥h₂tés, an athematic mobile dental stem, alternatively analyzed as a ti-stem *h₂énh₂-ti-s ~ *h₂n̥h₂-téy-s (de Vaan 2008: s.v. anas; Mayrhofer 1992: s.v. ātí-, 163). The precise development in Germanic is disputed. Ringe (2017: 98, 223) reconstructs PIE *h₂ánh₂t- > Proto-Germanic *anud-, interpreting the *u as a possible reflex of the syllabic laryngeal, while Fulk (2018: 89) is skeptical that the etymon even contained a syllable final laryngeal. Kroonen (2013: s.v. anad-, 26), on the other hand, reconstructs a PGmc root noun *anaþ ~ *anadiz, deriving *anad- from a remodeled pre-Germanic *h₂enh₂-et- with two generalized full-grades. He argues that there is no evidence for inherited *anud-. Hansen (2017b: 170–71) reconstructs *h₂enH̥t(-i)- ~ *h₂n̥Ht(-i)- and notes that the Latin evidence points to an original root noun, but that a transition to an i-stem must have already happened in PIE.

The Indo-European cognates are exceptionally widespread and include Sanskrit आति (ātí), Latin anas ~ anatis, Old Prussian antis, Lithuanian ántis, and Proto-Slavic *ǫty (de Vaan 2008: s.v. anas; Derksen 2008: s.v. ǫty; Derksen 2015: s.v. antis I; Mayrhofer 1992: s.v. ātí-, 163–64). Ancient Greek νῆσσα (nêssa), νῆττα (nêtta), and νᾶσσα (nâssa) are probably related but have undergone extensive remodeling: Rix (1991: 195–98) explained these as reflecting a remodeled Pre-Greek *snāti̯a-, subsequently altered to *snākhi̯a- under the influence of νήχω (nḗkhō); in modern laryngeal notation, this analysis corresponds approximately to: PIE *h₂neh₂t-ih₂ → *sneh₂t-ih₂ → Proto-Greek *snātya (cf. De Decker (2016: 157)).