Post-Nasal Deaspiration in Ancient Greek: Mirage or Reality?

While Proto-Indo-European voiced aspirated stops are normally reflected by voiceless aspirates in Ancient Greek, it has long been noticed that they sporadically surface as voiced stops after a nasal. A regular sound change was first posited by Miller (1977a, 1977b), who identified the conditioning factor in the position of the accent: aspiration was lost at an early stage of Proto-Greek (before the devoicing of PIE voiced aspirates) in post-accentual syllables only. Such a development would be typologically well-supported: loss of aspiration is cross-linguistically common after nasals, and aspiration is generally weaker after the accent than in accented syllables. Nonetheless, the rule remains overlooked in standard reference works, and its scope has not yet been clearly defined. A fresh look at post-nasal deaspiration is all the more desirable since this rule interacts with other important (Proto-)Greek sound changes (Rix’s Law, Grassmann’s Law) and has implications for the etymological analysis of the Ancient Greek lexicon, providing a possible internal explanation for consonantal alternations often attributed to Pre-Greek substrate origin. This paper aims to establish the conditioning of post-nasal deaspiration, based on a thorough revision of all the potential evidence.