Evidence of laryngeal coloring in Proto-Indo-Iranian

Abstract

Past scholarship has made almost no mention of the effects in the Indo- Iranian languages of ‘laryngeal coloring’, the putatively Indo-European development according to which */e/ is ‘colored’ into */a/ or */o/ by an adjacent */h2/ or */h3/, respectively. And for good reason: the merger of nonhigh vowels in Proto-Indo-Iranian would have effaced these distinctions in any case. In this paper I survey the etyma in which laryngeal coloring could have interacted with Proto-Indo-Iranian palatalization, which (in part) preceded the merger of nonhigh vowels, and find that palatalization in almost every case has not occurred to inputs involving */Keh2/ or*/Keh3/, where coloring may be assumed to have taken place. This strongly suggests that laryngeal coloring – not as a discrete ‘sound change’, but as a phonological rule which requires additional sound changes (such as palatalization) before it can ‘show itself’ by affecting the distribution of phonemes in the lexicon – was present in the early stages of Proto-Indo-Iranian, giving further support to the hypothesis that laryngeal coloring was a feature of Proto-Indo- European itself.