Mergers of Voiced and Voiceless Consonants in the Siamese, Lanna and Lao Languages

Abstract

The voicing of voiceless sonorants (VVS) and the devoicing of voiced obstruents (DVO) in Tai languages not only reduced the number of consonants but also increased tonal distinctions . While DVO in Siamese has been dated to the 1440s–1480s (Tangsiriwattanakul 2020), VVS in Siamese and both VVS and DVO in other Tai languages have received comparatively little attention. This research aims to date the mergers between voiced and voiceless consonants by examining correspondences between graphs and sounds of Siamese, Lanna, and Lao, and analyzing whether graphs attested in inscriptions represented the same or different phonemes. The findings reveal that confusion between graphs representing voiced and voiceless sonorants emerged around the 1370s, while voiced and voiceless obstruents emerged around the 1390s in Siamese. In Lanna, VVS and DVO were evident by the 1410s CE, supported by a 1411 inscription, whereas Lao shows confusions in voiced and voiceless graphs by 1444 CE. Considering that these observed time gaps may stem from limited earlier inscriptions in Lanna and Lao, this study proposes that VVS and DVO likely occurred at approximately the same time across the three languages due to their geographical proximity. These findings align with existing hypotheses (Li 1977; Gedney 1989b; Bickner 1992; Tangsiriwattanakul 2020), which argue that the mergers of voiced and voiceless consonants occurred after the foundation of the Sukhothai and Ayutthaya Tai polities in north-central Thailand.