Khmer onset voicing at the end of the 13th century

Khmer displays extensive diachronic phonological restructuring (Huffman, 1976), especially in the realization of initial stops. These changes include (i) devoicing and merger of voiced and voiceless stops, and (ii) the emergence of implosives from pre-vocalic voiceless stops. However, the details and the chronology of these changes remain unclear in many respects. This paper proposes a chronology of the two changes based on philological evidence: Chinese transcriptions of Khmer words in the Zhēnlà Fēngtǔ Jì (ZFTJ), a travel account from the late 13th century. Previous research on ZFTJ by Pelliot (1951) suggested that implosives had already emerged at the time of transcription, while the merger of voiced and voiceless stops had at least started. This is at odds with the general view of Khmer language history. I motivate a revised analysis of the Khmer transcriptions in the ZFTJ and show that the devoicing and merger of voiced and voiceless stops in fact had not yet occurred at the end of the 13th century. There is, however, not enough evidence from ZFTJ to confirm whether implosives had already emerged.