Middle voice in Cháoyáng Mǐn

Abstract Cháoyáng Mǐn, a dialect originating from Southeast China, exemplifies a unique instance of an isolating language featuring a middle voice. Historically, the middle voice in Cháoyáng Mǐn emerged as the culmination of a sequential transformation involving causative, passive, and anticausative constructions as intermediate stages. This unexpected alternation between valency-increasing and valency-decreasing constructions can be attributed to the grammaticalization of pro-drop (zero-anaphora) within serial verb constructions. The formal marker of all these constructions is the coverb khiɁ ‘give’ which is further merged with the expletive human pronoun i ‘s/he’ into khiɁi as a marker of the last two constructions: anticausative and middle voice. We claim that khiɁi serves as a genuine synthetical middle voice marker, whereas khiɁ fails one crucial set of data to qualify for an analytical middle voice marker.