Traditionally, the Korean voice has been analyzed as a system of opposition between the active and passive voices, and the causative voice has also been discussed. However, by viewing voice as a correspondence between various situation types and syntactic structures, it is understood as a continuum of different types in syntactic concept spaces. However, the monosemous approach had problems with misclassifying different voice types and presented contradictory explanations by treating other voice types as passive or causative. Therefore, identifying the differences in the event structures and participants’ semantic roles in the constructions revealed through the combination of voice affixes based on the semantic domains of each type appearing in concept spaces is necessary to describe the Korean voice types more precisely. It has been demonstrated that Korean voice affixes -i/hi/li/ki- are polysemous not only for causative and passive voices but also for antipassive, spontaneous, potential, conversive, body action middle, cognition middle, reciprocal, and resultative voices.
