This study is to discuss the characteristics of Korean by which English passives are translated into Korean actives. Cho (2005) suggests the three major characteristics of Korean responsible for the voice modulation found in English-Korean translations: i) free word order, ii) a non-subject topic in the sentence-initial position, and iii) ellipsis of subject. This statistical study, however, reveals that the first two features play no role and the third one plays a limited role. It shows that the voice modulation is significantly affected by other typological characteristics of Korean such as BECOME-language (Ikegami 1991) and high-context language (Hall 1976), and also by the use of the combination of ‘noun + delexicalized verb’ caused by lexical gaps between the two languages.