The thesis presents a syntactic and semantic analysis of selected problems in the analysis of transitivity in English, Japanese and Korean in terms of a system of 'thematic relations' adapted from the work of Gruber and Jackendoff. The system is introduced in Chapter One. One way in which it differs from those of Gruber and Jackendoff is that their ambiguous notion of Theme is replaced by two separate notions - Theme and semantic subject. Verbs are classified into 'univalent verbs', which have a single set of thematic relations, and 'ambivalent verbs', which have more than one set of thematic relations. It is claimed that syntactic alternations such as that between 'spray x on y' and 'spray y with x' are a reflection of the ambivalence of the verbs in question. Chapter Two reviews discourse approaches and semantic approaches to Dative-shift. On the basis of a comparison of English and Korean it argues that Dative-shift in these two languages is clearly a semantic process rather than pragmatically motivated. Moreover, like the 'spray' type alternation, Dative-shift is shown to be a reflection of the ambivalence of verbs expressing a change in possession. Chapter Three is devoted to double-nominative constructions in Japanese and Korean, which in the past have been considered to be syntactically and semantically equivalent. A careful study of data from the two languages shows that the double-nominative constructions in Korean are not homogeneous but are divided into two classes with distinct thematic relations, one of which is lacking in Japanese. Chapters Four and Five discuss the passive in Japanese and Korean, respectively, and propose a new classification of passives in the two languages. Three types are recognized: the passive of interest, the anti-causative passive and the attributive passive. Finally, Chapter Six presents some typological observations concerning the passive and Dative-shift. In particular, the English passive and the passives of Korean and Japanese are compared with regard to their relation to Dative-shift, and it is argued that Dative-shift and the passive are, in principle, independent grammatical processes.