This thesis investigates the consequences of historical changes of Japanese case particles, 'accusative' o and 'dative' ni. The change of o is characterized as becoming a structural Case-marker from the inherent Case-marker. The consequences of this change are manifested in; (a) the ni filling the gap of linking to the FROM-function vacated by the particle o, becoming the structural accusative Case-marker; (b) the development of the o causative construction; (c) the inability of topicalizing the o-marked object; (d) the disappearance of the sequence of NP-o-to in the coordinate structure; (e) the development of the double o constraint. The constancy of ni throughout the history of the Japanese language is characterized by its lexical content; the Locative ni has not changed. The development of the ni causative is, then, attributed to the development of the nominative marking triggered by the accusative marking, that is, from the ergative case to the nominative case. This ergative hypothesis of Old Japanese is supported by the distribution of the o-marked and non-overtly marked objects which is determined by the transitivity features proposed by Hopper and Thompson (1980).