| 摘要 |
Source sentences in the first language to be translated are entered to an input unit 12. Through the input unit 12 source sentences are entered to a polysemous word settlement unit 20 via an entered sentences analyzing unit 14. A characteristics table generation part 30 stores a characteristics table in which each meaning is preliminarily associated with each characteristic with respect to polysemous words. A characteristics record generation part 22 generates a characteristics record to compare the characteristics table and the characteristics record, thereby selecting and outputting the meaning of the polysemous word based on the characteristics in the source sentences. Each word whose ambiguity has been settled at the polysemous word settlement unit 20 is translated at a correlative translation processing unit 40 to be output to an output unit 46. Thus ambiguity of the meaning of a polysemous word in a first language having a plurality of meanings is settled so as to make source sentences in the first language correspond in the most appropriate way to the translated sentences in the second language. <IMAGE> |
