Abstract:Generally speaking, studies of spoken language （including bilingual communities) usually fail to find the real “source” of grammaticalization. However, the grammaticalization of written language induced by “language contact through translation” has the possibility of discovering the immediate “starting point” of a grammaticalization item. Based on this idea, this paper examines the origin of the format “ （NP1) dui对（ duiyu对于)NP2 zhi之N”, identifies its real “starting point” （the first groups of use cases), and describes its code-copying phenomenon as detailed as possible. The author believes that this format is a code-copying of the Japanese format “（NP1) NP2ニ对スル N”, and its earliest examples are found in such texts as The Annals of Japan（ Riben Guozhi《日本国志》) （1895), O n The Right of The Strongest（ Lun Qiangquan《论强权》)（1899) and The Compilation of Translations Works（ Yishu Huibian《译书汇编》)（1901), where the process of grammatical transfer was accomplished into Classical Chinese texts. Due to the differences in the code-copying strategies adapted by different translators, two sub-categories with similarities and minor differences have been/are formed in the target language. It is believed that it can provide a real and reliable case for the improvement of relevant theories.