History is a narrative in which the author selects references according to his intention. However, it is hard to realize what kind of reference has been used or excluded by an author while in the process of the history compilation. Sima Guang(司馬光), for example, referred to more than three hundred historical documents while he was compiling Zizhitongjian(資治通鑑), and cross referred to at least four to five sources for any one event in history. Zizhitongjian-Kaoyi(資治通鑑考異) explains why one reference was chosen over another when evaluating historical sources. It holds value in the fact that it shows the process of selecting the appropriate sources for compilation, unlike most other counterparts whereby authors choose certain sources without explanation. This thesis focuses on “Tangji(唐紀)” of Zizhitongjian-Kaoyi to understand the boundary of history in Sima Guang(司馬光)’s point of view. He did not include ‘official history(正史)’ just because they were deemed official; neither did he exclude narratives just because they were deemed unofficial such as ‘Xiaoshuo (小說, Miscellaneous notes)’. He had his own standard of evaluating documents, and included sources which deemed fit in his standard. His sources do show his own perspectives and assumptions, but he seeks to speak of history that teaches us lessons. This reflects how much literary narratives are included in the process of compiling and editing history.