本論文運用法國語言學家邦維尼斯特（Benveniste）言談（discourse）和故事（story）的觀念，分析《四史》傳文，指出故事文本中的言談及其作用。史傳之中，敘述故事（史事）的文字幾乎完全是無我的；至於有我的、言談性質的議論都集中在序、論、贊、或通篇似論的傳中（如《史記．伯夷列傳》）。但我的研究發現：傳文在對話之外，也不乏屬於言談的文字。其中或以「我」介入故事文本，或向讀者提示可參考相關傳文，或說明敘事取捨原因，或直接插話議論，或詮釋人物的話。如果從本文所指涉的時間著眼，不難辨別故事與言談。故事文本所指涉的是過去，過去發生的事；而言談指涉的是現在，寫作時候的議論、說明、解釋。此外，言談介入時，原本敘事中的時間就暫時靜止住，人物的動作也暫停，要等到言談文本（一般不會很長）結束，暫停進行的事件（傳主的生平）才能在文本�媊~續前進。史家撰史，不僅關心發生在過去的事件，更關心他正在進行的寫作行為，處處插話解釋。執筆的作者當然關心自已的文章，但這只是想當然耳，在文章中未必有證據；而我在文本中就找到證據。就寫作技術而言，有的言談可以改易，使得傳文維持一貫的故事狀態。但很多情況都是有必要的，否則讀者很難領會人物的話。由此可見，縱然歷史敘事給人似乎是無人講述的印象，但不介入的敘事方式有時而窮。而不能澈底的只用故事來完滿敘事，有時要倚賴言談的輔助，或許正是歷史敘事的特色。

　　　　　Historical biographies seem to be an a-personal, past-tense narrative except the introduction, the citations, and the comment at the end. I used Benveniste's concepts of discourse and story to analyze the texts of Shi-ji, Han-shu, San-guo zhi, and Hou-han shu, and discovered that there were different discourse narratives among the story narratives. Some discourses told the reader to refer to anohter biography if they want to know more about the thing mentioned. Some discourses explained the citation of a character's words. Some discourses told the reader why the writer recorded or didn't record a certain thing. Some discourses used “I” as a character in the event narrated. Some discourses made comments directly. All of them were written in personal, present-tense narrative. In other words, there was a speaker who wants to persuade the listeners in the text. The discourse appeared in story narratives dividing the text into two different levels. Perhaps we should reconsider wheather a historical narrative is indeed a story without discourse.