通常所謂科判，多為疏家施於文本之結構分析，鮮有文本自 設科判者。《大乘起信論》是其異類，文中明確而清晰地表現出 一種自我科判的努力。不過，《大乘起信論》雖然歷來以結構齊 整聞名，學界卻鮮見對其自我科判有全面而系統之討論。本文之 研究，即試圖填補這一方面的空白，專注探討《大乘起信論》之 自我科判。這一探討，主要著眼於《大乘起信論》科判兩個方面 的問題：一則旨在分析其科判之具體表現形式，二則旨在考察其 各種形式科判在組織文本上的有效性。本文試圖說明，《大乘起 信論》之科判中，不僅有明確表述的顯性科判，更有若干暗藏的 隱性科判，兩方交錯，構成一套相當複雜的科判體系；同時，這 一體系並未完全貫通，故《大乘起信論》文中仍有相當之結構關 係未得到有效的說明。

This article examines the self-imposed textual organization (or kepan) of the sixth century Buddhist treatise entitled the Awakening of Faith in Mahāyāna (or Qixinlun in its popular Chinese abbreviation). While the treatise is generally acknowledged as a well-organized text, its practice of kepan (i.e. the practice, rather than the content, of kepan) has not been examined in a comprehensive and systematic manner: how, or in what forms, does the treatise organize itself? In what scope are such kepans applied? And to what extent are such forms of kepan effective in charting out the structural relationship of the treatise? Responding to these questions, this article proposes to address this neglected issue of the kepan practice in the Qixinlun. More specifically, it seeks to identify and outline some of the major forms of the Qixinlun kepan and, where necessary, to discuss a few possible problems in such kepan systems.