以數位分析進行《金剛經》註疏科判脈絡探索— —兼論俞樾《金剛經注》之科判結構特色

摘要本論文研究對象《金剛經》為中國佛教史上最廣為流傳的經典之一，縱觀學界對《金剛經》的討論，大抵皆著重在版本、經義、思想、宗派立場上，並且多為單一註疏或少數幾部註疏間的思想，但對於註疏間的傳承脈絡討論較少，這個現象可能來自於傳統研究方法對於大量資料比對的上的困難。科判為闡明一經要旨之注釋方式，藉由科判內容，便可大致理解該註疏家對於該部經論的整體與細部觀點。本研究嘗試透過數位分析大量進行科判比對，以探索註疏間的思想傳承與演變脈絡。筆者從《金剛經》註疏科判下手，意欲找出四十二部《金剛經》註疏間的關聯性。但因資料量較為龐大，因此採用資訊技術，透過演算法的運算，計算出《金剛經》註疏兩兩間的科判相近程度，再透過階層式分群演算法(Hierarchical Clustering)，對此四十二部《金剛經》註疏進行分群，梳理出四十二部《金剛經》註疏科判型態的全貌。研究結果顯示，首先，科判為一隨著歷代發展而逐步趨向分科精簡的演變過程;其次從各群的群聚現象分析結果顯示，若兩註疏間的作者有師承關係，或同時引用參考到相同註疏時，均會明確反應在其註疏的注解或斷點上，也就是後人確實從前人處得到傳承，並且多會在前人的基礎上對斷句做調整。另外，透過本研究之數位比對方式，當某部註疏的注釋方式明顯異於大部分傳統的注解方法時，將會很清楚的被辨識出來。雖然利用數位分析，我們可以相當容易的進行大量資料的比對，但對於深入解釋數位分析的結果，往往必須藉由細節比對文獻內容才能完成。因此，本研究特別挑選其中一部科判結構具有極大差異的作品，即清末儒者俞樾所著的 X0506《金剛經注》，進行傳統文獻研究的細節研讀方式進行俞樾的 X0506《金剛經注》的科判探討，找出其異於其他《金剛經》註疏作品的原因，並以此示範一個完整的由數位分析方式找出值得被關注的問題，做為整合數位技術與人文思考之成功研究範例。

關鍵字

金剛經；科判；階層式分群演算法；結構；俞樾

並列摘要

The study target of this paper, the Diamond Sutra, is one of the most widely circulated scriptures in the history of Chinese Buddhism. But most academic discussions of the Diamond Sutra only have focused on editions, meanings, ideologies, and religious stances. And it is mainly based on a single commentary or ideas between a few commentaries, and less attention is paid to the development and inheritance among commentaries. This phenomenon may be because traditional research methods are not easy to use for comparing large amounts of data. Kenan is an annotation method that can express text structure and content topic. From the content of Kepan, we can get a general idea of the commentator's overall and detailed view of the scripture. This study attempts to compare Kepan through a digital approach to explore commentaries' development and ideological heritage. We started with Kepan of the commentaries on the Diamond Sutra, trying to find the connection between the forty-two commentaries on the Diamond Sutra. Our innovative algorithm calculated the similarity of the two Diamond Sutra commentaries. Then with a hierarchical clustering algorithm, the forty-two commentaries on the Diamond Sutra were grouped, and we sorted out the overall picture of the forty-two "Diamond Sutra" commentaries. The study results show that, firstly, The development of Kepan has undergone an evolution towards simplification over time. Second, the clustering results show that if the authors of two commentaries have a mentoring relationship or cite the same commentary simultaneously, it will be reflected in the content or breakpoints of the commentaries. That is, later generations will often adjust the sentence segmentation based on their predecessors. In addition, through the digital comparison method proposed in this study, it will be identified when a commentary is markedly different from most of the other commentaries. In addition, through the digital comparison method of this study, when the annotation method of a particular commentary is significantly different from most traditional annotation methods, it will be identified. While digital analysis methods allow us to compare a large number of pans easily, the in-depth interpretation of the results of digital analysis often requires a detailed comparison of the content of the commentaries. Therefore, this research selects a work whose Kepan structure differs from other commentaries, X0506 " Diamond Sutra Commentary," written by Yu Yue, a Confucian scholar in the late Qing Dynasty. We adopt the traditional literature research method to find out why it differs from other commentaries. From the above discussion, we have fully demonstrated a complete digital assistant problem-solving process which firstly discovered worthy of the attention issues by digital method and solved the problem with traditional humanities research methods. We believe this successful research example of integrating digital technology and humanistic thinking.