數位人文「文本空間化」的實證研究：以詩歌、傳記、日記為例

早期人文學者對於文本內容的分析與研究往往是建構在知識的博覽，在熟讀相關文獻資料後，將知識內化於研究應用。在進入資訊化社會後，爆量資訊讓研究者必須思考如何調整研究方法，以順應科技帶來的改變，以及改變後所衍生出的相關議題。所幸近年來受益於數位人文多元思維的啟發與影響，帶給傳統人文領域的研究學者諸多創新思路，特別是在研究方法的應用上，不僅突破了既有的理論框架，效率化的研究工具與系統化作業程序，擴展並促成了愈來愈多跨領域學科的合作。本研究嘗試從「文本空間化」的思維角度出發，針對《全臺詩》、《蔣經國先生大事長編初稿》與《雷震日記》等不同類型的文本素材進行分析，並結合資訊技術與空間分析方法，建置相應的文本內容檢索與分析系統，實證「文本空間化」在數位人文領域之應用模式。其中，對於《全臺詩》擷取詩人出生地與詩作中的地名，並藉其時空特性來解析詩人與詩作的相關性；在《蔣經國先生大事長編初稿》資料處理上，則結合時序性文本紀錄之轉換，以空間化與量化資料的圖表視覺化方式，來重構文本內容，分析其不同地點相互關連的發展脈絡；而針對《雷震日記》的文本特色，設計日記文本分析系統，以快速查找不同時期所記述地名及其相關內容，分析人名出現頻率與記主的社會網絡關係。上述實作也揭示了「文本空間化」於拓展數位人文研究視野的可能性。

關鍵字

文本分析；空間人文；自然語言處理；地理資訊系統

並列摘要

Textual content analysis explores and internalizes knowledge into a research methodology after reading related references. We live in the information age and should assess how to change the existing research methods. Due to the inspiration and influence of digital humanities, scholars with a traditional research philosophy could adopt many innovative ideas into their existing theoretical frameworks. It extends and enables increasing interdisciplinary collaborations through efficient research tools and systematic procedures. This study focuses on proposing a method of spatializing textual data. As case studies, we use three textual data types: The Complete Collection of Classic Taiwan Poetry, Chronological Memorabilia of Chiang Ching-Kuo, and Lei Chen's Diary. Based on the characteristics of these text types, combined with information technology and spatial methods, we construct relevant analysis systems to demonstrate textual data spatialization in digital humanities. First, for The Complete Collection of Classic Taiwan Poetry, we analyze the poets' birthplace and the place names in poems using information technology and spatial methods to explore the relationship between poets and poetry from the perspective of time and space. The context of interaction between places can be analyzed through graphical visualization methods that spatialize and quantify the data when dealing with the Chronological Memorabilia of Chiang Ching-Kuo. In addition, we establish a diary text analysis system based on diary text features to analyze Lei Chen's Diary. The place names and detailed context in the diary can be quickly retrieved. The above implementation also reveals the possibility of the "spatialization of textual data" in expanding the horizons of digital humanities research.

並列關鍵字

Textual analysis ； spatial humanities ； natural language processing ； geographic information system

參考文獻

李宗信、張育誠、劉庭羽（2021）。《劉福才日記》中的社會關係網絡—另一個觀看日記的視角。臺灣師大歷史學報，65，157-213。doi:10.6243/BHR.202106_(65).0004

張素玢、李宗翰、李毓嵐、李昭容、顧雅文、柯皓仁、謝順宏（2018）。從 CBDB 到 TBDB：以《新修彰化縣志．人物志》為試金石。數位典藏與數位人文，2，91-115。 doi:10.6853/DADH.201810_2.0005

羅鳳珠、白璧玲、廖泫銘、范毅軍、鄭錦全（2014）。唐代詩人行吟地圖：李白、杜甫、韓愈。圖書館學與資訊科學，40(1)，4-25。doi:10.6245/JLIS.2014.401/634

Che, W., Li, Z., & Liu, T. (2010). LTP: A Chinese language technology platform. In C.-R. Huang & D. Jurafsky (Eds.), Proceedings of the 23rd International Conference on Computational Linguistics: Demonstrations (pp. 13-16). Stroudsburg, PA: Association for Computational Linguistics. doi:10.5555/1944284.1944288

Garbin, E., & Mani, I. (2005). Disambiguating toponyms in news. In R. Mooney, C. Brew, L.-F. Chien, & K. Kirchhoff (Eds.), Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (pp.363-370). Stroudsburg, PA: Association for Computational Linguistics. doi:10.3115/1220575.1220621

主題瀏覽