Identifying Keywords in the Buddhist Canon

This paper discusses approaches to keyword identification in Buddhist texts. The paper first defines what is meant by a keyword and how it differs from terminology and other related concepts. Systematic approaches to keyword extraction with natural language processing methods are described as well as visualization techniques to help researchers identify keywords. Examples from the Heart Su̅tra are explored followed by extension to the larger body of the Taisho̅ Tripitaka. The use of term frequency, co-occurrence networks, and collocations is also discussed. The results show that term frequency filtered by terms labelled as Buddhist gives a reasonable set of keywords with subjective assessment based on linguistic and stylistic arguments. Ranking of collocations can indicate predominant word senses. Finally, potential future areas of application of the tools and methods are discussed.