The present study applies character-frequency stylometry and principal component analysis (PCA)to a comparison of the Kumārajīva and Xuanzang translations of the Diamond Sutra (VajracchedikāPrajñāpāramitā Sūtra). Using overlapping fixed-length textual segments drawn from digitized CBETAeditions, the study examines whether statistically detectable stylistic distinctions can be identified betweentwo parallel translations of the same Buddhist scripture. The analysis shows that the two translations formdistinguishable, though partially overlapping, clusters in PCA space despite their broadly similar doctrinalcontent. Principal component loadings reveal that Kumārajīva's translation is more strongly associated withcanonical dialogue markers and formulaic scriptural discourse, whereas Xuanzang's translation exhibits agreater concentration of connective narration, expository prose structure, and technically precise Buddhistterminology, including a stronger retention of transliterated Indic expressions. Shannon entropy analysisconfirms that the two translations are broadly comparable in overall lexical diversity, indicating that thestylistic separation identified by PCA reflects differences in discourse organization and terminologicaldistribution rather than global differences in textual complexity. The findings suggest, within this limitedparallel corpus, interpretable stylometric analysis can indicate translator-associated differences that meritfurther philological comparison.
Kuo‐Chin Jong (Mon,) studied this question.