Abstract：This essay argues that the excerpts of Bai Juyi’s poems collected in Wakan Rōeishū (《和汉朗咏集》) provide new evidence for reconstructing this anthology’s compilation process, and for investigating the textual history of Bai Juyi’s poems, and for the overseas pathways of its transmission.There is a series of systematic textual correspondences which links the excerpts collected in Wakan Rōeishū to the seventy-volume edition of the Baishi Wenji (《白氏文集》) which has been circulating in Japan. The scope of selected passages, their sequencing, thematic classifications, and the scroll assignments in the former largely conform to the ordering and internal regimen characteristic of the seventy-volume compilation. It is found that Fujiwara no Kintō’s selection strategy inherits the editorial practice exemplified by Senzaikaiku (《千载佳句》), which uses the Baishi Wenji in the Heian period as a lexicographic repository of poetic lines. The selected material also shows a clear preference for seven-character verse over five-character verse. In addition, the inclusion of song lyrics, prefaces, records, and fu compositions suggests that contemporary readers’ literary tastes and scope had undergone certain changes.Further textual tracing reveals that, compared with the current circulating editions of the Baishi Wenji, the excerpted lines of the excerpts from Bai Juyi’s poems preserved in the Wakan Rōeishū are textually closer to those of the old manuscript copies of the Baishi Wenji. The lines that coincide with ones found in the handwritten manuscript of Changhen ge （《长恨歌》） by Prince Son’en are identical, suggesting that they may have derived from the “imperial copies” transmitted through successive imperial households, sharing the same origin with the “seventy-volume edition of the Wenji” preserved in the hereditary tradition of the hakase families. The exchange of books between the court collections and Fujiwara family libraries facilitated the sustained preservation and use of the seventy-volume edition of the Baishi Wenji among the Japanese aristocratic circles. Although Northern Song printed editions of the Baishi Wenji had already been transmitted to Japan by that time, they did not replace the Tang manuscript copies. Instead, the two traditions coexisted, giving rise to a multi-lineage textual ecology.The distinctive textual ecology formed by printed and manuscript editions in Japan was certainly related to the rarity of Song printed editions during the Heian period. More importantly, however, it stemmed from the cultural practices whereby hereditary scholarly households transmitted their learning from generation to generation on the basis of Tang manuscript copies. By the late Heian period, ?e no Masafusa had begun to consult the Northern Song printed editions and used them to collate and annotate the relevant poetic excerpts. The differing attitudes toward the Baishi Wenji adopted by Fujiwara no Kintō, the compiler of the Wakan Rōeishū, and by ?e no Masafusa, its annotator, reveal the broader process by which Heian-period Japanese literati read, collated, and annotated Tang poetry.Although China and Japan belonged to the same Sinosphere, the formation and transmission of literary classics did not necessarily follow identical principles or follow the same path in the two countries. In China, Tang manuscript copies, including those of the Baishi Wenji, were superseded by Song printed editions for various reasons. Japan, by contrast, continued to regard Tang manuscripts as authoritative textual standards. This was closely related not only to the distinctive modes of intergenerational transmission within the scholarly learning of the hakase families and to the historical circumstances that made Song editions difficult to obtain, but also, conversely, to the fact that after Tang manuscripts were transmitted eastward to Japan, anthologies such as the Senzaikaiku and the Wakan Rōeishū were successively compiled. Based on Tang manuscripts, these two anthologies came to function as new classics, replacing individual collections and comprehensive anthologies as principal vehicles through which Japanese literati of the time and later periods studied Tang poetry. When ?e no Masafusa subsequently incorporated Song printed editions into his critical collation and annotation of Tang poetry, and medieval Japanese literati expanded their reading to include Su Shi and Huang Tingjian, this transition from compilation and selection to collation and annotation reflected the actual circumstances of the book exchange between China and Japan. It also demonstrates how overseas readers engaged with Chinese texts during the period when Song printed editions and Tang manuscripts converged and interacted in Japan, thereby reconstructing the overall trajectory of the book circulations between China and Japan in that era.
刘莹. 《和汉朗咏集》的编纂与《白氏文集》七十卷本关系考论[J]. 浙江大学学报(人文社会科学版), 2026, 56(8): 60-74.
Liu Ying. The Compilation of Wakan Rōeishū and Its Textual Relationship with the Seventy-volume Edition of the Baishi Wenji. JOURNAL OF ZHEJIANG UNIVERSITY, 2026, 56(8): 60-74.
2026, Vol. 56 
