Send message to Author

Paper short abstract

The goal of this paper is to reconsider the standard three-fold taxonomy of Genji manuscripts by shifting the unit of variant analysis from the whole work to the individual chapter. To this end, it attempts to produce exploratory “mappings” of the witness corpus on a chapter-by-chapter basis.

Paper long abstract

A myriad of variants may color the countless texts of the Tale of Genji that survive, but textual scholarship has discovered in this chaos a surprising degree of order. Specifically, there exist two clear textual lineages, the Teika-bon and Kawachi-bon lines, and additionally a group of texts belonging clearly to neither, conventionally called the “Beppon” (Other Texts) group. Applying this taxonomy, scholarly consensus today assigns extant texts of any of the Genji’s fifty-four chapters, through observation of their characteristic variants, to one of these three groups. At the same time, it has long been clear that the degree of variation itself varies significantly from chapter to chapter, raising doubts about the feasibility of applying any single framework to a group of subtexts so diverse. Yet given the sheer difficulty of surveying such a corpus, detailed discussion of each chapter’s unique lineage-assigning criteria, and in particular their evidential basis, has remained relatively minimal since the current consensus took hold.

In recent years, however, with the mass digital provision of ever more Genji manuscripts, it has become dramatically easier to consult the text of many specific important witnesses. This has opened up new opportunities for the reassessment of consensus, above all the possibility of refining the overarching taxonomic framework itself. Future advances in our understanding of the text will require deepening our grasp of its variance as encountered in actual witnesses, whose fundamental physical unit is not the full work, but rather the individual chapter. As such, supplementing the qualitative findings of traditional philology with the findings of quantitative analysis, our pressing task is to first produce “mappings” of the witness corpus on a strictly chapter-by-chapter basis.

The goal of this paper is to reconsider the received taxonomy through the exploratory production of precisely such “mappings” for a limited number of select chapters, based on statistical analyses of massive quantities of textual data. Focusing respectively on differences between the Kawachi-bon and the Beppon in early chapters, and in later chapters on differences between the Teika-bon and the Kawachi-bon, the paper throughout surveys various outstanding problems in manuscript classification.