When Che Xilun, preeminent Chinese scholar of baojuan (and incredibly kind man, if our email correspondence is anything to go on), compiled his catalogue of baojuan in China, Zhongguo baojuan zongmu 中國寶卷總目, he did so in part by collating previous catalogues into composite entries for each title. Given that his catalogue includes entries for over 1500 different baojuan, it is unreasonable to expect that Che would have personally examined every text included in his work. This means, of course, that he had to choose whether to trust previous cataloguers’ work or correct what seem to be errors to streamline the entries in his updated work.

Working from his entry on Liu Xiang baojuan has taught me that these inconsistencies are often because the historical reality of these texts’ publication was more complicated and fragmented than neat lists have so far managed to detail. I don’t fault Professor Che for this. His catalogue is the best starting point from which to begin doing research on the textual and publication history of baojuan. I couldn’t have done much of what I’ve been able to do without it as a guide. I wouldn’t have known which libraries to visit in China on my tactical short term visits there. I would be wholly without reference points in an understudied, messy field.

But his catalogue is still best as a starting point. In the field, things get a little more complex.

For example, it says that the earliest extant edition of Liu Xiang is from 1774, and is kept at the library of the Chinese National Academy of Art. I’ve been there twice. Their card catalog, yes, a literal card catalog, is from three separate collections, organized by different logics, stored together in two cabinets. Since I couldn’t find a card for the 1774 edition, I couldn’t request it from the closed stacks. Blindly requesting call numbers adjacent to baojuan I could find in the catalog actually turned out to be more beneficial to my research than searching the musty cards, but still no 1774 edition. Che’s entry for the edition comes from an earlier catalogue, published in 1961. Perhaps that entry comes from even earlier? I’ve yet to track that down.

Moving forward. Che’s catalogue lists one edition from 1833, published by a certain Shen Maozun. Rostislav Berezkin, another friendly baojuan researcher, sent me some photos he took of an edition from 1833. Now, Berezkin researches baojuan from a different era than I do, so for him to have taken any photos of an early edition of Liu Xiang baojuan and shared them with me was a godsend. The fact that it was from the private collection of the 1961 catalogue compiler made it all the better.

But when I started looking at them again, I realized that something was amiss with this edition’s entry in the catalogue as well. Namely that the 1961 catalog listed two 1833 editions, while Che’s lists only the one. The pages Berezkin photographed of the 1833 edition don’t show a date, but his research is of such a detailed nature that I trust his notes.

Another of my research tactics is to periodically browse Chinese auction websites for old books that come up. Liu Xiang baojuan being a popular text 100-200 years ago, quite a few editions do emerge. Most are familiar ones from my archival work, but when I see an unfamiliar one, I save what images of it I can and file it away for future reference. Most of these I will never be able to identify and are thus just a magpie-like collection of shiny things.

But when I hit upon the issue of two 1833 editions rather than just the one, I went back through all the auction site images, just in case. This image, which I saved as “supposedly 1833 edition,” didn’t look at all like the images Berezkin sent me, which had previously made me assume that the auction site’s attribution was false.

But armed with the 1961 catalogue telling me that there might be two 1833 editions, I took a closer look at it.

There, after flipping the picture, and messing around with brightness and contrast, the character Shen bleeds through from the title page, as does the date: 1833. The 1961 catalogue was right, there were two editions! Berezkin has personally seen one, and I fear that the other one is now lost somewhere in a private collection in China. The data, just a photo from an auction site, isn’t much to go one, but it’s more than I had before! And, if the 1961 catalog was right about two 1833 editions, then maybe it was also right about the existence of the 1774 edition. A researcher can hope, right?

Moral: Research is messy and more complicated than I might want it to be, but people out there, for the most part, are kind. And sometimes the messiness is much more interesting!