Classification

11 January 2012

At this year’s EASL conference in Prague in early September, amongst the interminable talk of e-resources and the disappearance of budgets, it was a pleasure and a surprise to hear Marc Winter (from the Ostasiatisches Seminar in Zürich) give a succinct presentation on the background of the sibu 四部 classification system. At one time, this would have been the common knowledge of anyone who studied China. Now, it is almost esoteric.

When the web interface to my pre-modern catalogue is up and running, everything it contains will be accessible through a search 檢索 mechanism. But the whole will be presented in conspectus by an arrangement very similar to the browse 瀏覽 mechanism found in the International Union Catalogue of Chinese Rare Books 中华古籍善本国际联合书目系统 recently mounted by the National Library of China (this data is actually derived from Soeren Edgren’s Chinese Rare Books Project). Here, as in all catalogues of pre-modern Chinese books, whether electronic or printed, the data is presented and made comprehensible through sibu classification, a knowledge of which is indispensable to anyone who studies Chinese bibliography.

Inspired and informed by Marc’s timely presentation, I will first document the main stages in the development of the system. In a second posting I will deal with modern uses of the system, and how it may be adapted to embrace materials for which it was not originally designed.

In 26 BC, Emperor Cheng of the Han Dynasty 漢成帝 ordered the scholar Liu Xiang 劉向 to organise the imperial book collection and produce a summary catalogue of it. The catalogue was called Bielu 別錄 (the term bie 別 is taken to mean that the works were separated out into different categories – or classified, as we would now say), and consisted of 20 juan 卷. Liu Xiang died in 8 BC, and the Emperor Ai 哀帝 ordered his son Liu Xin 劉欣 to take over his father’s job as imperial librarian. Liu Xin moved all the books into a new library called the Tianluge 天祿閣, and summarised and re-shaped his father’s work into what is now regarded as the first classified Chinese book catalogue, the Qilue 七略 (“seven summaries”), which was presented to the throne in 6 BC. In the preface to his Taoist books in the libraries of the Sung period (London, 1984), Piet van der Loon takes the completion of the Qilue as the beginning of Chinese bibliography.

Both the Bielu and the Qilue were lost at the end of the Tang Dynasty, but the Qilue had been used by Ban Gu 班固 to compile the bibliographical treatise of the Hanshu 漢書藝文志 (漢書, 卷30), which was completed by his younger sister Ban Zhao 班昭 in AD 111, so both its arrangement and contents are preserved. Also, during the movement to reconstitute lost works that preoccupied a number of Chinese scholars in the first half of the 19th century, a few fragments quoted in other texts were reassembled and published, notably by Ma Guohan 馬國翰 in the largest of such endeavours, his collectaneum entitled Yuhanshanfang jiyi shu 玉函山房輯遺書.

The Hanshu treatise lists all the books that had reached the the imperial library by the end of the Former Han Dynasty (AD 8), and is arranged in Liu Xin’s sixfold classification (the seventh “summary” is in fact the general introduction), which is as follows:

1. 六藝略: 易、詩、書、禮、樂、春秋、論語、孝經、小學 – 9 classes, 103 titles.
2. 諸子略: 儒家、道家、陰陽家、法家、名家、墨家、縱横家、雜家、農家、小說家 – 10 classes, 189 titles.
3. 詩賦略: 屈赋之属、陸赋之属、荀赋之體、雜赋、歌詩 – 5 classes, 106 titles.
4. 兵書略: 兵權謀、兵形勢、兵陰陽、兵技巧 – 4 classes, 53 titles.
5. 數術略: 天文、歷譜、五行、蓍龜、雜占、形法 – 6 classes, 190 titles.
6. 方技略: 醫經、醫學方、房中、神仙 – 4 classes, 36 titles.

From this it is clear that the major elements of the sibu system are already in place, but differently named and differently arranged. What would later be called “classics” (jingbu 經部) comes first, with its major divisions more or less are they are now. The “philosophers” (zi 子) are also there, as are “belles-lettres” (ji 集). Only “history” (shi 史) is absent

The earliest codification of the sibu system in its present form is attributed to the Wei 魏 Dynasty official Zheng Mo 鄭默 (AD 213–280) in his Zhongjing 中經, which was used bu Xun Xu 荀勗 (d.289) to compile his Zhongjing xinpu 中經新簿, in which all books were listed in the four-part classification jia 甲, yi 乙, bing 丙, ding 丁, corresponding to the later jing 經, zi 子, shi 史, ji 集. The present order (switching shi 史 and zi 子) was established by Li Chong 李充 (fl.320s) in his Jin Yuandi shumu 晋元帝书目 (sometimes called Jin Yuandi sibu shumu 晉元帝四部書目). All three texts cited in this paragraph are lost, and their content known only in later descriptions; only fragments of Xun Xu’s Zhongjing xinpu are preserved in Ma Guohan’s collectaneum referred to above.

In the bibliographical treatise in the Suishu 隋書經籍志 (隋書, 卷32-35), compiled at imperial behest in AD 636 by Wei Zheng 魏徵, the system was given its current nomenclature, and is almost in the form in which it was to remain until the present day. The only significant difference was in the treatment of Buddhism and Taoism. These were not included in “philosophers”, but appeared in separate sections at the end with only the total number of titles given, not the names of texts.

The arrangent of the Suishu bibliographical treatise is as follows:

1. 六藝經緯: 易、尚、詩、禮、樂、春秋、孝經、論語、異說、小學 – 10 classes, 627 titles.
2. 史: 正史、古史、雜史、霸史、起居注、舊事、職官、儀注、刑法、雜傳、地理、譜系、簿錄 – 13 classes, 817 titles.
3. 諸子: 儒、道、法、名、墨、從橫、雜、農、小說、兵、天文、曆數、五行、醫方 – 14 classes, 852 titles.
4. 集: 楚辭、別集、總集 – 3 classes, 554 titles.
5. 道經: 經戒、餌服、房中、符錄 – 4 classes, 377 titles (but not listed).
6. 佛經: 大乘經、小乘經、雜經、雜疑經、大乘律、小乘律、雜律、大乘論、小乘論、雜論、記 – 11 classes, 1950 titles (but not listed).

The final step of incorporating Daoist and Buddhist works into the sibu 四部 system was taken in the Northern Song 北宋 Dynasty, some four centuries after the appearance of the Suishu treatise, with the completion of the Chongwen zongmu 崇文總目, which was submitted to the throne in 1042. This is catalogue of the imperial archive (called the Chongwenyuan 崇文院), and was produced by a team of scholars led by Wang Yaochen 王堯臣. The catalogue contains 3,445 titles from the early Song.

The apotheosis of the system took place in the 18th century with the compilation of the catalogue of the imperial manuscript library Siku quanshu 四庫全書, which was first submitted to the throne in 1781. The first edition of this work was published in the imperial palace in 1789, and is entitled Qinding siku quanshu zongmu 欽定四庫全書總目 (in 200 juan), but modern editions often use the title Qinding siku quanshu zongmu tiyao 欽定四庫全書總目提要 or a shorter version of it, such as Siku tiyao 四庫提要. This is because it is a catalogue raisonné which not only lists the works, but also described their contents in some detail (these descriptions are the tiyao 提要, which means “noting the essentials”).

Three official editions of the catalogue were produced during the Qing Dynasty. The palace edition of 1789, which is the best; a re-cutting of it in Hangzhou 杭州 in 1795, which is the most widely distributed; and a re-cutting of the Hangzhou edition in Canton in 1868. To these are added a number of commercial reprints, which are the editions normally found in western libraries.

The four divisions now embrace 44 classes, containing not only the 3,461 titles that that were copied into the manuscript library, but a further 6,793 titles that were considered important, but not meriting inclusion (the cunmu 存目). A total of 10,294 were thus catalogued, classified, and described. But large as this figure sounds, it is less than double the number of titles in the Suishu treatise produced over a millennium earlier, and by no means represents the totality of the national literature of the time. The arrangement of the catalogue is not significantly different from that of the Chongwen shumu of seven centuries earlier.

One Response to “Classification”

  1. Endymion Wilkinson Says:

    David, Although the Suizhi is the first surviving catalogue to use the Sibu (Jing, Shi, Zi, Ji) classification system, it was evidently in use at least half a century before as mentioned by Yan Zhitui 顏之推 (please see my outline of the development of the Sibu system in Chinese History: A New Manual, Harvard, 2015, section 71.2.2. Endymion


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: