Abstract
Background:
The traditional Chinese medical literature provides a substantial resource for natural products research. When undertaking systematic searches, investigators need to assess the scope, content, and relevance of collections both singly and in comparison.
Objectives:
This study examines eight collections of pre-modern literature, develops an approach to quantitatively assess their content, and compares their relative inclusivity.
Methods:
All publications listed in each collection were collated and a uniform scoring system was developed to account for variant editions, incorporation of multiple books under a single title, addition of commentaries, and other factors affecting content. Comparisons were undertaken between collections to determine the degree of overlap, genres of literature represented, and relevance to natural products research.
Results:
Following adjustments for differences in how books were listed, duplications, and variant editions, Zhong Hua Yi Dian (Encyclopaedia of Traditional Chinese Medicine) (4th edition CD) is the largest collection of complete books containing 1009 different books. Zhong Guo Ben Cao Quan Shu (The Complete Collection of Traditional Texts on Chinese Materia Medica) contains 2026 titles including multiple editions and extracts, so after adjustments the number of different complete books is approximately 740. No collection was fully inclusive of the others, but Zhong Hua Yi Dian includes between 52.1% and 91.5% of the books in the six smaller collections.
Conclusions:
Comparing traditional collections is complex due to variant editions and multiple titles. This necessitates examination of the text rather than title alone. Prior to undertaking systematic searches, the characteristics of collections need to be investigated, duplicated books identified, and differences between editions should be assessed. Its size, electronic format, and broad spread of genres makes Zhong Hua Yi Dian suitable for systematic searches, but due to the lack of bibliographical detail on included books, cross-referencing to other collections is recommended.
Introduction
Considering the time invested in searching and extracting information from traditional books, when undertaking searches of collections it is important to have prior knowledge of any duplications within a collection, the identity and origin of the included works, whether the books in one collection are the same as the versions in other collections, and the degree of overlap between collections. In this study, eight collections are examined in detail in order to accomplish the following: 1. Develop a method for comparing collections; 2. Identify issues requiring consideration when undertaking searches; and 3. Determine the degree to which collections overlapped.
The following eight collections were examined: 1. Zhong Hua Yi Dian (ZHYD) “Encyclopaedia of Traditional Chinese Medicine” 4th edition
2
2. Zhong Guo Ben Cao Quan Shu (ZGBCQS) “The Complete Collection of Traditional Texts on Chinese Materia Medica”
3
3. Zhong Yi Fang Ji Da Ci Dian (ZYFJDCD) “Great Compendium of Chinese Medical Formulas”
4
4. Si Ku Yi Xue Cong Shu (SKYXCS) “The Four Treasuries of Medical Works”
5
5. Ming Qing Ming Yi Quan Shu Da Cheng (MQMYQSDC) “Complete Collection of the Works of Famous Doctors from the Ming and Qing Dynasties”
6
6. Tang Song Jin Yuan Ming Yi Quan Shu Da Cheng (TSJYMYQSDC) “Complete Collection of the Works of Famous Doctors from the Tang, Song, Jin, and Yuan Dynasties”
5,7
7. Zhong Guo Yi Xue Da Cheng (ZGYXDC) “Great Collection of Chinese Medicine”
8
8. Huang Han Yi Xue Cong Shu (HHYXCS) “Collection of Imperial Han Medical Books”
9
All are in Chinese, seven are printed books, one (ZHYD) is on CD only, 2 two (SKYXCS and HHYXCS) have print and electronic forms, 5,9 –11 and one (ZYFJDCD) is an indexed compendium. 4
Materials and Methods
The procedure for investigating collections was as follows: 1.. Complete versions of each collection were accessed; 2. Constituent titles were listed in a spread sheet together with author and publication year (if available); 3. Title lists were compared with the books included in the collection and any issues investigated using reference works,
3,12
–21
other collections, and Internet searches; 4. Books were examined to determine whether they were complete, were divided into parts, or incorporated other books; 5. Scores for the number of books included within a collection were adjusted; 6. Issues relating to variant editions and commentaries were investigated and resolved; 7. Scores were totaled and crosswise comparisons undertaken; 8. Issues relating to relevance for herbal medicine were considered; and 9. Comparisons with the genres in ZHYD were undertaken.
Criteria for scoring
Since informational content was the key concern, variant editions were considered equivalent when their content was the same, irrespective of publication date, publisher, or other bibliographic considerations. Pre-modern was defined as written/published before 1950. Each book in each collection was examined individually and scored in the same manner. This required a consistent approach to defining and scoring the following features of collections: 1. Division and incorporation 2. Duplication and editions 3. Fragmentary works and extracts 4. Variant titles and authorship 5. Versions with commentaries and supplements 6. Errors and omissions 7. Genre and relevance to herbal medicine
Scoring was undertaken by two investigators (BM, SC), results were checked and any issues resolved by a third investigator (ALZ). Notes were kept on decisions, and examples are provided below and in Table 1.
Division and incorporation
Books could be included within other books or appear as separate titles in book lists. For example, SKYXCS and TSJYMYQSDC include separate listings for the Lei Jing “Classified Classic,” its illustrations (Lei Jing Tu Yi), and its appendices (Lei Jing Fu Yi), but the last two titles are combined under Lei Jing Tu Yi in ZHYD. In other cases, ZHYD splits up large books into all their constituent titles. For example, the well-known Qing dynasty work Yi Zong Jin Jian “Golden Mirror of Medicine” does not appear in the ZHYD table of contents, but each of its 15 constituent texts does. It would appear that this approach was to facilitate classifying these works into the genres of ZHYD, but the lack of the principal title can be confusing. When books are divided or combined, the baseline figures for comparisons between collections are altered. Thus, where one collection included a book separately, the approach was to score it as a separate book in all collections. For example, the complete Yi Zong Jin Jian was considered 15 books and Lei Jing was scored as 3 books.
Duplication and editions
In ZGBCQS, there are three editions of the Ben Cao Gang Mu, but these are essentially the same book, although the illustrations can vary. Therefore, for the purpose of comparison, these were regarded as the same work and counted only once. Five (5) editions of the Dian Nan Ben Cao appear in ZGBCQS and one in ZHYD. However, each of the five in ZGBCQS is a distinct book with different content, and the ZHYD version turned out to be a composite of five versions, so these were considered six separate books.
While it was expected that the same book in different collections might have variant titles, this phenomenon is surprisingly common within collections. ZHYD includes You Ke Zheng Zhi Zhun Sheng and Zheng Zhi Zhun Sheng: You Ke, both of which refer to the same book. Since duplications affect the results of systematic searches, these are all listed in Table 1.
In ZHYD, the Yi Zong Bi Du by Li Zhong-Ji was counted as two books, since it includes the Ben Cao Zheng Yao, but there is also a separate entry for Ben Cao Zheng Yao in the Materia Medica section of ZHYD; however, the layout of the two texts is different and there are differences in content, so these were counted as separate books.
Duplicated books were counted once only. All sections of the books were examined and when the content of the main text was the same, the books were considered duplications even when there were differences in prefaces, titles, punctuation, footnotes, illustrations, and minor differences in the order of sections.
Fragmentary works and extracts
ZGBCQS contains books that survive only as fragments and extracts selected by the editors as relevant to materia medica. In adjusting for duplications, when a partial work was found to be identical in content to a more complete work, it was treated as a duplicate within ZGBCQS, but when comparing collections, complete and incomplete books were scored separately. For the Materia Medica and Formula Book sections this was feasible, but for some books in other genres it proved difficult to determine whether the book was complete or not, since no other version could be obtained for comparison, so scores are only reported for these two genres (Table 2).
No. books published post 1949.
comp. bks, complete books; MM, Materia medica genre; F, Formula book genre; est, estimate; ZHYD, Zhong Hua Yi Dian
‘Encyclopaedia of Traditional Chinese Medicine, 4th edition’; SKYXCS, Si Ku Yi Xue Cong Shu
‘The Four Treasuries of Medical Works’; ZGYXDC, Zhong Guo Yi Xue Da Cheng
‘Great Collection of Chinese Medicine’; TSJYMYQSDC, Tang Song Jin Yuan Ming Yi Quan Shu Da Cheng
‘Complete Collection of the Works of Famous Doctors from the Tang, Song, Jin and Yuan Dynasties’; MQMYQSDC, Ming Qing Ming Yi Quan Shu Da Cheng
‘Complete Collection of the Works of Famous Doctors from the Ming and Qing Dynasties’; ZYFJDCD, Zhong Yi Fang Ji Da Ci Dian
‘Great Compendium of Chinese Medical Formulas’; HHYXCS, Huang Han Yi Xue Cong Shu
‘Collection of Imperial Han Medical Books’; ZGBCQS, Zhong Guo Ben Cao Quan Shu
‘The Complete Collection of Traditional Texts on Chinese Materia Medica’; ATN, Acupuncture/Tui Na books; T/NNR, total no. of ATN books/no. ATN books not relevant to herbal medicine.
ZYFJDCD comprises extracts rather than full books, so when title and author corresponded to those in another collection, this was counted as an instance of commonality. Doubts were resolved by comparing the texts of extracts. Nevertheless, there is a greater likelihood of error in the results for ZYFJDCD compared to other collections.
Variant titles and authorship
The same book may appear under variant titles, and the same author can have multiple names. For example, the Ling Lan Yao Lan in MQMYQSDC is the same as the Chong Ding Ling Lan Yao Lan in ZHYD. Similarly, the Yan Hou Mai Zheng Tong Lun by Xu Shan-Lin appears in ZHYD and the Chong Ding Yan Hou Mai Zheng Tong Lun by Xu Lian is in ZGYXDC. Both names refer to the same author and the content of the books is the same.
In contrast, ZHYD contains Xue An Bian Shu by Xue Ji, while SKYXCS contains Xue Shi Yi An, also by Xue Ji. Since the author was the same, and both works were in the case history genre, it appeared they were the same book. However, examination of the texts revealed quite different layouts with some overlapping content but also major differences, so they were considered different books. The approach taken to variant book titles was to regard them as the same book when the main texts had the same content.
Versions with commentaries and supplements
Books may contain commentaries or other additions to the original text. For example, ZHYD contains the Jin Gui Yao Lue Fang Lun by Zhang Ji and the Jin Gui Yao Lue Qian Zhu by Chen Shou-Yuan. These have the same basic content, but the second has additional commentary. Similarly, the Wai Ke Zheng Zong by Chen Shi-Gong appears in ZHYD and ZGYXDC contains a book titled Xu Ping Wai Ke Zheng Zong, which is the same book plus a commentary by Xu Ling-Tai. SKYXCS contains the Zheng Lei Pu Ji Ben Shi Fang by the Song dynasty author Xu Shu-Wei, whereas ZHYD contains the Lei Zheng Pu Ji Ben Shi Fang Shi Yi, which is a Qing dynasty version of this book with a commentary by Ye Gui. These books have similar content but there are a number of differences. The approach taken was to score books whose content varies due to addition of a commentary to the main text (excluding prefaces and modern footnotes) as separate works.
More complex examples
The Yang Lao Feng Qin Shu, which is attributed to the Song dynasty author Chen Zhi, and the Shou Qin Yang Lao Shu, with no author listed, both appear in ZHYD. The second book has a preface containing the name Chen Zhi, so these texts were examined for similarity. There is considerable overlap in content but differences in arrangement, so they were considered to be different works. SKYXCS contains a book entitled Shou Qin Yang Lao Xin Shu, but it is attributed to the Yuan dynasty author Zou Xuan, so this appeared to be a later commentary on Shou Qin Yang Lao Shu. However, the first part is identical to the Yang Lao Feng Qin Shu, and the remainder is additional. So, as it turned out, there was overlap in content between the books that had differing titles, rather than between those with similar titles. However, due to their differences, the three books were considered as separate.
Errors and omissions
In ZHYD the book entitled Ben Cao Zheng is attributed, incorrectly, to Zhang Shan-Lei but it is identical to the Ben Cao Zheng, which forms part of Jing Yue Quan Shu. Zhang Shan-Lei is the author of the similarly named Ben Cao Zheng Yi, which is included in ZHYD as a separate entry. Discrepancies were found between the titles listed in the contents and the books included in ZHYD and ZGBCQS (Table 1). Comparisons were based on corrected lists of included books.
Genres of books
The genres of Chinese medicine (CM) books are not well defined and the decision criteria for including a book in a particular genre are not provided in ZHYD or ZGBCQS. For example, Qian Jin Yao Fang is included under “Comprehensive medical works” in ZHYD but in ZGBCQS it is classified under “Formula Books.” Since the ZHYD genres encompass a broad spread of common genres, these were used as the basis for comparison. When a book in another collection was also included in ZHYD, it was scored under the corresponding genre (Table 3). When books had received scores greater than 1 due to adjustments for splitting, the components were scored under their respective genres. It was not feasible to classify books not included in ZHYD.
Titles/complete books.
Relevance to herbal medicine
Since this study was undertaken from the viewpoint of natural products research, books on acupuncture and other manual methods such as tui na (massage) would not appear relevant. However, some such books also contain content on herbal medicine (for example, Zhen Jiu Feng Yuan and Tui Na Jue Wei also include herbal formulas). Similarly, some books on diagnostics such as Mai Jing “Classic on Pulse Diagnosis” contain major components on herbal treatments. Therefore, genre alone is not a good measure of relevance to herbal medicine.
The above examples are only a selection of the numerous instances in which the full texts were examined in detail to determine whether the books should be considered to contain the same content. It was found that it was not sufficient to compare books in terms of titles alone or even in terms of title plus authorship; the actual text needed to be examined.
Results and Discussion
ZHYD has 10 duplicated books and after adjustments for division and errors, contains 2009 different books (Table 2). ZGBCQS lists 2027 titles but contains approximately 740 complete books. The Materia Medica section contains multiple editions, so it contains 293 different books, of which 282 are complete. Of the 16 works in the Formula Book section, 3 were considered incomplete since they are fragmentary works recovered from archaeological sites such as Ma Wang Dui. In ZYFJDCD, 31 of 715 books were not written or published prior to 1950, as was 1 book in ZGBCQS.
In ZHYD, of the 51 books in the Acupuncture/Tui Na genre (ATN), 22 were considered to have no content directly relevant to herbal medicine. Similarly, no relevance to herbal medicine was found for 8 of 43 books in the Diagnosis genre, 12 of 40 in Health Enhancement, and 1 of 6 in “Other.” So 43 of 1009 books were considered not directly relevant to CHM, but even among these, the discussions on disease names and symptoms may be of interest. Since relevance was difficult to calculate, comparisons between collections were based on adjusted totals without exclusions relating to relevance, and scores for ATN are reported separately in Table 2.
Comparisons between collections
Overlap was greatest between ZHYD and MQMYQSDC with 214 of 234 books in common (91.5%), but was much lower for its companion collection TSJYMYQSDC at only 67.6%. It had been expected that the smaller, older collections would be subsumed within ZHYD but 86.5% of SKYXCS, 82.0% of ZGYXDC, and only 52.1% of HHYXCS was included. Only 395 (55.2%) of the books referenced in ZYFJDCD are also included in ZHYD. When the 31 modern books are excluded, this increases to 57.7%. Based on adjusted figures, ZGBCQS and ZHYD have 153 titles in common, but only 67 of these are complete books.
Comparisons with ZHYD genres
The smaller collections SKYXCS, ZGYXDC, and HHYXCS show overlap with ZHYD across almost all genres, with the largest number of books being in the “Clinical & specialty” genre. In MQMYQSDC, this genre also dominated but there are only seven books in TSJYMYQSDC. This may be because this genre expanded greatly from the Ming dynasty onwards. Only a little more than half the books in ZYFJDCD (395) are represented in Table 2, but 73 of the 117 “Formula books” in ZHYD (62.4%) are common to both. There is also substantial overlap in the “Clinical & specialty” (148) and “Comprehensive medical books” genres (65) and also some books in each of the other genres.
Considering the size and scope of ZGBCQS, it was expected that all 67 of the Materia Medica books in ZHYD would be included in ZGBCQS, but the overlap is only 48 books. In a few cases, different versions are in both collections, for example, the Wu Pu Ben Cao in ZHYD is not the same as the Wu Shi Ben Cao in ZGBCQS; all the versions of Dian Nan Ben Cao differ; and ZHYD contains a separate Ming Yi Bie Lu, the content of which is incorporated into Ben Cao Jing Ji Zhu in ZGBCQS. However, the remainder are different books.
Although there is considerable overlap in the contents of these eight collections, none was completely subsumed and even the genre-focused ZGBCQS and ZYFJDCD were not comprehensive within their genres. When all the full text collections were considered, including the ZGBCQS Materia Medica and Formula Book genres but excluding ZYFJDCD, 1346 different complete books were identified based on adjusted figures. Therefore, ZHYD contains 75% of the books found in the other six collections of whole books, making it both a large and representative sample of the CM literature. Considering its searchable digital format, it is an attractive option for systematic searches but it still suffers the drawback of lack of good bibliographical data on the included books. This deficiency can often be remedied by reference to the print collections discussed here, but these still cannot account for 337 books. In comparison, ZYFJDCD also references a large number of books, 45% of which are not in ZHYD, and is amenable to searching via its indices. It is, however, a collection of extracts and shares the drawback of lacking good bibliographical data on the included books. So, again, full text collections are needed for reference.
Conclusions
Comparing traditional collections is complex due to variant editions and multiple titles. This necessitates examination of the full texts to determine whether books are the same or different. Following adjustments, the largest collection of full books is the ZHYD CD (4th edition) with 1009 different books. Although ZGBCQS contains more than twice the number of titles (2027), many are incomplete. Nevertheless, with approximately 740 complete books, it remains the largest collection in the materia medica genre. None of the eight collections completely subsumes any of the others, but ZHYD includes between 52.1% and 91.5% of the books in the six smaller collections across a variety of genres and 75% of the total number of complete books identified. Its electronic format makes it suitable for systematic searches. Nevertheless, when one undertakes searches, duplicated books need to be identified and the results should be adjusted. Also, the lack of bibliographical detail on included books means that cross-referencing to other collections is recommended.
Footnotes
Acknowledgments
We thank the Preventive Health Flagship Program, Commonwealth Scientific and Industrial Research Organization (CSIRO) for providing partial funding support. The project is also partially supported by an International Grant from the Guangdong Provincial Academy of Chinese Medical Sciences, Guangdong Province, China. We thank Dr. Angela Yang for her work in data extraction and compilation, and David Lu and Daniel Rechter for assistance in resolving the numerous questions. We also wish to thank the library staff at Hong Kong Baptist University and at the Guangzhou University of Chinese Medicine for their kind assistance with the literature search process.
Disclosure Statement
The authors declare that they have no competing interests.
