-
Prosopography and its Potential for Middle Period Research
Workshop on the Prosopography of Middle Period China: Using the China Biographical Database
(University of Warwick, December 13–15, 2007)
Those of us whose research touches in some way on the period between the Five Dynasties and the late Ming dynasty have now for a long time heard about the existence of a database, compiled by Robert Hartwell (1932–1996), said to include the biographical data of thousands of Middle Period China civil servants. After Hartwell's death, his estate passed into the care of the Harvard Yenching Institute, where Professor Peter Bol initially took on the challenge provided by the geographical data in the estate. Hartwell's geographical data inspired the development of the China Historical GIS.1 In contrast, the biographical data in the estate, designed by Hartwell to be used in conjunction with the geographical information, initially received less attention. Questions posed at the annual meetings of the Song and Conquest Dynasties Group only seemed to confirm that this database would not be ready for use soon. Behind the scenes, however, Professor Michael Fuller worked tirelessly at developing the biographical data included in what became known as the China Biographical Database (CBDB). [End Page 161]
When Hartwell began to compile biographical data—initially including only officials serving in the financial administration of the Song dynasty, but later including other Middle Period civil servants as well as their associates—he designed his own database program, over time repeatedly migrating the data to new generations of programs. Fuller's tasks included structuring this highly complex data and making it available for users first in FoxPro and then in MS Access. More crucially, however, Fuller sought to redesign the tables in the database so as to ensure their usefulness for Middle Period scholars, and here an impasse emerged: it seemed impossible for scholars unfamiliar with the database to indicate how the data should be organized and what questions they might wish to ask of the data without actually using the data, yet Fuller could not prepare the database for wider use without making some decisions on its design that would have long-term implications. There were other problems, too: it was difficult to judge the accuracy of the data without extensive checking of that data, and hard to decide how to deal with the obvious gaps in the coverage of the database without a clear sense of its future usefulness. Moreover, it seemed important to know what lessons might be learned from comparisons with similar prosopographical databases currently under development.
While Fuller continued his work on database design, Bol concentrated on developing collaborations between the Harvard-Yenching Institute, the Center for Research on Ancient Chinese History at Peking University, and the Institute of History and Philology at Academia Sinica in Taibei to add new content to the database. Meanwhile, Naomi Standen (Newcastle University) and I secured funding from Warwick University to facilitate systematic comparisons between the China Biographical Database and similar databases in the UK.2 The full report this funding made possible, written by Andrew Wareham and Artemis Papakostouli, then both of the Centre for Computing in the Humanities at King's College, London, entitled "Prosopographical projects in the pre-modern world: CHS and its counterparts," can be downloaded from the website of the Society for Song, Yuan and Conquest Dynasties Studies (http://www.sungyuan.org).
These developments, useful as they were, did not provide a way out of the impasse: without actually using the database it remained impossible to know [End Page 162] how one might wish to use the database, but without that knowledge Fuller could make little progress in preparing the database for wider use. With this in mind, I conceived of the idea of organizing a small workshop at Warwick University. My plan was to provide access to the database in its preliminary (MS Access) form to a number of junior scholars, who would be asked to carry out small research projects using the biographical data, and to invite a number of senior scholars to comment upon these papers. Taken together, these research projects and the discussions they would generate could then serve to inform those involved with designing and developing the database. Both the British Academy in London and the Chiang Ching-kuo Foundation in Taibei were willing to fund this workshop, and the event took place on December 14–15, 2007. Abstracts of eleven of the papers presented at the workshop follow this introduction, and full versions of these papers are available at the website of the Society for Song, Yuan and Conquest Dynasties Studies.
In the remaining pages, I will briefly discuss the papers, highlight some of the changes that have been implemented as a result of the workshop, and offer some reflections on the implications of these projects for research in Middle Period China. It is hoped that these papers will generate interest in the database and help colleagues negotiate the still rather treacherous terrain of Middle Period prosopography.3
The Technique of Prosopography
So what exactly is prosopography? The term primarily refers to a listing of persons who share certain characteristics: they might all be of a certain rank, have the same occupation, belong to the same period, or share a connection to a single individual. For example, Debra Nails' 2002 study, The People of Plato, provides a listing of every person referred to in the Platonic dialogues, including a short life with information where available on the career, the family, and the writings of each person. Nails has written her book for those who teach and research the philosophy of Plato; her painstaking work carried out over many years and published in a single volume makes further [End Page 163] prosopographical analysis possible.4 Here we encounter the term prosopography in a second, "applied" sense, referring to studies where prosopographical methods have been used. Such methods include manipulating and analyzing data gathered in a prosopographical listing to acquire further knowledge and insights.5 It is in this sense that Lawrence Stone discusses it in his now classic eponymous study of the subject.6 In a third meaning, used for example by the Islamicist Chase Robinson, prosopography is understood to indicate the biographical characteristics of individuals that mark their membership in a given social group, as opposed to the characteristics that make them stand out as individuals.7
We owe Hartwell a huge debt of gratitude for creating a prosopography for our use. The task of collecting the biographical data of tens of thousands of individuals he undertook to produce this prosopographical listing, allows us to move straight on to using prosopographical analysis as a research tool. That said, we have not yet resolved how best to use this tool for Middle Period research. Prosopography relies on asking the same set of questions of a large group of people so as to further our understanding of the members of that group. Its success depends on the formulation of an effective set of questions, part of the basic training of any social scientist, but a challenge to those of us who work mostly in humanities-based disciplines. So what questions to ask? To give an example from a different field, in her study of women traveling to the Middle East in the long eighteenth century, Billie Melman asks about each woman's family status, her social origins, her occupation, the nature of her travel, and the kind of publication her travel gave rise to so as to understand [End Page 164] the structures and patterns that shaped their experiences.8 As she readily admits, to understand these experiences themselves and to see the social and cultural dimensions of the practice, we cannot do without textual analysis. Our readings of texts, however, are all too often shaped by assumptions, in Melman's case about the social status of such women travelers or the kind of texts they produced. Quantitative data will not supersede the need for textual analysis, but can provide sometimes surprising insights that change the ways we read these texts.
The Warwick Workshop
We were fortunate to be able to invite colleagues who have expertise in working with a number of other online databases. These included: Dr. Alex Burghart, Research Officer for the Prosopography of Anglo-Saxon England (PASE http://www.pase.ac.uk/); Dr. Andrew Wareham, associated with the Centre for Computing in the Humanities at King's College, London, and more recently Director of the Hearth Tax Project based at Roehampton University; Professor Grace Fong, Project Editor of the Ming Qing Women's Writings Database (MQWW http://digital.library.mcgill.ca/mingqing/); and Dr. Dagmar Schaefer of the Max Planck Institute in Berlin (http://www.mpiwgberlin.mpg.de).
Dr. Burghart demonstrated the way in which the PASE database works, highlighting the fact that it aims "to cover all of the recorded inhabitants of England from the late sixth to the end of the eleventh century."9 All existing sources including inscriptions and coins are combed through for references to named individuals, and each assertion about an individual in any of these sources (a 'factoid') is included in the database. A total of over 84,600 factoids leads to records for about 11,700 persons, based on over 2,000 sources.10 In comparison, there are over 34,200 persons listed in the main biographical table of the 2007 Access version of the CBDB database.11 A significant number of sources has been used in the compilation of the CBDB, but by no means all [End Page 165] sources (the muzhiming 墓誌銘, for example, have not yet been systematically analyzed.)12 In other words, even though the extant sources for Anglo-Saxon England are far more limited than they are for Middle Period China, users of that database can be confident that any individual who appears in the sources will be represented in the database. A significant difference between PASE and CBDB is the representation of material culture in PASE. Gifts exchanged, belongings inherited, and donations received are all included, and can be searched under the category "property." One can find out, for example, that two pilgrims who journeyed to Rome returned to England with "silk and cloth dyed in purple." They also had volumes of monastic rules, relics of the saints and images of the virgin in their possession, suggesting by association the high status of the purple silk. The database allows one to trace these two anonymous men through the extant sources, to build a picture of all those they were in contact with in their lifetime, and to find extant editions and translations of those sources.13 Of course the information is frustrating in its tantalizing nature: if only we could know where the silk had come from and where it had been dyed, or what happened to it when it was brought to England. What matters is the level of detail possible when the extant sources are limited. The sheer volume of the Chinese sources probably makes the inclusion of such detail unrealistic on any significant scale, but as a result of the workshop, the category of "possessions" (cai chan 財產) has now been added to the existing data tables.
Professor Fong's database, a joint project of McGill University and Harvard Yenching Library, is built on a different premise. MQWW includes writings by women of the Ming and Qing dynasties (1368–1911) held in the Harvard Yenching Library, including materials in its rare books collection, and digital images of the full texts in this collection. The real strength of this material, however, lies not only in the access to textual materials the database provides, but in the research it facilitates: the "comprehensive, searchable database supports statistical research on women's social and marital status, ethnic identity, their geographical location, family and regional networks, male involvement in women's publication, and many other issues related to the social and cultural [End Page 166] history of women."14 Like CBDB, and unlike PASE, MQWW does not aim to include every known female in the extant written material—the material is simply too abundant for that—so the nature of the statistical research possible is thus very different. Prosopographical research, however, seeks to find out more about a specific population group, and in the case of MQWW that group is determined by its shared activity of writing.15
The work done at the Max Planck Institut für Wissenschaftsgeschichte under the direction of Dr. Dagmar Schaefer is useful in this context for yet a different reason. The aim of one of the projects directed by Schaefer is to facilitate the matching of data, including data extracted from CBDB, to maps powered by Google. Examples of the work carried out in Berlin include, for example, the mapping of all known kiln sites on timed map series, so that one can see how the centers of kiln technology shift over time. By combining kiln data with, for example, geological data, one can investigate the extent to which the location of centers of ceramic production was dominated by access to geological resources. As the CBDB website clarifies, extracting CBDB data and including the data as a layer on any GIS-generated map is made possible by the provision of x/y co-ordinates for all entries in CBDB. For ease of access and user-friendly interface, however, Google maps are hard to beat, and it seems likely that the Max Planck Institute will continue to provide extremely valuable and user-friendly research tools.16
The Warwick Papers
Specifically, then, what did the case studies reveal? Using prosopography as an effective tool relies on a number of important steps: to select from the complete database a group defined by a set of shared characteristics for interrogation; to formulate a set of questions that can be meaningfully asked of that group; and to integrate the quantitative information such a query yields into wider analyses. The papers presented at the workshop revealed that for a number of reasons, it was not always easy or even possible to complete these steps. Chang Woei Ong's paper focuses on the Zhe family in Northern Song, and [End Page 167] seeks to challenge the conventional wisdom that the Song dynasty fell because certain policies, enshrined at the dynasty's outset, undermined the military in the long term. His textual analysis reveals the ways in which members of the military elite effectively mobilized military strategies to strengthen their local and national standing. But Ong's paper raises important questions for the usefulness of the CBDB for research of this nature: his analysis highlights the significance of specific events, such as the granting of rights to those stationed on the border to manage border affairs, and the implications of such events over time. A biographical database such as CBDB does not include events as a matter of course, because events are difficult to integrate into a relational database.17 The Prosopography of Anglo-Saxon England project has painstakingly included each event noted in the texts, and related it to each individual involved in the event. In response to the workshop discussions, the category of "events" has been added to the online version of CBDB, so that it will be possible to link individuals in the database to the major historical events in which they played roles and to map connections between individuals established through shared participation in those events.
Related to the difficulties associated with the representation of events in the database is the spatial dimension of those events.18 Ideally, we would like to know what happened in a person's life, as well as where those events took place. If someone was granted or withdrew from a particular position, we would like to know where this took place. In principle, the database is designed to accommodate a number of different types of place associations in a person's lifetime (categorized rather imprecisely as "actual residence," "moved to," "household registration," and "basic affiliation"), although in practice, the most common place reference is "basic affiliation." Place associations of women are usually based on the place associated with their fathers, whether that is the place where their fathers served or were born. The database would ideally provide for each individual the places of his/her ancestry, birth, residence(s), career posting(s), and burial, including the dates for which those places are valid, allowing one to plot the movement of each individual through space and time. This would then facilitate the building of a picture not only of the movements of individuals, but where the paths of individuals intersect with [End Page 168] others. We can map, for example, the "basic affiliation" of the known members of Wang Anshi's social network, or the place of origin of all the prefects serving in Ji'an prefecture over time, but we probably will not be able to map exactly where each man in Wang Anshi's network ever served, or where the men who served in Ji'an lived just before they accepted the post of Ji'an prefect.
The issue is highlighted specifically in the paper by Fang Chengfeng, who uses the case of Zhou Bida's 周必大 (1126–1204) family as they migrated from the north during the 1120s to test the value of the CBDB. Fang's analysis throws light on the pattern of that migration, revealing other factors beyond the Jurchen invasion that scholarship of the "southward migration" (nan qian 南遷) thus far has taken little account of: the significance of civil service postings in determining the path of migration, the role of postings of relatives in providing a base for widows and orphans, and the role of the family graves in the process of resettlement and the claiming of a connection to new localities. Fang's precise tracing of the movements of the various members of the Zhou family relies on extracting details distributed throughout literary collections and local histories. Moreover, it relies on mapping the location of individuals through time and matching that mapping with other relatives. As Fang points out in his conclusion, for CBDB to be able to do this work for him, each event in an individual's life would need to be dated and provided with x/y co-ordinates.19 The striking difference between the CBDB's representation of Zhou Bida as a man from Jizhou and Fang's research into the complex pattern of migration that led to Zhou Bida's construction of a Jizhou identity has significant implications for what we can expect from CBDB. Such finegrained time-space detail is rarely available from standard biographies but depends instead on tracing a life history from all possible documentation. CBDB is biography based, but once all available biographies have been digested researchers may find it worth their time to add value through such detailed studies.
One of the main issues that emerged at the workshop was that of data capture. Can (and should?) this database ever provide sufficient data so as to satisfy all of our research needs? The papers that dealt with the Tang dynasty can serve as example here. Hartwell only began to add Tang data at a later stage of the project, and Tang data coverage remains slight. Anthony DeBlasi [End Page 169] set out to test the data included for holders of Tang dynasty high office by comparing the CBDB data with what is available in published Tang reference works. Finding the data available for Tang men extremely small, DeBlasi suggests a significant amount of data would have to be added to make statistical analysis meaningful.20 The limited nature of the coverage seriously hampers the scholar interested in Tang prosopography, although DeBlasi is able to pose a number of queries about the manner of entry into the bureaucracy for high officials, the kinship relations of high officials, and the background of prefects. Chang Wook Lee evaluated the role of the CBDB for a study of Hanlin academicians during the period 783 to 1082. He pursued two separate lines of enquiry: one on the political aspects of the Hanlin academicians and one on the social relationships of the Hanlin academicians. As in the case of DeBlasi's study, the thin coverage of Tang and Five Dynasties materials in general, and the current exclusion of newly excavated materials such as Tang dynasty funerary inscriptions (of which Lee claims there are over 8,000 extant) specifically, makes the use of CBDB for statistical analysis difficult. Through a close analysis of Ouyang Xiu's career, Lee demonstrates convincingly the value of including extensive detail on the various types of posts held, often concurrently, including honorary and titular posts, and the bestowal of various honors and their remunerations. Lee draws on the nearly 60 letters of appointment (zhi ci 制詞) in Ouyang Xiu's literary collection to make his case. While these two studies clearly point to problems in the data coverage for the Tang, they also highlight the potential for Tang dynasty research once the available database architecture has been populated more substantially.
The issue of data coverage also emerges in Gerritsen's attempt to evaluate the usefulness of the database for research on women in Middle Period China. Focusing on women of four different prefectures (Jizhou 吉州, Fuzhou 撫 州, Mingzhou 明州 and Wuzhou 婺州), the paper compares data on their lifespan, age at marriage, number of children, the comparative status of their fathers and husbands, and the address of their husbands and fathers. This last comparison was intended to test the hypothesis that a shift took place between Northern and Southern Song in the selection of marriage partners: from selecting partners that strengthen ties to the capital to using marriage alliances as a strategy to strengthen local ties. The design of the database in [End Page 170] principle allows for such investigations: sons-in-law are often included in an individual's network. In practice, however, one can only detect a trend over time when there is an evenness in the coverage over time. In this case, the available data is small (with samples on average containing 17 individuals) and heavily skewed towards the Southern Song, rendering any detection of a historical trend meaningless.
Naomi Standen set out to investigate the socio-political transformation of rulership and governance during the Five Dynasties (907–960). Venturing into territory that was of little interest to Hartwell, Standen found the existing data coverage thin to non-existent, and the categories that shaped the associations between individuals too static for her purposes. In response she moved to create her own spreadsheet-based database that allowed for greater flexibility and ease of input. Her work here, which potentially forms the initial stage of a much larger research project that broadly investigates the ways in which power moved from individual powerholders to institutional structures of power during this period, focused on the composition of groups associated with the court and the emperor. To compensate for the lack of data, Standen not only created new sets of data but also organized the data into new categories, classifying the types of associations she found as "givens" (including native place and kinship), "circumstances" such as a martial encounters or joint appointments, "deliberate creations" such as marital ties and adoption, and "relationships of unequal power" such as refusals and rescues. Only through such categorizations do the power dimensions of those associations become manifest. She strongly argues for greater involvement of the users of the database, through, for example, the creation of user-defined categorizations of associations, and the uploading of data sets, which the online version is intended to facilitate.
Mark Strange focuses on the networks that surrounded Sima Guang and Wang Anshi. He makes use of the database to locate significant experiences with border administration in the life experiences of the associates of the two men and of prosopographical analysis to identify the group characteristics of the networks of associates that surrounded these two powerful opponents. His analysis of the two networks confirms some of our assumptions about the existence of differences between the associates of the two men, but also challenges those assumptions where the backgrounds and experiences turn out to be largely shared between the two networks. His paper thus shows how prosopographical research can provide the impetus for new textual research. [End Page 171]
Robert Foster's paper investigates the social network of Lu Jiuyuan 陸九淵, testing the information available in CBDB by comparing it to information compiled in a number of other sources.21 His research reveals that, although many of the associations that appear in other sources are conspicuously absent in the database, several associations are only visible in the CBDB. Most importantly, these include the associations created by marital connections in the extended Lu family. The analysis of network connections, if one uses the filters of kinship, sex, and the nature of the connection, can reveal the role marital connections played in the consolidation of local ties, in this case in Jinxi 金溪 county in Fuzhou 撫州 prefecture, highlighting the potential of CBDB-based research where it is used in conjunction with other datasets.
Some of the papers, instead of investigating networks of individuals, created groups based on shared participation in socio-cultural practices. Chen Wenyi's paper, for example, investigates the exchange of "presentation prefaces" (zengxu 贈序), which she has argued elsewhere had become popular first in the Tang dynasty and again in the mid Yuan dynasty.22 Whereas her dissertation work relied on an analysis of the content of the prefaces, her approach here supplemented that data with quantitative analyses of temporal trends, local variations and generic differences. The CBDB data is limited, to be sure, yielding only one-tenth of the number of prefaces exchanged compared to a count of prefaces included in extant wenji collections, but the lack of coverage is compensated for by the fact the database facilitates relatively quick comparisons across time and space. Chen Wenyi is thus able to compare the data for the exchange of prefaces with epitaphs, or the data for Jiangxi with Zhedong, or the Southern Song data with Yuan dynasty data, or the status differences between authors and recipients of prefaces through time and space.
Inspired by the relevance of the sociology of social networks to longue durée questions in imperial Chinese history, Hilde De Weerdt set out to undertake a case study of correspondence networks (focusing on Mingzhou Prefecture, [End Page 172] now Ningbo) to gauge the utility of CBDB in mapping the frequency, distance, social make-up, and content of correspondence associations. Her inspiration here was the promising presence of a number of types of associations in the database, such as X "corresponded with" Y, or "sent a congratulatory note to" or "replied to an official letter from." Lack of detail for correspondence and postings data in CBDB, however, thwarted her attempt at mapping correspondence associations (a query into correspondence exchanged between residents of Mingzhou and beyond yielded zero results), but led her to test the coverage for all types of associations in Mingzhou and across jurisdictions in CBDB. Despite the absence of evenly-spread geographical detail on the majority of postings and a dearth of data on sub-circuit level postings, De Weerdt was able to develop hypotheses about the differential geographical distribution of Mingzhou residents' first- and second-order connections thanks to the richer coverage of the exchange of biographical writings (muzhiming) and ritual texts (jiwen). She suggested, for example, that Mingzhou residents were mostly connected with other Mingzhou residents though first-order ties (person to person), but had second-order (i.e., via another individual) of a political nature almost exclusively outside of Mingzhou.
Chen Song reflects the approach of someone intimately familiar with the database itself, as Chen Song serves as one of the CBDB project managers. His investigation of changes in the socio-political roles of elites in Sichuan relies on sophisticated interrogations of the available data and deftly demonstrates the changes over time through the projecting of this data onto historical maps. Using network visualization software, he is able to show the impact of changing personnel policies implemented from the mid-eleventh century onwards, when native incumbency began to be possible for prefecture and circuit level office. His case study of the Sichuanese elite shows the political motivations behind the social changes we usually refer to as "localization" and associate with a weakening central state from the late eleventh century onwards. His paper is a powerful demonstration of the potential of this database for testing and adjusting the hypotheses we work with in Middle Period history.
Methodological Considerations and Implications
The Warwick workshop clearly had its uses: the pilot projects and the more or less successful outcomes of the research they represent can, if nothing else, serve as guidance to others to see what might work and what does not. Clearly, [End Page 173] the database presents problems in terms of coverage, and each participant found gaps in the specific areas he or she investigated. The plan to develop a system of automated data capture under human editorial supervisions will go a long way towards greatly increasing comprehensive, rather than selective, coverage of extant sources, adding significant numbers to the extant data. Nevertheless, gaps are bound to remain, because the sources themselves were selective to start with and because only a proportion of the sources have survived. More valuable, perhaps, would be a further consideration of the questions that the database can answer despite its imperfect coverage. The methodology of prosopography is, after all, frequently used in precisely those areas where coverage is imperfect at best. The advantage of working with a limited amount of extant material, as prosopographers of the Byzantine world or the Roman empire do, is the confidence this generates amongst its users that all materials have been included, despite the fact that this material can represent only a fraction of the material produced at the time. The sheer size of the extant material for Middle Period China (not to mention the potential for growth emerging from ongoing archaeological work) means we have much more material to work with, but it remains only a sample of the total population. The contentious issue of the representativeness of the sample afforded by CBDB will likely remain for the foreseeable future, and we must remain alert to the problems this throws up. As Robinson warns, "treating prosopographies as data banks is not without its risks, especially if the answers are taken to be representative for society at large."23 Of course the biographies this database relies on come from a narrow slice of the social and political elite, as do many other prosopographical databases. Over time, all Liao, Jin, and Yuan biographies will be included, but the database can only reveal what is in the historical record; coverage of the non-Han population who rarely interacted with the compilers of biographies will always remain incomplete, as will coverage of members of the lower socio-economic strata for whom biographical records rarely were composed. The closer the questions are focused upon the members of that elite themselves, the more accurate those results are likely to be. Our challenge will be to devise queries that yield meaningful results despite the unevenness in the coverage and the lack of broad social representativeness. Ultimately, China's historical record will [End Page 174] give us one of the most detailed and continuous records of a political elite available for the last two thousand years. At the very least the existence of a deep and broad prosopographical database means that we can investigate changing patterns across tens of thousands of cases. These abstracts, and the full papers downloadable from the site, will potentially go some way towards suggesting ideas and possibilities for such investigations.
I am grateful to the British Academy (London) and the Chiang Ching-kuo Foundation (Taibei) for their generous funding of the workshop at which these papers were presented and to Warwick University for its support of the conference.
References
Footnotes
1. Hartwell's original GIS datasets are available through the CHGIS website but were not integrated into the China Historical GIS (from January 2007 in its fourth version). CHGIS datasets are freely available for downloading (http://www.fas.harvard.edu/~chgis/)
2. The project was awarded a grant from the Warwick Research and Teaching Development Fund in 2003.
3. Note that the authors of these papers used a version of the database that has now been superseded by the updated and enlarged online version, and that a number of the issues raised in these papers apply only to the MS Access version of the database, and not to the online version. At present, the online version is not yet available to the public. The MS Access version can be downloaded from the website (Google "CBDB Harvard").
4. Debra Nails, The People of Plato: A Prosopography of Plato and Other Socratics (Indianapolis, IN: Hackett Pub., 2002), xxxviii.
5. Ollie Salomies, "Names and Identities: Onomastics and Prosopography," in John P. Bodel, ed., Epigraphic Evidence: Ancient History from Inscriptions (London: Routledge, 2001), 75.
6. Lawrence Stone, "Prosopography," in Historical Studies Today, ed. Felix Gilbert and Stephen Richards Graubard (New York: W.W. Norton, 1972), 107–140.
7. Chase Robinson discusses the difference between biography and prosopography in Arabic texts produced between the eighth and sixteenth centuries, the formative and classical periods of Islam. In his particular usage, which he admits differs from the way the term is commonly understood, prosopography is distinct from biography: "Whereas biography is about exemplary or otherwise distinctive individuals, prosopography compiles and organizes those items of biographical data that mark an individual's belonging to a group" (Islamic Historiography [Cambridge, New York: Cambridge University Press, 2003], 66).
8. Billie Melman, Women's Orients: English Women and the Middle East, 1718–1918: Sexuality, Religion, and Work (Ann Arbor: University of Michigan Press, 1992).
9. http://www.pase.ac.uk/, accessed on May 27, 2008.
10. http://www.pase.ac.uk/redist/pdf/PaseInventory-0505.pdf, accessed on May 27, 2008.
11. The number continues to grow now that it is possible to add new entries online; as of spring 2008, there were 39,000 biographical entries. For the latest figure, see the CBDB website.
12. For details on the sources used, see the CBDB website.
13. As it happens, not much is known about these two anonymous men; the author of the text they appear in, Aethelwald, is also rather unknown, and the text, in manuscript until it was published in 1919 by Rudolphus Ehwald, undated. The point is that the database is designed to provide such detail wherever available.
14. "Introduction to the Online Digital Archive," http://digital.library.mcgill.ca/mingqing/english/introduction.htm#text4, accessed on May 27, 2008.
15. Further funding has recently been awarded to this project to build an interface between MQWW, CBDB, and CHGIS.
16. As of the summer of 2008, the CHGIS website also offers the possibility of downloading Keyhole Markup Language (KML) for use with GoogleEarth.
17. See also the discussion below of Naomi Standen's paper.
18. This is specifically mentioned as a problem in the papers by Hilde de Weerdt, Chen Wenyi, Chang Wook Lee, and Anne Gerritsen.
19. The updated version, in response to the workshop, now includes the category of "visited to or went to" to facilitate such research.
20. The data on Tang women is even smaller, as DeBlasi points out, with only 52 women included for the Tang dynasty.
21. These include both published secondary sources and Foster's PhD dissertation on Lu Jiuyuan. See Robert Foster, "Differentiating Rightness from Profit: The Life and Thought of Lu Jiuyuan (1139–1193)" (Harvard University, 1997).
22. In CBDB, such prefaces are referred to as "departure notes" (贈行序 or 送行序). Chen Wenyi explores this genre more fully in her PhD dissertation. See Chen Wenyi, "Networks, Communities, and Identities: On the Discursive Practices of Yuan Literati" (Harvard University, 2007).
23. Robinson, Islamic Historiography, 71.
Paper Abstracts
"Social Writings from the Song and Yuan: The Recipients of Prefaces by Jizhou and Mingzhou Writers"
(Chen Wenyi, Academia Sinica)
This paper explores one aspect of literati social connections by analyzing the practices associated with a literary genre known as the "presentation preface" (zengxu 贈序). I try to test how the CBDB can be used to study the general features of one specific practice. By treating the recipients of "presentation prefaces" as a group of actors for prosopographical analysis, this paper investigates what kinds of personal relationships and social networks were expressed in preface writing: who the recipients were, what relations they had [End Page 175] with the author, and what geographical range they occupied in opposition to the author.
Elsewhere, through a detailed analysis of the content of numerous prefaces dating from the Yuan dynasty, I have established that there was a general mode of practice surrounding the "presentation preface." In this paper, I expand on that earlier work by using CBDB to study concrete patterns of this practice, such as temporal trends, local variations and generic differences. I first draw data from CBDB to provide statistics for analysis. I then supply a complete list of prefaces written by Jizhou and Mingzhou authors during the Southern Song and Yuan (falling roughly between 1100 and 1350). (Note that it is almost impossible to collect all the recipients from a specific area without the help of the database.) In combining these two kinds of information, I have two objectives. First, I aim to get a better sense of the distance between the unfinished database and a completed data set. Second, I hope the experiment based on the numbers from the two more detailed case studies can suggest the potential results we can expect to achieve when the database is complete.
This paper concludes that CBDB is more valuable in generating new issues for study than in simply providing answers to existing questions. For my research on the practices associated with prefaces, the simple fact that CBDB can, when it is completed, provide information about the recipients and make it possible to analyze them as a group is of great help. Moreover, the CBDB should be able to provide valuable information about obscure figures, such as different types of specialists (fortunetellers, physicians, craftsmen, etc.), and thus facilitate the study of these hitherto neglected segments of society. It is also very useful in testing various temporal trends and geographical distribution, two factors that should serve to modify/qualify generally accepted conclusions and stimulate new directions for further research.
I make two suggestions regarding the future development of CBDB. First, concerning the data, I recommend that we include ALL prefaces in the database because prefaces are important indicators of social relationships. Second, for the queries, I assume there will be a way to collect information by association (as the "geo-reference" part of the query in the current CBDWin). It would be useful to be able to establish correlations among various types of association, especially since this database consists of so many different types of association. Finally, it would be very helpful if we could also analyze the networks with criteria of certain types of association. [End Page 176]
"Military Commissioners and Grand Counselors: Testing Tang Era Data Capture in the China Biographical Database"
(Anthony DeBlasi, University at Albany)
This essay assesses the utility of the CBDB in its current state for Tang-era prosopographical research by examining its coverage of three concrete populations. On this basis it offers suggestions for bringing the database closer to its potential. By comparing the information already entered in the database for Tang individuals with that readily available in published biographical indices and compilations, the paper identifies some patterns in the CBDB data. This data comparison also makes clear that the CBDB's useful database architecture remains under-populated and relatively data-thin, thus limiting its present value for doing Tang-era prosopography.
The data tested concern the occupants of different bureaucratic posts, a common method for prosopographical research. Compiling diachronic lists of government officials holding three different posts and then comparing these lists and the data in the database associated with the identified officials provides a solid basis for understanding the current limitations of the CBDB. The three lists are: (1) Grand Counselors that served during the Tianbao 天寶 (742–755) and the Yuanhe 元和 (806–820) reign periods; (2) Military Commissioners (jiedushi 節度使) of the Xuanwu 宣武 command; and (3) Prefects (cishi 刺史) of Lianzhou 連州. These offices were chosen first because they represent different levels of the bureaucratic hierarchy. Second, they are all represented in previously published compilations. The list of Grand Counselors was drawn from the tables in the Xin Tang shu 新唐書. The Military Commissioners were taken from Wu Tingxie's 吳廷燮 Tang fangzhen nianbiao 唐方鎮年表. The Lianzhou prefects were identified using Yu Xianhao's 郁賢皓 Tang cishi kao quanbian 唐刺史考全編.
The above comparisons reveal that relatively few of the individuals appear in the database. This fits the overall coverage of the Tang in the CBDB. Whereas the standard Tang biographical index, the Tang Wudai renwu zhuanji ziliao zonghe suoyin 唐五代人物傳記資料綜合索引, contains 30,000 individuals, only 2285 of 34,212 records currently in the CBDB are Tang records. For the three populations examined, the extent of the coverage declines with the prominence of the post. 63% of the Grand Counselor sample appears in the CBDB compared to the standard index. 28% of the [End Page 177] Xuanwu Military Commissioners are included in the CBDB data, and only 11% of the reconstructed list of Bianzhou prefects is in the database. Overall, the combined adjusted number of individuals in our three samples is 117. Of those, the CBDB has entries for 35, or 30%. The Tang suoyin has entries for 112, or 96%. Ultimately, the examination of these three small sample populations suggests that the Tang portion of the CBDB is best at capturing high profile, capital positions. As one moves into lower profile offices, the coverage drops rather dramatically.
The utility of the database depends also on the amount and quality of the information that populates the data fields associated with each individual. The architecture of the database promises much research potential because of its well-developed field structure relating to career history, entry into the bureaucracy, personal associations, and kinship relationships. At the moment, however, the amount of information provided for the three test populations used in this essay is too limited to provide any useful statistical entry into Tang history. There are nonetheless some patterns that emerge in what has already been entered in these fields. The data for entry into the bureaucracy suggests a bias toward entry via examination for Tang individuals. With regard to the kinship data, the more prominent the post, the more data that appears for those holding it. For example, the CBDB contains an average of 2.1 kin per record, but the Grand Counselors have an average of 2.4 kin per record. Currently, the association data is too thin and unreliable for research purposes. Even very well known historical relationships are missing in the data.
The analysis of the sample populations suggests that the first step for improving the Tang CBDB data is to mine the Tang Wudai renwu zhuanji ziliao zonghe suoyin to increase the number of Tang figures included. Utilizing the almost thirty thousand entries in the index to augment the slightly more than two thousand Tang entries in the CBDB would require a serious investment of time and resources, but it is a well-defined task that would require only widely available sources. It would further allow a geographically dispersed cadre of individuals to add the data to the CBDB. More generally, the value of utilizing pre-existing compilations to more quickly populate the database is clear from the samples examined. Once that is accomplished, attention could be turned to the expanding corpus of published Tang era inscriptions. [End Page 178]
"The Migration of Zhou Bida's Family during the 1120s and 1150s: An Evaluation of the CBDB"
(Fang Chengfeng, PhD candidate, History, Peking University)
Historians frequently have to face the reality that people are constantly on the move. Theoretically, the CBDB can easily trace the movement of a given person by looking at the entries of addresses in the database. This paper focuses on the issue of addresses by tracing the migration of Zhou Bida 周必大's family during the 1120s and 1150s. Zhou Bida's basic affiliation was Luling 廬陵, Jizhou 吉州, Jiangxi Circuit 江西路, but before the collapse of the Northern Song, his family resided in Guancheng 管城, Zhengzhou 鄭州 in the north. Thus the Guancheng-Luling migration seems to be an example of the southward migration 南遷 studied by historians of the Song.
The starting point of the Zhou family's migration, however, was not Guancheng. Zhou Shen 周詵, Bida's grandfather, was the vice prefect 通判 of Jizhou when the war among the Song, the Liao, and the Jurchens broke out in the 1120s. His son Zhou Lijian (1) 周利建, Bida's father, was an official in Kaifeng 開封. He and his wife then moved to Pingjiangfu 平江府 in 1126, before the fall of Kaifeng. Zhou Lijian (2) 周利見, Shen's eldest son, was in Xiazhou 峽州 and then the Sichuan Basin when the Jurchens invaded. In sum, family members were already in the south before the collapse of the Northern Song. Therefore, instead of asking about the process of southward migration of the Zhou family, we might instead ask: how did the dispersed family members reunite and become a new family in the Southern Song?
During the process of reunion, Zhou Shen, Zhou Lijian (1) and other adult members successively died. Zhou widows and orphans always followed in the footsteps of their relatives who held offices, and those who were officials also constantly moved following changes in their appointments. From the 1120s to the 1150s, Zhou Bida and the other family members seldom resided in Luling. Although the Zhous had seldom resided in Luling, in the autumn of 1150 Zhou Bida passed the prefectural examination of Jizhou, which meant he was regarded as a Luling person by the government. How did this happen?
The family graveyard played a significant role in shaping the identity of Zhou Bida and his family. Evidence from another two families demonstrates that the establishment of the family graveyard could be the beginning of resettlement. It was Zhou Lijian (2) who built the family graveyard and maintained the concentration of the tombs of the family members (zuzang [End Page 179] 族葬). According to Patricia Ebrey, the proximity and the order of tombs are vital for the grave rites. Since none of the Zhou family members had permanently resided in Luling before the 1150s on the one hand, and the family was regarded as Luling-affiliated on the other, the family graveyard became the most important means for the Zhous to show their new affiliation. Luling in this way was shown to be the new home of the Zhou family, whose ritual identity was confirmed by the dead, not by the living, who constantly moved across the empire.
My research drew mainly on the writings of Zhou Bida rather than on the CBDB. The comparison between my resources and the data in the CBDB reveals something significant. First, how to integrate addresses from different fields? In the CBDB, addresses are in several fields, namely the field of Addresses, the field of Office, and the field of Association. The reporting system needs to find a way to gather all sorts of addresses at once. Furthermore, since the addresses come from different fields, it will be problematic to organize the sequence of the addresses into one report.
The CBDB is good at analyzing spatial distribution when it works together with CHGIS. Determining spatial movement, however, presents a challenge that goes beyond those involved in figuring out distribution. It requires the knowledge of sequence in which respective movements occurred. The CBDB deals well with the data for which accurate times are known, but it is not flexible enough to deal with the data dated ambiguously. And even when exact dates are not known, such ambiguous data can be very useful in reconstructing the traces of the biographees.
Second, the CBDB needs to find a way to figure out how various people's movements are linked. My paper shows that people move with others. For instance, Zhou Bida, his siblings, his mother, his grandmother, his maternal uncle and his maternal grandmother were all in Hengzhou around 1132. The CBDB has to be able to 1) trace person A's mobility by retrieving the various addresses related to him or her; 2) search among people who have a relationship (kinship or association) with person A and are related to the addresses retrieved.
Finally, the CBDB's data is not adequately detailed. There are two reasons for this. First, current work is based on Wang Deyi's index, which contains very limited information. Second, though it would be better if we input into the CBDB original biographical texts, my paper shows that some crucial [End Page 180] evidence cannot be found in biographies, but in prefaces and diaries. This indicates that only when different kinds of materials are well digested and organized, can the information be maximized for the CBDB. Actually, the most valuable and well-organized materials lie in the large number of existing second-hand works, such as genealogies, local histories, literary history and political history, including institutional history. Therefore, the most reliable contributors of data to the CBDB should be the scholars who use it.
"Comparing Sources for Lu Jiuyuan's Social Network"
(Robert W. Foster, Berea College)
While researching material on Lu Jiuyuan for my dissertation, I compiled a non-systematic database of individuals encountered in Lu Jiuyuan's collected works, local gazetteers, biographical compendia, the Song-Yuan xue'an, and other sources. Most of the 160+ individuals were disciples, family members, those mentioned in the yulu and nianpu of Lu's collected works, people for whom Lu Jiuyuan had written funerary inscriptions, or with whom he had corresponded. The goal of the current paper is to compare family connections and social network information compiled from four sources: the CDBD, the dissertation database, Robert P. Hymes's detailed study of Fuzhou, Jiangxi, found in Statesmen and Gentlemen: the Elite of Fu-chou, Chiang-hsi, in Northern and Southern Sung, and from the list of disciples compiled by Xu Jifang based on the Song-Yuan xue'an 徐紀芳 (Lu Xiangshan dizi yanjiu [Taibei: Wenjin chubanshe, 1990]). With rich sources external to the CBDB we can see how useful the database can be as a single source of information or used in conjunction with other materials. It is clear that each source has its strengths and weaknesses, but none provides as clear a picture as information garnered from all of them together.
The integration of the Lu family with local society becomes clear when we look at the descendents of Lu Jiuyuan's grandfather, Lu Jian, compiled from information generated by querying the CBDB. Though there are some quirks and the database returns are constantly shifting due to on-going data entry, the key benefit of the CBDB is that it broadens the scope of relationships to include marriage in ways that my gleanings from textual sources did not. By further generating data for marital connections for Lu Jiuyuan's family (particularly his five brothers) we find deeper connections with the local elites of Jiangxi that might further explain the position of the family in local projects [End Page 181] such as granary building and the militia. Such findings should encourage us to challenge the view put forward by Robert Hymes that the Lus were not actively involved in local society to the same extent as other elite families.
One particularly interesting link is with the Wangs of Linchuan, connected to Lu Jiuyuan through the marriage of his brother Jiuling to a great-great-granddaughter of Wang Anshi. When asked to write a record for the rebuilt school in Wang's home county of Linchuan, Lu Jiuyuan produced a respectful, though critical, account of Wang; and later, when criticized for writing the piece, Lu retorted that a truer assessment of Wang did not exist. Earlier, I had credited Lu Jiuyuan's positive assessment to his agreement with Wang's stress on returning to the "rightness" inherent in the models of ancient sage-kings. In part, this is true, but the CBDB also demonstrates strong family connections to the Wangs of Linchuan and to other Linchuan shidafu families. Linchuan was also the residence of eleven of the 82 men linked to Lu Jiuyuan in the Song-Yuan xue'an, and for ten of the 117 people linked directly to Lu Jiuyuan in the CBDB.
Turning to Lu Jiuyuan's larger social network, we can compare the listing of 82 men linked to Lu Jiuyuan in the Song-Yuan xue'an, as compiled by Xu Jifang, with the 117 linked directly to him in the CBDB. We find that 46 of the men from the Song-Yuan xue'an are not in the CBDB. One of the more interesting people to not be in the CBDB is Fu Mengquan, leader of the Huaitang group comprising 65 out of the 82 men in the Song-Yuan xue'an.
If we turn to the people linked to Lu Jiuyuan in the CBDB who are not in Xu Jifang's listing, we find 23 men. But, if we winnow away those who do not have established kinship relations with Lu Jiuyuan, we find eleven non-kin men mentioned as close associates in the CBDB, who seem to be intellectually connected to Lu, but who are not mentioned in the Song-Yuan xue'an. So the CBDB is providing new information on connections beyond family that are not clear in the Song-Yuan xue'an. However, the CBDB does not have information on an even greater number of men linked to Lu Jiuyuan intellectually.
The interaction between Lu Jiuyuan and Zhu Xi was a key moment in the intellectual history of Neo-Confucianism. The two exchanged letters and, at the behest of Lü Zuqian, met for the "Goose Lake Debate" to discuss their intellectual differences. According to Xu Jifang, of the 82 men in the Song-Yuan xue'an group, 22 were also linked to Zhu Xi's school. Combining [End Page 182] the data gathered from searches for those directly connected to both Lu Jiuyuan and Zhu Xi produced 353 records. Yet of that group, only thirteen were reduplicated. Of Lu Jiuyuan's brothers, only Jiuling, whom Zhu met at Goose Lake, is found in Zhu Xi's direct social network, while Lu Jiushao, who was also in correspondence with Zhu Xi and seems to have opened the intellectual debate that led Jiuling and Jiuyuan to Goose Lake, is not. It would be useful to develop a means by which people can reliably contribute to the information in the database, since at this stage I was unable to enter links such as that between Lu Jiushao and Zhu Xi.
Clearly the CBDB has some gaps, but it does provide a great deal of information and opens further avenues of exploration. Since the database is not fixed, but is constantly being amended, these gaps and issues are continually closed and fixed. However, it is only fair to point out the difficulty of using the database if one is not well-schooled in Microsoft Access. It is clear that this paper barely scratches the surface of the compiled information, but for the CBDB to become a truly useful tool for research there are at least two key issues to face: the creation of a user's manual, or a more user-friendly interface, on the one hand, and an efficient and trustworthy means of opening the database to further data entry. It might help to determine how scholars and students of Chinese history can individually contribute to this resource.
"Use of the Hartwell Database to Study Tang-Song Hanlin Academicians, 783–1082"
(Chang Wook Lee, PhD candidate, Binghamton University)
During the Tang-Song transformation, one of the most significant political and social changes was the establishment of new imperial secretaries and the transition in the power of the ruling class from its traditional base in the old aristocratic families to the emerging hold of the newly rising elite class. In 736, the Tang government created the new imperial secretaries called Hanlin Academicians 翰林學士, modeled upon the pre-existing Chief Secretaries 中書舍人 of the Secretariat and the various other types of academicians. This event reflected changes in elite society and it significantly influenced the development of elites in subsequent imperial history.
This paper discusses political and social issues related to the Hanlin Academicians during the Tang-Song transition by using the Hartwell Database. Because this paper concentrates on the political aspects of the Hanlin Academicians during the Tang-Song, its main focus is on the various different [End Page 183] official posts in the Chinese government, such as special commissions 差遣使職, titular offices 階官 and prestige offices 散官. From the seventh century on, both the Tang and Song governments made frequent use of special commissions in order to create new official posts and to supplement the weakness of the pre-existing governmental structure. But they could not substitute the existing governmental structure with the newly created special commissions since no other special commissions could completely replace the centuriesold Three Departments 三省 and Six Ministries 六部 in the government. Although the Yuanfeng reforms 元豊官制 (1085) ended this complicated governmental organization by re-establishing the Three Departments and Six Ministries, the use of special commissions continued into the Ming-Qing period.
This paper examines the various different governmental official posts based on information concerning the Hanlin Academicians in the Hartwell Database. Second, this paper also uses the Hartwell Database to investigate the social relationships of Tang and Song Hanlin Academicians, especially those formed through marriage, sponsorship and the recommendation 薦擧, but also those reflected in ancestral place 祖籍, current residence 籍貫 and funeral locales 葬地.
"Mapping Communication from Mingzhou: Networks of Correspondence"
(Hilde De Weerdt, University of Oxford)
Responding to the call of John Unsworth, who wrote convincingly about the need for reports on failed research among humanities researchers, my report presents a record of the questions and methods that guided my first foray into the CBDB, a rationale for research on correspondence networks, a review of the rather unsuccessful results qua content, an examination of the reasons behind the mixed results, and, finally, reflections on the utility and extensibility of CBDB in pursuing research on social networks.
The scope of its biographical data and the fine granularity of the stored information make possible multi-factor analyses of social and cultural developments on both large and small spatial and temporal scales. The typology of associations drew my attention in particular. The wide range of social, political, and textual relationships the database proposes to track has the potential of cross-pollinating research on imperial Chinese society, the sociology of social networks, and relational models of social history. The first section of [End Page 184] my report therefore briefly reviews the history of social network analysis and the interaction between network models and social history. Recent work in the "new science of networks" and the application of relational analysis in the scholarship of Charles Tilly present productive challenges for historians working on the social, political and intellectual lives of imperial Chinese elites. Can major transitions in imperial history such as the reorientation towards local society in the twelfth century, the spread of Neo-Confucianism, and the economic and cultural rise of the south be explained by or correlated to changes in social network ties at regional and/or empire-wide scales? Can outstanding questions such as the level of political organization and the nature of political participation be answered by systematically investigating and mapping associations among elites of various kinds? My interests in intellectual history and in the dissemination of political information led me to a pilot project designed to trigger a larger investigation into the history of correspondence networks.
Why investigate networks of letter-writing? Mapping the frequency and distance of correspondence across time will result in findings regarding the connectedness and centrality of particular places, individuals or groups of individuals. Taking account of the status (degree-holder, non-degree holder, bureaucratic ranking and actual post) or kinship relations of individuals in plotting letter writing may result in new findings regarding patterns of social and political interaction among (groups of) elites. When letters are differentiated by genre and content, conclusions can be drawn regarding what types of information spread at what times, to what extent, among whom, and in which directions. The nature of the plotted networks may also help explain the reach and speed of particular types of information. Results obtained regarding changes in frequency, distance, genre and content can also be correlated to known events and developments in the capital or in the provinces. Networks of correspondence could also be compared to networks resulting from other types of associations, especially those involving the exchange of writing, in order to determine the weakness or strength of ties and the role of different types of associations in the formation of social boundaries.
What strategies can we use to tackle some of these questions with CBDB? I narrowed the parameters of my inquiry down to mapping the exchange of correspondence in Song dynasty Mingzhou (Ningbo). As a first step towards constructing a correspondence network centered on Mingzhou during Song [End Page 185] times I set out to compile a table of individuals who were resident in Mingzhou at a given point in time either as natives of the area or as officials posted there. Step two was to query the associations by correspondence of the individuals listed as working or living there. The report's tables and charts evaluate the results of these queries and test them against a broader sample of postings and association data. The tests highlight the following conclusions: 1) the lack of specificity for the geographical location for the majority of postings data; 2) the unevenness in the quantity and quality of the biographical data; 3) the virtual absence of association by correspondence data; 4) the unevenness of the quantity and quality of association data, with some types being tracked in relatively large numbers (e.g., the exchange of biographical writing) and others in insignificant proportion.
As a first step towards mapping networks that show the extent of direct as well as indirect connections, the report looks not only at relationships of the first order (those linking Mingzhou individuals to others directly) but also at second-order relationships (those removed at a distance of two edges). It compares the distribution of association types by degree of separation, by place and by time. The results produce creative hypotheses about the connectedness among prefectures, their centrality, differences in the distribution of social ties among counties, intra-county fragmentation, and the historical trajectories of the frequencies of different types of association.
My report concludes with some suggestions and desiderata whose implementation may contribute to the fulfillment of CBDB's promise as the basic infrastructure for future network research on imperial Chinese social life. Those concern preferred strategies for the acquisition of data, their modification, and modes of access.
"Factionalism and the formation of eleventh-century military policy"
(Mark Strange, D.Phil candidate, University of Oxford)
The eleventh-century statesmen Sima Guang and Wang Anshi defined their political relationship in terms of factional opposition. That image has stuck for much of the past thousand years: clear-cut groups of conservatives and reformers have been fashioned around them. Among the central themes of this opposition has been their response to a foreign threat to Song's imperial integrity and, as a corollary, their formation of military policy. Sima Guang [End Page 186] and his conservative associates advocated restraint in the use of force; Wang Anshi's network cluster tended towards belligerence—that is the picture that has taken root.
Assumptions of divergence between the two men provide the starting point for this paper. The rhetoric of political opposition, especially in the context of military policy formation, demands treatment in its own right. It goes beyond our concerns here. Instead the present focus is on the biographical data of Sima Guang and Wang Anshi, and the lived experiences and group identities of their associates. Three tasks present themselves: to identify differences in the military experiences of these two individuals and their associates that might account for conflicts in policy; to examine the value of characterizing Sima Guang and Wang Anshi as the driving forces of two opposing political factions; to propose ways in which prosopographical research might be used as a foundation on which to construct further study of military policy. A sampling of biographical data, taken from the CBDB, supplies the context for analysis.
All this comes with an early caveat. The CBDB is necessarily schematic: the forms of tables and lists through which it presents its data show that most readily. They simplify individual experience and distort complex social and political realities. They also reflect current organizing principles as much as any contrasts or similarities in eleventh-century society. And for the purposes of the present paper they implicitly perpetuate rather than question assumptions of a factional opposition between Sima Guang and Wang Anshi, and their two network clusters. But to conduct this sort of large-scale prosopographical research and to gain insights into the formation of eleventh-century political identities, there is little choice but to use the data as it stands, remaining alert at all times to the values inherent in its graphic form.
Certainly differences do emerge between Sima Guang and Wang Anshi, as well as between the network clusters that developed around them. These have been well documented; a handful of examples makes the point. The majority of Sima Guang's associates came from regions along the Yellow River, like Sima Guang himself, or from the area around Chengdu; most members of Wang Anshi's network cluster came from the mid-Yangzi valley and its southern tributaries, as well as the Yangzi delta. The two men's family backgrounds differed. Sima Guang came from a high-ranking official family and gained his first post through the yin privilege system; Wang Anshi [End Page 187] came from a family of local officials and was not eligible for such hereditary benefits. With a close focus on military experience, a larger proportion of Wang Anshi's network cluster than of Sima Guang's associates came from regions under high-level military administration. Yet it was Sima Guang who served as an official in frontier regions and in posts with military jurisdiction; Wang Anshi, by contrast, tended to take local civil posts in the circuits of the mid- and lower-Yangzi valley, far from the military tensions of the frontier.
But if we sharpen our focus on military experience, the same data brings out convergences. A north/south divide in native place yields little as an organizing principle here. Frontier regions—the most likely settings for military conflict—are more robust as an organizing principle for comparison. And in this the two network clusters are strikingly similar: each comprised low proportions of individuals from frontier regions; neither network's members had experienced much sense of military urgency in their native regions. This extends to their official careers. The proportion of individuals who held posts that might have afforded first-hand experience of the practical implications of court-directed military policy—military administrative posts in frontier regions—is low for both networks.
The educational backgrounds of members of the two network clusters are similar. Both contain equally high numbers of jinshi degree holders and yin privilege recipients, despite broad contrasts at the time between the regions from which they came. Of greater relevance here, though, is the fact that not a single member of either network entered officialdom through military channels. Both attracted individuals who started their official careers with a civilian focus.
In the types of association that bound their members together, Sima Guang and Wang Anshi's network clusters show parallels. Most associations formed around literary exchanges and both network clusters located themselves exclusively in the context of civil governance. Five types of military association appear in the CBDB; not one features here.
The clearest indication of convergence is the participation of individuals in both network clusters. And more than that: some individuals even appear able to move back and forth between the two clusters, changing their allegiances with the pull of personal acquaintances and external political circumstances. Here, more than anywhere, is a sense of the fluidity of eleventh-century political [End Page 188] relationships and a warning against the schematic discourse of factional opposition.
That is not to reject outright the idea that there were two coherent and integral network clusters that surrounded Sima Guang and Wang Anshi; an above-average density of associations linked their associates. But such cohesion does not provide evidence of factionalism. Unexpectedly, given the broad sweep of its vision, prosopographical analysis throws up exceptions that destabilize the general image of opposition. Despite the density of the bonds of the two men's network clusters, what stands out is how much convergence there is between their collective identities, how much divides between them blur, and how uneasily the discourse of factionalism sits with the bio-data of Sima Guang and Wang Anshi, and their associates. Long-held binary oppositions—proposed even by Sima Guang and Wang Anshi themselves—are inadequate as an analytical assumption for the study of complex political relations in the eleventh century.
This paper does not propose convergence in place of existing representations of divergence. Quite the opposite—it aims to reject all such generalized models and to acknowledge the contingency and malleability of political identities at this time. The discourse of factionalism suggests homogeneity and permanence, and factional oppositions like the one commonly attributed to Sima Guang and Wang Anshi appear clear-cut as a result. Under close scrutiny, though, external political boundaries blur and internal associations become loose and unstable. The CBDB shows clearly that the structures of eleventh-century political and social groupings were shaped by the particular circumstances that attended their formation. And that is where its value lies. In affording opportunities to look at the broad canvas of eleventh-century political activity, it paradoxically nudges us towards a consideration of close, individual details. We gain new impetus to sharpen our understanding of policy formation in the eleventh century and to give the hazy people-like outlines that emerge from the CBDB's data some of the texture and individuality that cannot be gained from the sweeping categories of prosopographical research. From these crude but suggestive outlines, we therefore turn to the textual record for an insight into the inner machinery of eleventh-century political relations rather than just their external forms, as we attempt to replace schematic simplicity with a sense of the complexity that underlay eleventh-century political debates. [End Page 189]
"The Zhe (She) Military Family in the Northern Song"
(Chang Woei Ong, The National University of Singapore)
The set phrase "Cherishing the civil, ignoring the military" (zhongwen qingwu 重文輕武) has been used by scholars and laymen alike to explain why the Song was not the powerful Han or Tang. According to this conventional wisdom, the Song founders, after witnessing how the loss of control over the military had cost the rulers of the Five Dynasties their empires, became wary of the possibility of themselves falling prey to the recurring problem. They quickly took measures to prevent regional separatism from becoming a real threat. As a consequence, the Song army was weakened to the extent it could not defend against foreign invasions. It is therefore not surprising that the Song would eventually collapse, first under the Jurchen invasion in 1126 and later under the Mongol invasion in 1279.
Despite several attempts over the past few decades to rectify this perception, this view has remained prevalent. But it is essentially teleological, as it explains historical phenomena (e.g., the organization of Song military institutions) by looking at how they fit with the final outcome (the fall of the Song). In doing so, time and space are suppressed to justify the claim that the tragic ending to the dynasty was predictable by looking at the initial Song policies towards the military; regional variations and temporal changes are neglected in this process.
The purpose of this paper is thus to restore the temporal and spatial dimensions of the military aspect of the Song. I examine how historical actors made choices under specific circumstances and what were the factors affecting their choices and how did they change over time. In particular, I trace the historical experiences of a Zhe (or She) 折 family, stationed at the so-called hewai (河外, lit. beyond the Yellow River) region, whose members, making use of the unique conditions there, excelled in military services generation over generation.
This is therefore a study of social mobility and elite strategy, which is the paramount issue in the study of social history of the Song and later dynasties. But such studies have been predominantly about the literati, or shi 士, class. In comparison, the military elite attract much less attention from historians. The limitation of sources is one possible reason, as unlike the shi, the military elite seldom (but not never, as we will see later) produced literary works. Also, the view that stressed the diminishing importance of the military sector during [End Page 190] the course of the Song dynasty discussed above might have contributed to this bias. To correct the imbalance, it is therefore necessary to explore how military resources were being exploited by ambitious men to achieve local and national prominence.
The paper is divided into three parts. It first describes the peculiar situation of Hewai from the tenth to the early twelfth centuries and documents the rise of Xi Xia and its impact on the region. The second part surveys the Zhes' relationship with the court and the findings indicate that the Song court was more often than not willing to put the task of defending the borders in the hands of the Zhes, who had managed to secure hereditary rights to the post of prefect of Fuzhou in Hewai since the beginning of the dynasty. The third part examines the various strategies that the Zhes adopted to stay afloat in an increasingly Song world. They began by marrying local strongmen, then by marrying prominent military families of national importance. Over the course of history, they also adopted certain practices of the literati class and intermarried with established literati families. But a notable point is that, despite the Zhes' effort to explore various means to stay competitive, military endeavors remained the most vital route for their success.
A detailed study of the Zhe family, I believe, should inspire us to rethink some widely accepted perceptions of the Song dynasty. First of all, the Zhe story did not fit the conventional Song story of "Cherishing the civil, ignoring the military." On the contrary, what the Zhe case exhibits is the court's overwhelming concern for border defense. Also, the Zhes' pattern of rising to prominence and sustaining their success does not resemble any group of Northern Song elite, both civil and military, that has been substantially discussed in secondary scholarship. In many ways, the Zhe are unique in a Northern Song setting.
On the other hand, the Zhe story reflects a typical Song problem: how concern over national security and border defense could be incorporated into the greater goal of building an imperial state governed by a bureaucracy that was based on a civil order. For its part, the court had to strike a balance between maintaining a workable defense system and preventing the growth of excessive military power. The relationship between the court and the Zhe, and the willingness of the Zhe in helping the court to contain an aggressive Xi Xia regime for more than a century attests to the Song ability to achieve such a goal. [End Page 191]
Research for this paper was done entirely through "conventional" means of flipping through historical materials in their printed forms. The Hartwell database, at its present stage, is not very useful for conducting a study of this nature. It draws information mainly from standard biographical materials, which constitute only a very small part of the rich materials pertaining to the Zhe family. There are tremendous amounts of information about this family in Xu Zizhi tongjian changbian and Song huiyao that are not extracted and included in the database. Furthermore, the current database lacks a function for the users to search "events" (e.g., the Xi Xia "rebellion" of 1038), which makes it next to impossible to identify major figures related to major events. But the potential of the database is huge. If fully developed, it will certainly change the way we study Chinese history. I think the most important development is that the database will make apparent some connections between people and events that we tend to neglect if we simply browse through the printed materials.
"Networks and institutions in the CBDB: Adding the non-kin network of Liu Zhiyuan of the Five Dynasties"
(Naomi Standen, Newcastle University)
Working with the CBDB has helped me to identify just what data I need for my project, to develop methods for collecting that data, and to provide a framework within which to think about methods for ensuring systematically robust analysis of the data. My interest is in the socio-political transformations of rulership and governance in the Five Dynasties (907–960). The warlord regimes of this period were in the process of evolving from the personalised politics of leaders and retinues of the late Tang (characterised by charismatic adult rule) into the institutional politics of memorialising officials of the Song (characterised by the ability to sustain rule by a minor), but it remains unclear just how and when these transformations happened. We should first distinguish ruling authority (who was in charge) from governance (how things got done). Then the question is: what was the (certainly uneven) trajectory by which political power shifted from a ruling group comprising the emperor's direct personal adherents to become incorporated instead into established institutions of government under whoever happened to be staffing them at the time. At what point did handpicked individuals stop being the locus of a bureau's power and the institution become itself the source of authority [End Page 192] for those placed in charge of it? To trace the change we need to be able to compare what kinds of relationships bound ruling groups together (or not) in both the presence and the absence of effective and regularised institutions. Five Dynasties leaders often adopted their followers as sons or married them to daughters, but there were many more ties than this. What was the full range, how were they forged and broken, which were more important, when and why? And of the many factors binding people to leaders or each other, how important was cultural background? Chinese, Shatuo and many others regularly served together, married each other and pledged allegiance to each other. Was such behaviour statistically unusual? Did culture modify the influence of other factors upon choices made? Did any of this change over time? Rather than assuming the primacy of culture, we must examine cultural factors within a framework in which other possibilities can be given equal consideration.
To trace the vast proliferation of connections between individuals a database is a essential. In order to trace the relative significance of personal bonds (centred on the emperor) and institutional/structural bonds (where the emperor is one actor among many) it must be possible to examine the same dataset from both egocentric and whole-network perspectives at any given moment. The annals (which are the emperor's biography) of the Later Han (947–950) founder Liu Zhiyuan provide a suitably limited body of material for this preliminary study; the biographies of people with links to him are the next phase. None of the relevant data is in the existing CBDB and entering it is a major task. Just under two juan of annals produced over 430 relationships, of which over 300 are associations with Liu Zhiyuan.
It quickly became apparent that it would be more effective to create spreadsheets for the new data, which can ultimately be uploaded directly into the CDBD without the need for re-keying everything. With a few exceptions (notably natural phenomena and the movements of the emperor) almost all of the information in both annals and biographies can be expressed as an association between two people. Triadic relationships where X sends Y to do something to Z must be expressed as a set of three pairs to capture all the associations.
The two basic types of data around which my analysis will need to revolve are these associations between people, and the official posts held by individuals at any given moment. However, the almost entirely non-literary types [End Page 193] in whom I'm interested are not the focus of the existing CBDB, so I found major omissions in its lists of associations and official posts. Furthermore, the relationships that individuals have with their emperor (the ones I'm interested in) are very different from those they tend to have with each other (the ones collected by Hartwell). Hence over a hundred new relationships needed to be added to the CBDB list of ASSOC_CODES. There is also a need to be able to date associations much more precisely than a year or reign era. The CBDB records named individuals, but sometimes associations important for my purposes involve unnamed people or groups like military officers or 'the court', and a method is needed for recording these, not least lest they subsequently become nameable through further work. Data on official posts could particularly lend itself to automated collection, were that to be possible.
Where the CBDB is rather unwieldy to work with, spreadsheets are flexible. Being able to experiment freely with new fields and types of data allowed me to establish more fully and precisely what kinds of information are required to answer my research questions. Better still, I found that simply tabulating data into a spreadsheet revealed that associations fell naturally into different types. I made a preliminary identification of four broad categories on the basis of how links were formed and what the power relationships were in that relationship (a very preliminary and tentative listing). The method of turning the materials into data has thus contributed to my analysis, and so will help to shape my further research:
• native-place
• kinship
• cultural group
Givens:
• martial and other encounters—fighting alongside or under command of, etc.—or possibly against; prior service with or under authority of or in authority over; visits
• joint service, incl. discussions with
• shared experiences incl. joint appointments to same post or bureau
• relationship with relatives, e.g. posthumous titles for fathers, etc. [End Page 194]
Circumstances:
• adoption (none in this material)
• repeated/multiple post-giving (not just single posts or occasions, though confirmation and retirement posts may be included)
• marriages
• urging leader to power (taking throne, seizing cities, etc)
Deliberate creations:
• submitted or gave allegiance to (in battle, during siege, etc)
• provoked response in (anger, admiration, advice taken or rejected)
• sent or received envoy/ message
• refusals—of posts, acceptance of mandate, etc
• given task—signs of especial trust; or just given orders?
• transfer of position
• rescues etc
Relationships premised on unequal power (potential relationship-breakers?):
I can now consider the significance of individual and perhaps collective agency in the formation of groups by categorising together links such as marriage ties and, inter alia, instances where one individual is given a disproportionate number of posts to be held simultaneously. One might not naturally place these together, and it is difficult to imagine how such a categorisation might have been arrived at by more traditional methods of handling the source material. This promises to be a powerful way of approaching the same dataset, and was facilitated by the ability to group associations according to project-specific criteria.
Suggestions towards methods and guidelines for researchers adding new data to CBDB
• Those adding new data should use spreadsheets, and may suggest new fields for possible inclusion in CBDB.
• New data should be added in cohesive blocks for thoroughness. For example, full coverage of a specified set of annals, or a set of biographies chosen on clear criteria. New inscriptions will need to be added as they become available, although that could create (or compound) the CBDB's problems with consistency of geographical and chronological coverage. [End Page 195]
• New sources should be mined comprehensively, including all relationship data and not just what a particular researcher is specifically interested in.
• Automation is desirable for basic information like official postings, which account for a large proportion of the annals, but manual checking/adding would still be required too.
• All the above implies that CBDB requires a permanent maintenance staff to guide those adding new data, ensure consistency, check and integrate spreadsheets provided by researchers, etc.
Desirable features:
• Extra fields providing more detail on dates and official positions.
• Method for pulling out time-slices of who was in which government department/bureau in which year/month.
• A method for noting triadic links, and the direction of links.
• Permit user-defined categorisation of associations to suit specific research purposes.
The basic structure of the database seems robust enough to collect and manipulate most, and probably all, of the types of data that I need, and CBDB should eventually be very useful for work such as mine. Furthermore, CBDB's value lies not only in its collection and organisation of a vast amount of data, but in the ability of the discipline and practice of working with the database to themselves generate new research questions and approaches. In terms of encouraging wider use, this last point may be where the greatest potential lies.
"Native Incumbency and Elite Networks in Song Dynasty Sichuan: Evidence of the Mid-Eleventh Century from the China Biographical Database"
(Chen Song, PhD candidate, Harvard University)
Political and social transformations of the late eleventh century have been described by Robert M. Hartwell as (i) the increasing importance of regional divisions in public administration and bureaucratic rotations and (ii) the demise of a professional political elite from diverse regional backgrounds and the breakdown of their nation-wide marriage alliances. On the one hand, the balance of governmental power shifted from the central government to large regional commands, and there was a corresponding shift in career patterns from interregional ones in specific branches of the bureaucracy to intraregional ones that crossed diverse fields of administration. On the other hand, from [End Page 196] late Northern Song on, elites became more reluctant to migrate out of their home areas and more willing to marry locally. The implication is that both in government structure and in elite society, during the last two centuries of Song dynasty, China came to exist more as a federation of regional divisions than as a national entity.
Drawing upon existing data in the CBDB, this paper studies these transformations in Sichuan from the eleventh to the thirteenth century. Through an investigation into the incumbency of local government positions in Sichuan and the consequences for its native elite networks, this paper seeks to demonstrate that native incumbency and intraregional patterns of bureaucratic rotations in Song dynasty Sichuan provided the political impetus for localized social networks and marriage alliances.
The geographical isolation of Sichuan, its recent history of political independence in the aftermath of the Tang collapse, and repeated rebellions following the Song conquest of the region had all raised great suspicion among political leaders in the Song court. During the first hundred years of the Song dynasty, Sichuanese elites were not allowed to serve in positions of vice-prefects and above within Sichuan, and those who came out of Sichuan to serve in the court were forbidden to return to Sichuan even after they became aged. Although compromises and short-lived rescissions were made in 1018 and 1030, the tide was not effectively turned until the mid-eleventh century, by virtue of the confluence of a variety of social, political, and intellectual currents.
Government restrictions on native incumbency of local government positions in Sichuan were relaxed around the mid-eleventh century, and by the 1060s instances abounded of Sichuanese elites posted to significant local positions in Sichuan. Meanwhile, by the 1060s, there had emerged a circle of Sichuanese elites, among whom a discourse had been developed and circulated. This discourse, which exculpated native scions of Sichuan from its recent history of political separatism, was evoked in 1068 by a Sichuanese official, in his memorial to the newly enthroned Emperor Shenzong (r. 1067–85), in defense of the political allegiance of his landsmen and in support of postings of them to local government positions near their home areas.
Legacies of policy change since the mid-eleventh century and of the sociopolitical movement among native Sichuanese elites in advocacy for their own political allegiance were consolidated in the method of "delegated appointments" (dingchai fa 定差法), promulgated in 1070, which made consecutive appointments in Sichuan possible. [End Page 197]
Consecutive appointments in Sichuan, or intraregional patterns of bureaucratic transfers, cannot be substantiated by existing data in CBDB, for it requires the documentation of full office-holding histories for a significant number of Sichuanese elites, which CBDB has not covered yet. But existing data in CBDB does report a soaring number of native Sichuanese who, since the latter half of the eleventh century, held significant positions in prefectural and circuit governments of Sichuan, which had been hitherto almost exclusively staffed by officials of non-Sichuanese origins.
Native incumbency and consecutive appointments in Sichuan introduced into the historical scene a new group of sociopolitical actors, who were scions of influential local elite families and managed to penetrate into local governments. By the mid-eleventh century, officials of non-Sichuanese origins, who had nearly monopolized government positions in Sichuan, were seldom connected to each other and could hardly be identified as one cohesive group. Instead, each of them was oriented towards more cosmopolitan circles centered on prominent cultural and political elites based in the court far away from their administrative jurisdictions in Sichuan. Politically aspiring Sichuanese elites in early Northern Song wished to enter into these circles, and governors posted to strategic Sichuan prefectures played an important role as patrons and sponsors for these aspiring locals.
But since the latter half of the eleventh century, along with native incumbency, interwoven webs of social connections among Sichuanese elite families, often formed independently from the political structure of the ruling dynasty and sustained over many generations, crept into the administrative apparatuses in Sichuan as scions of these families were appointed to local government positions. These densely knit networks among Sichuan officials of native origins, which rivaled the cosmopolitan ones, were further cemented by agnatic and affinal ties. Prominent lineages managed to place many of their sons in administrative positions of counties and prefectures, and as such for generations. Intermarriages were actively pursued among these locally based office-holding elites, including prefectural administrators, circuit intendents, as well as regional commissioners in civil, fiscal and military affairs.
Social and political change in Sichuan since the mid-eleventh century unfolds a historical dynamic alternative to the received wisdom on the localist turn, a dynamic in which the consolidation of local power bases and the pursuit of political success may not exclude each other. The shift in the balance of [End Page 198] power from the central government to large regional administrations, native incumbency of local government positions, and the emergence of intraregional patterns of bureaucratic rotations, all made possible a cross-fertilization between office-holding in local governments and the consolidation of local power bases. Thus, the localized elite networks in Sichuan revealed not the dissipation of political aspirations among native elites, but rather a reorientation of such aspirations from the court to regional and local governments.
The macroscopic analysis of this historical change which CBDB facilitates, in spite of the paucity and historiographical bias of its data at the moment, attests to the strength of such relational databases in revealing macro-historical patterns and in testing and refining propositions pertaining to a long temporal range and a large geographical area.
"Using the CBDB for the study of women and gender? Some of the pitfalls"
(Anne Gerritsen, Warwick University)
When Robert Hartwell began his vast project of compiling biographical data for what would become the CBDB, women were not his priority. He was, however, interested in marriage alliances and migration patterns, so gradually data on women was included in the database. In my paper, I explore what the database can tell us about women, highlighting some of the pitfalls of such a project. I have largely relied on the data provided in the database, and only in very few instances supplemented the CBDB data with information from elsewhere. My point was to see what the database could yield in terms of gendered biographical research, and to test the accuracy of conclusions drawn from the database alone. One of the problems the database presents for anyone wishing to use it to study women and gender relations during this period of history remains that women were included on the whole only as daughters and wives, rather than as persons in their own right.
The associations and connections between individuals that characterize this database include only a small number of women. If we briefly take this group of women for whom we know they had "connections" as our focus, what can we find out? It turns out that most of these "connections" that involve women are based on a man writing a grave inscription (the database uses the term "epitaph") for a woman. The group also includes a small number of empresses and other exceptional females to whom certain men had largely [End Page 199] political connections, and their presence in the focus group problematizes our ability to make any statements about the group as a whole. We find, for example, that the group of women with associations had longer life spans than the women in the database in general, that a majority of women with associations (61%) had these with individuals with a different place of origin (the database uses the term "address"), and that most of the epitaphs (70%) were written by men with an address that differed from their recipients. The question remains, however, whether this group of women defined by their "having associations in the database" is a suitable object for meaningful prosopographical queries. Similar problems present themselves when we separate for analysis the group of women for whom we can trace kinship data in the database. Again, compared to the amount of data on kinship available in the database as a whole, the data involving women is very small. It mostly involves information about fathers and husbands, and, for example, hardly any data about kinship relations between women.
A more fruitful avenue of inquiry was to identify groups of women on the basis of their address. That strategy led to small sample groups, as the table below illustrates, but at least they form to some extent a meaningful group for separate analysis and comparison. For Jizhou 吉州, for example, we have data for 885 individuals, of whom only 15 are women, while the Jizhou population as a whole counted, according to Hartwell, around 22,400 households in 742, and 251,200 households in 1542. I asked a number of identical questions for the women in the database from a number of prefectures, including their dates (separating them into dynastic groups), their average life span, their age at marriage, and the number of children they bore. I also compared the status identification assigned to the women's fathers and their husbands. Finally, I investigated the nature of their connection to the men they married, asking whether they shared the same address, and where not, how and where the association between the two families was formed. This yielded the following table:
| Jizhou | Fuzhou | Mingzhou | Wuzhou | |
| size of sample (no. of women) | 15 | 11 | 24 | 19 |
| Dynastic group | 66.6% SS | 63% NS | 87.5% SS | 74% SS |
| median age | 66 | 49.5 | 66.5 | 70 |
| median marital age | 18 | 14 | 19 | 20 |
| number of children | 3 | 3.5 | 5 | 3 |
| Status | mostly same | inconcl. | mostly same | mostly same |
| F marrying outside M | 27% | 40% | 25% outside | 21% outside |
[End Page 200]
Clearly, the size of the samples makes drawing conclusions rather hazardous. Another problem that presented itself during the course of this research was the nature of the assignment of an address to an individual. My comparison of addresses between women and the men they married was intended to test the Hartwell/Hymes hypothesis that marital strategies changed between Northern and Southern Song, and that men started to marry locally. One could argue that the table above confirms that, considering that most of the low percentages (27% and lower) are based on Southern Song samples, and the higher percentage (40%) reflects a Northern Song sample. The problem, however, lies in the assignment of an address that this is based on, as a brief comparison of the data for one individual I worked with, drawn only from the database, with the data for that same individual drawn from conventional textual research presented in the paper by Fang Chengfeng, revealed.
The database is an extremely powerful tool for researching the socio-political world of pre-modern China. It also holds some extremely important information about women and their roles within that world. Some of that information, however, is not readily available as much of it has been included in the "notes" sections in the main biographical table and therefore cannot be captured by the queries one formulates to analyze the data. The 'LookatNetworks' in its current Access form can only handle a single individual at present. It might be useful to be able to explore the networks (within 2 nodes) of a group of people, whichever way that 'group' was composed. Finally, I found that even within this small study of women, the error level seemed high, especially for women, where the database only provides partial names. [End Page 201]
