Retracing Professional Mobility: Historical Network Analysis through CERD
By Thorben Pelzer
If sinology was a religion (which, despite its dogmas, it probably is not), the 15-volume Cambridge History of China could well serve as its holy scripture. Published over the course of over forty years, each of its well over a hundred chapters provides an authoritative overview of a particular topic. Authoritative, yes, but not necessarily consummate. Take for instance E-tu Zen Sun’s 1986 introduction to science and education in the Republican era (1912–1949). As scholars of Chinese history are well aware, the Chinese education system received its major overhaul in this period. Students of ‘Western learning’ found that science and technology could help get their country back on track. It was, as Frank Dikötter put it, an “age of openness” to foreign ideas but also an era where these ideas were applied under the premise of nationalism and nation-building. In her Cambridge History chapter, Sun lamented how little historians knew about the academic protagonists of her story, about their “personnel configurations, groups, factions, and cliques.” With vision, she hypothesized the spatial findings of a yet unwritten, more in-depth study. For instance, she assumed that college towns and urban centers must have drained aspiring experts from the countryside and peripheries. Nevertheless, she was well aware that her sample size of roughly seventy students was too small to substantiate her assumptions.
For the time, her amount of data was still remarkable, and we must remind ourselves of the cumbersome manual methods used to survey and tabulate. But since then, the introduction of digital humanities (DH) has amplified what scholars consider a manageable sample size many times over. DH as a term is peculiar in that it means neither a discipline nor a method. Instead, it describes a skillset of exploiting modern machines for knowledge production. As long as the historical data is available and the source material printed or written down orderly enough for machines to read it, it no longer matters if we deal with a hundred, a thousand, or tens of thousands of biographies. As long as we know how to assess the data, we can trace a manifold of long-dead individuals at once, stalk their movement through space, and discover where they met and collaborated.
Such is the goal of the Chinese Engineers Relational Database (CERD). It is part of a research project on engineers in pre-war China (1906–1937), operating within the Leipzig Collaborative Research Centre (SFB) 1199 on “Processes of Spatialization.” The idea of the database is to provide a pool of digitized and cross-referenced data that researchers can then analyze to answer their own specific research questions. For the Leipzig project, these questions concern the ways in which engineers operated under, but also shaped, formats and visions of space.
As the name of the database suggests, our aim is to gather information about every Chinese individual who worked as a professional engineer in the first half of the last century. Where available, their biographic data may include time and place of birth, educational institution, major, degree, and year of graduation, and the dates and location of each work commitment. When inputting data, we always take the individual as the starting point. In this respect, the database can act as a simple encyclopedia to conduct research on over 15,000 persons. However, the network that has evolved in the process is much more than that. CERD provides insight into the rise and fall of popular universities, the scope of important infrastructural projects, and the organizational structure of historical companies. As each location is georeferenced, we can also map academic mobility and workplace migration on different regional scales, for instance on a national, provincial, or county level.
See, for instance, this demonstration of temporal shifts in workplace allocation, using but a small set of preliminary data from last year. You can use the slider at the bottom to pick a year and hit pause to hover over the heatmap for more information. Scholars of Chinese history may recognize, for instance, the early importance of Hankou, the gradual development of Shanxi, or the migration to the Southwest after the invasion of Manchuria.
In a sense, the main contributors to our database are as deceased as the historical engineers themselves: when in 1941, the National Resource Commission, a powerful institution of China’s nationalist command economy, compiled a voluminous register of all the engineers at their disposal, they did much of the work for us. However, we have since learned that an academically sound historical database takes more than just rendering existing registers machine-readable and applying natural language processing to tabulate the information. Spelling and printing errors are frequent, and the interviewed engineers often referred to the same employer using a handful of different names. Merely collecting records and assessing them in a ‘big data’ fashion would significantly distort the outcome. Manual editing, including proofreading, merging, and linking entries, is indispensable. While minor typos and spelling differences can be merged semi-automatically, the history in which towns, schools, and companies occasionally changed their names is so complex that no algorithm can replace a historian who researches the necessary background information.
Also indispensable is the know-how. When conducted seriously, historical network analysis involves more than creating an Excel file and calling it a database. Despite a rise in accessible tools, setting things up is not a one-click-affair. When learning how to use citation software is the most digital that curricula in the humanities usually get, and our friends from the computer sciences can make a fortune outside of academia, many projects are prone to fail due to the lack of skilled personnel. Inputting data through collaborative software like Heurist is easy and can be undertaken by any student assistant, but planning the database design at the onset and, even worse, querying, re-arranging, and assessing the collected data down the road are skills one cannot acquire during a lunch break. Despite the time investment, the training still pays off as the machine-supported arrangement of data allows scholars to explore questions that cannot be answered by traditional, analog methods.
CERD will make for some interesting papers, including research questions the team and I have not yet thought of. That is because the data is publicly available—an invitation not just to review any claim we may make in the future, but also to export and assess the data for one’s own research projects. You have no interest in engineers, but are an expert in family name distribution? You could use the data to check if the claim that the name Wang is most prevalent in the North holds true and how workplace migration may have helped to disperse it. More importantly, we are in contact with research projects that assess biographic data from other parts of society. Combining and comparing these datasets could eventually complete the picture, if appropriate funding was secured. Did political, infrastructural, and artistic actors share centers of activities, and did their movements correspond to the same world events? By keeping track of their mobility, we can understand how their movements helped shape different Chinese regions and how they played a role in reordering them in a transregional context.
With all 15,000 biographies from the 1941 source put into the system, CERD presents a good idea about the approximate spatial distribution of educational degrees and employment. Nevertheless, there are always more primary sources waiting to be inputted to help complete the picture: biographic information found in alumni magazines, membership lists of professional societies, and archival material such as photographs. This is why, although the full 1941 dataset is now available, we will continue to release updates for years to come.
CERD is browsable through a web interface. As an exportable dataset, the latest version of CERD is published here.
 Dikötter, Frank, The Age of Openness: China before Mao (Hong Kong: Hong Kong University Press, 2008).
 Zen Sun, E-tu, “The Growth of the Academic Community 1912–1949,” in Republican China 1912–1949: Part 2, ed. Fairbank, John K. and Feuerwerker, Albert, The Cambridge History of China 13 (Cambridge: Cambridge University Press,  2002), 399.
Dikötter, Frank. The Age of Openness: China before Mao. Hong Kong: Hong Kong University Press, 2008.
Pelzer, Thorben. Chinese Engineers Relational Database (CERD) Design Manual. Working Paper Series SFB 1199, no. 21. Leipzig: Leipzig University Press, 2020.
Zen Sun, E-tu. “The Growth of the Academic Community 1912–1949.” In Republican China 1912–1949: Part 2. Edited by John K. Fairbank and Albert Feuerwerker, 361–420. The Cambridge History of China 13. Cambridge: Cambridge University Press,  2002.
Thorben Pelzer researches Chinese history at Leipzig University. He graduated in Chinese Studies, Japanese Studies, and East Asia Studies at Ruhr-University Bochum. He also studied at Osaka University, at Tongji University in Shanghai, and at National Taiwan Normal University in Taipei. He is an alumni of the German Merit Foundation. As a social and cultural historian of the Republican period (1912–1949), his dissertation follows the professionalization of Chinese civil engineers after their graduation from US colleges. He currently researches Chinese engineers at the Collaborative Research Center (SFB) 1199 in the context of spatialization.
Further articles in the Digital Humanities Interface series on TRAFO:
Ninja Steinbach-Hüther and Thomas Efer, The Digital Humanities Interface – An Introduction, 28 September 2021
Citation: Thorben Pelzer, Retracing Professional Mobility: Historical Network Analysis through CERD, in: TRAFO – Blog for Transregional Research, 11.11.2021, https://trafo.hypotheses.org/31210
[…] Pelzer, Retracing Professional Mobility: Historical Network Analysis through CERD, 11. November […]