Workshop: Network Perspectives on the Study of Chinese History

at Fudan’s ICSCC, Sept. 23-24, 2019

Welcoming remarks from Prof. Jin Guangyao
(Department of History, Fudan University)

Even though metaphors of networks and guanxi permeate scholarship on all periods of Chinese history, the theories and tools of network research are only gradually entering the field of Chinese studies. This workshop that I organized as part of my stay as a visiting scholar at Fudan’s ICSCC brought together a small group of scholars to discuss the advances and potentials of using new relational datasets and tools to study Chinese history and literature.

As the presentations demonstrated, network research offers new ways to engage with well-known sources or to make use of previously unmanageable quantities of texts. However, the required level of data curation as well as the necessary data and network literacy (not only of the researchers involved but also of their peers), poses challenges that can hardly be overcome by individual scholars. In general, digital humanities rely on interdisciplinary work. For the study of Chinese networks, however, we also need cooperation across national borders interlinking, for example, know-how in natural language processing with DH labs, or critical network scholarship with databases curated in foreign academic institutions.

This small event could only offer a glimpse into the emerging field, yet I hope that the ties forged here and the positive responses I received from many Chinese scholars will provide a starting point for further engagement with an even more diverse community of scholars and students.

And again many thanks to the team of the ICSCC for all their organizational support!

Workshop poster 中国历史研究的网络透视

Mapping Gaps in Wikipedia with Wikidata

A New View of the Tarantula Nebula. NASA via Wikimedia Commons.

I didn’t even know there were stars to look at to not see. If you don’t know that they’re there, you don’t know that you’re missing them. (Neil deGrasse Tyson, Source)

Internet users looking for China-related information on Wikipedia are very likely to come across empty search results, and even though the quality of Wikipedia pages on China-related topics increased in recent years, the quality of these less frequented Wikipedia pages is often not up to par. For example, the English page of Yang Xiaodu, a Politbureau member, was only created in 2015, the German version was created in November 2018 as part of the “Asian month” hackathon, one year after Yang got elected to the Politbureau.

While some scholars in Chinese studies might be relieved that is more difficult for their students to copy-paste homework from Wikipedia, for the general public this means that many realms of China’s history, culture, and politics remain invisible. Even within Chinese-language Wikipedia there is a clear bias: because of censoring mechanisms and the prevalance of Baidu in the People’s Republic of China, the majority of Chinese-language pages are created from users in Hong Kong and Taiwan. In short, in times of a steadily increasing international interest in China, Wikipedia’s representation of all aspects of China’s past and present is patchy.

I (like most other fellow sinologists I know) lack the time to create a Wikipedia entry every time I find that an entry is missing. However, I have decided to use Wikidata to at least “map the gaps”. When I notice that a Chinese person’s name, or the title of an important Chinese book is missing, I create a Wikidata entry. This way, I can always decide to go back and create Wikipedia pages to these items later and other users may find them and decide to create entries themselves.

For Chinese-language entries, you might have to change your settings, so I will explain how to do this below.

In pectore: Some thoughts on the Debate Over the New Journal of Controversial Ideas

No publisher or publication date is set yet, but the announcement of the launching of a Journal of Controversial Ideas to offer a platform for pseudonymous publishing stirred up heated debates.

A philosophy professor at Oxford, Jeff MacMahan, first proposed this idea in a radio interview with BBC. He stressed that the selection and peer review process would abide by the same standards as any other academic journal.

As was to be expected, the idea of pseudonymized academic publishing caused heated debates across the web. However, as I followed arguments and counter-arguments across the spectrum, I wondered why no one had added a perspective that might move beyond the comparative comfort of academic publishing in democratic societies.

I want to reframe this controversy and add a non-eurocentric view to the proposal of a pseudonymized journal. But first, I go through those arguments, before I add my thoughts:

Pro: Too much conformity stifles controversial debates at universities

According to MacMahan, the journal “would enable people whose ideas might get them into trouble either with the left or with the right, or with their own administration at their own university and so on to publish under a pseudonym. […] In current conditions, something like this is needed. We don’t want to allow it to be the case that people who have important contributions to serious issues are deterred by fear from giving their reasons and arguments for their views.” (Cited from the radio program “Do universities have enough diversity of political opinion?” by Matthew Flinders, BBC Radio 4, 25:30 min)

The whole idea of the journal is thus based on the belief that political polarization and online mobs endanger the freedom of academic discourse.

Con: Controversy needs bravery

As was to be expected, the idea of publishing anonymously, or in this case pseudonymously, raised concerns that authors would no longer take responsibility for their arguments.

As Nesrine Malik, a columnist at The Guardian, put it:

No one enjoys the feedback that anyone with any public platform receives these days. You can engage with it or dismiss it, but you should not claim martyrdom and hide from it, especially if you are going out of your way to incite it.

The underlying problem is a clear imbalance of the rules of the game of academia and the web: A scientist guarantees with his or her name for the scientific integrity and accuracy of their work. Once their research is disseminated beyond the academic realm (a genuinely positive process), however, those criticizing their work can hide behind pseudonyms and the anonymity of the web.

With its emphasis on upholding scientific integrity through a clear attribution of works to authors, the debate resembles the discussion on open digital collaboration in which scientists

“seldom frame participation in such projects
as a process of negotiating risk.” (Forte, Andalibi, and Greenstadt 2017)

Whether it be in collaborative online projects such as Wikipedia or in academic fields, external obstacles rather than intrinsic disinterest determine whether one chooses not to participate in a particular “controversial” debate.

To put it more bluntly, where does bravery end and risk start? In non-democratic societies publishing under your name might not be an option. So could anonymized academic publishing be an opportunity if we broaden our outlook beyond our somehow comfortable liberal institutions?

We, therefore, have to differentiate between those who might like to publish pseudonymously because they shy away from public debate and personalized criticism and those who have to fear persecution.

In short, those enjoying the privileges of academic freedom should stand by their work, yet we need to raise our awareness that for those colleagues working under conditions, where “controversial” equals “potentially dangerous”, a pseudonymized journal published within a democratic country might offer a platform to publish valuable work that might otherwise remain within the drawers of their desks.

Sources

“Do universities have enough diversity of political opinion?” by Matthew Flinders. BBC Radio 4.

Forte, Andrea, Nazanin Andalibi, and Rachel Greenstadt (2017): Privacy, Anonymity, and Perceived Risk in Open Collaboration: A Study of Tor Users and Wikipedians. CSCW.

Academy Award Sample Dataset – Part 3b: Importing your Dataset into Visone

[Click here to read part 1, part 2 and part 3a]

First, save your “nodes” sheet as a csv-file: Go to “save as” and then select “csv” as the data type. Name your new csv file for example “academy awards nodes.csv.” Do the same with the “ties numerical” sheet and name it accordingly.
Now open the network analysis software visone. (Download links and instructions can be found here).
In visone you find the file symbol (“open network from file”) on the top left. Click on it, chose your “ties numeric” csv file and confirm with “ok.” The program will then ask you to “specify your data format.” Select “CSV files.” In the next window, set everything as you can see below and then click “ok”.

Now, visone can already visualize your network.

I will later add a blogpost on how you can import attribute data and work with visone.

Academy Award Sample Dataset – Part 3a: Importing your dataset into NodeXL

NodeXL is a free extension of Excel, so if you use Excel, you can download the NodeXL Excel Template. Open the template, and you will see one datasheet named “edges,” here you can copy paste the content of your “ties” datasheet.

Note: If you do not use Excel, but for example OpenOffice, you can also go to part 3b and try out the free software visone.

In the “vertices” sheet you copy paste the content of your “nodes” table.

Click on “show graph” and you will get your first network visualization.

As I mentioned above, our dataset does not only consist of people, but also of films, different award types and award years. In NodeXL you can make these differences visible in the “vertices” section by adding visual properties. Klick into a cell and you will see a drop-down menu where you can select a shape for different node types (disk, square, etc.). You can add different colors for these shapes by right-clicking on a cell in the “color” column (B) and then on “select color.”

Academy Award Sample Dataset – Part 2: Creating a Spreadsheet that Represents a Network

Visualization in NodeXL, Oscar Award silhouette by Graph+sas [CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0)], from Wikimedia Commons

There are several possible ways to structure relational data, for example in a matrix, but the method of creating two columns of “sources” and “targets” might be the most intuitive, especially for historical research.
Before we structure our data in this way, we have to ask ourselves what exactly it is that we are connecting. Let’s look at the first two columns (A and B): Obviously, there is a link between a person and the “award edition.”

Academy Awards Sample Dataset – Part 1: Creating a Dataset for Network Analysis

Disclaimer: Below I will describe how I created a sample dataset for a network analysis workshop (which is now available here). There might be smarter or more efficient ways to do so, but I did not find any comprehensive guide online. So any comments or suggestions from more tech-savvy readers are very welcome!

1988 Academy Awards red carpet.JPG
Academy Awards Red Carpet, photo by Alan Light [CC BY 2.0 (https://creativecommons.org/licenses/by/2.0)]

These days, I was searching for a test dataset that I could use for a student workshop on social network analysis in Chinese studies at Freiburg University. However, even though many datasets are freely accessible online, I found that none was suited for a beginner’s workshop. Many sociologists work with pre-formatted datasets, but in most historical projects have to select, clean, and structure your data manually. So far, there is one helpful guide for historians by Marten Düring, where he also offers sample datasets and explains how to import data into Palladio. However, I decided to create a new dataset based on a Wikidata query, so the following documentation of this process can be seen as an addition to the guide on the Programming Historians site. It comprises three parts: 1) How to export data from Wikidata and how to structure data; 2) How to import data in NodeXL and 3) how to import data into visone.

Continue reading on page 2…

Encoding historical data – Why it is useful


In my daily work, I am constantly switching between languages, talking to colleagues in German, reading sources in Chinese and writing in English. The sources themselves are thus only accessible to those with the necessary language skills. But does my data have to be limited in this way? To make my data usable to people not familiar with Chinese characters, or to those users who prefer to read non-simplified Chinese script, I decided to encode the biographical data that I am collecting. This means that I ascribe every name of a person, institution, organizational unit and position a numerical value in a specific order.
Of course, ascribing written text, a numerical value has further advantages:

  1. I am tired of typing in the same characters over and over again. If I would get a penny for every time I wrote 常务委员会委员 (standing committee member), I might not have to write a grant proposal for the rest of my life.
  2. It can prevent typos, especially when typing Chinese names. (A short explanation for all non-Chinese speaking readers: When typing Chinese characters on your computer, you use the pinyin transliteration system, so that you can use the alphabet on the keyboard. However, several words can share the same Romanization, so when typing in “tongji” you might quickly end up with the characters 同级 instead of 同济. When typing in numbers, of course, typing errors might occur, but they are less likely and can be checked more easily.)
  3. It can force you to review whether a specific name, institution or organization has appeared in sources before (maybe in an abbreviated name) and decide which name to choose. Therefore it is only necessary to go back into your primary file and change the name that corresponds to a particular numerical value.

An example of coding historical data can be found here:

Shih, Victor, Wei Shan, and Mingxing Liu. “The Central Committee past and present: a method of quantifying elite biographies.” Sources and Methods in Chinese Politics (2010): 51-68.

Political and Academic Networks in the PRC

This blog presents a research project on the development of academic and political networks of Chinese scientists since the 1940s. It seeks to document the evolution of the project and offers readers a hands-on introduction to theories and methods of historical network analysis in Chinese studies, including sample data and visualizations. It aims to provide a platform for discussion with historians, political and social scientists as well as an international public.