Network analysis, chronologies and the history written around Chinese students abroad
At the beginning of May I attended a series of talks on historical network analysis organised by DH Ratio at the Free Université of Brussels. Historical network analysis is, simplistically put, studying connections between objects (mostly people, but can also be places, documents, words) that are traced or signalled in historical sources. The connections can be any-thing: from someone following someone else on facebook to borrowing money to mentioning someone in a letter to people posing in the same photograph to names co-occuring in a biography to migration paths between villages. The primary goal of network analysis is not to find out whether there are connections between things, but how a particular object is situated in a particular set of connections (network). It concerns questions like: Where are the centers? Which objects form the centers of all connections? Which objects bridge the different centers? Which one is isolated? etc. Basically, network analysis is like interpreting a geographical map, except the map is more rudimentary, broken down into dots (nodes) and all the lines between them (ties), with measuring metrics and visual aids at one's disposal to dissect the structure(s). It is a useful method in my research as, for example, I want to know about the organsations that facilitated Chinese-Belgian educational exchanges and in particular how they related to each others in terms of sharing people, revealing clusters (or lack of) in institutional coordination for Chinese student migration in Belgium.
One of the talks, which is given by Sébastien de Valeriola, dealt with the incorporation of time in the visualization of networks. To render structures of networks visible, all nodes and ties are typically drawn in one big space. This kind of visual mapping is sometimes problematic because it ignores the temporal aspect of the connections being depicted. It makes it difficult to understand how a certain hub of connections is constructed over time. One tie after the other? All at once? Or something in-between? The talk was about how to mitigate that problem, presenting a series of experiments on how to include chronology in network analysis, which I won't detail here, because to my knowledge the experiments have not been shared pubicly in any form.
Following the presentation, I started to think about another issue related to chronology. My question in its most basic form was as follows: if we include time in our analysis, what kind of time are we using? Because, despite what the suffix claims, chronology is never 'logic,' but something people culturally agree upon and practice. I mean, what is more proof of that than the familiar question "What time are we now?" If we structure our data on a timeline to observe patterns, would it be appropriate to apply a modern timeframe to interactions of people in the past who structured their life with a different understanding of time? And, when I talked about this issue with others during this session, another layer of complexity came into frame: The historical sources. They, too, have their own logic of time, whether it is a diary that grows one entry at the time and the iterations of a manuscript.
After the talks, the problem of "chronological frictions" in analyzing historical relations stayed with me and marinated in my head throughout May. When I think about my initial question in this blog post, I am aware that selecting a timeframe is not a problem inherent to historical network analysis, but also to the writing of history itself. The timelines used to bring order to past events have their own story, some contested, others unopposed. I see the frictions also playing out in the historiography on Chinese students in Belgium. For example, Belgian scholars start the story of the Chinese students in Belgium in 1904 whereas Chinese scholarship says 1903. The difference? 1903 is when Chinese students left China. 1904 is when they reached Belgium. Both years reveal the author's point of reference (China or Belgium) in their writing and they derive their point of references mostly from governmental sources, meaning their timeline is mostly informed by state actors who tend to manage Chinese overseas students instead of the students themselves.
I think chronological discrepancies are not something to be solved, but something to shine a light on for others, whether it is to analyse a part of the past or make it meaningful. Don't know how to put it into practice, but I have time to experiment with - well - time in my writing.
Well, these are just my two cents. Thank you for reading and have a good day.
Nora