Does Digital History Need a Theory?

Whoever has studied history has surely heard about the great controversies on historiographical theories and schools with their linguistic and epistemological ideas. It seems to me that the 20th century was full of radical position fights and historians walking from “turn to turn.” But today, they shy away from a theory based “digital turn” or – even a step forward – a “paradigm shift” of methodological thinking and narratives. Therefore, I wonder: do historians really need a “theory of digital history?” Do they even want it? Is it necessary?

Everyone who has read the essays of the so-called Fried-Althoff-controversy may get a feeling for how scientists fought without compromise. The conflicts between social history and cultural history showed that the epistemological interest of different historiographical theories lead to a polyvalent picture of “history.” So, this leads us to the trivial premise of a “theory of historical science:” there are different historiographical theories, because they have different epistemological interests.

Does digital history have a specific epistemological interest? If so, it would obligatorily need its own theory. The interest would need to be justified and we would have to develop digital methods to find insights aligned to these interests. As much as I think about it, my first impulse to answer it is: “We don’t need that!” Just a little example: If one would like to explore social networks in the modern era and uses techniques of “social network analysis (SNA),” doesn’t it fit into traditional positions of social history? Of course, it does. Network analysis is a method to find arguments for traditional issues and many computational methods could fit into theories or could enrich other conventional methods in the same way. Imagine an analysis of resources in an antique society based on papyri with bills and grocery lists (it seems to be a stupid example, but museums and papyrologists have definitely tons of paper bundles with information about a number of goats or corn, etc.).

Okay, all jokes aside! The actual point is, if we connect traditional ways of thinking with digital methods, we will get a sharpened picture of history which leads us away from “anecdotal based” to “data based” history. Digital techniques help historians develop data models with their sources to get statics, graphs, charts, and so on, which they can interpret regarding conventional theories. Has someone ever done a topic modelling with UN resolutions linked with geographical information about initiators and addressees and/or temporal changes? I expect huge insights for political history or Global History. You see, good old “history” works with digital methods.

The question will become much more complicated, if we replace the word “theory” in our question with the word “sources.” A while ago, I read an article on HSozKult about the “digital turn.” The author mentioned the danger of historians using more and more digital, accessible sources, as published sources or autographs would fall into oblivion. I feel sorry for historians who only sit on their computer and do historiographical research only by using Google and subject-specific database. Does “digital turn” really mean that historians stay at home in sweatpants, because sources have become available at all times?

My dissertation project is not a typical work of “digital humanities,” because when I started, I had no clue which way I should go. I had 20 boxes with manuscripts by Heinrich Bassermann, which weren’t even sorted. My first steps to “digital” were made by using an excel sheet with all the metadata I collected manually while I had reorganized every single paper into a new archival system. “Digital” started with chaos, dust, and dirty fingers. Who would expect that digital could be so analog… But in this case, through my work, I produced metadata, which then led to insights. This leads me to the question, if my sources are really sources or if data is my (new) source. Funny, isn’t it? Could “digital turn” eventually mean that sources produce sources, so historians can use more for their work?

You might say: “Silly! Sources are sources; data is data!” But the problem lies deeper. The traditional way from source to insight reads as follows: source(s) – inner/outer analysis – interpretation – insight. When adding “data,” this way changes: many sources – data modelling – analysis – interpretation – insight. The most important part of historiographical work to get insights – analyzing and interpreting of sources – hasn’t changed through digital methods. Well, the methods change. They need data. If analyzing and interpreting data leads historians to insights, does that mean that data is a new type of source?

While talking about sources, many different types of documents, etc. come into my mind. Letters, charters, epigraphs, books, and so on. In fact, we use collections of sources for our work already: editions! Would anybody question that editions are sources? At the library of the “Historisches Seminar” at the University of Heidelberg, editions are collected in the so called “sources room” (Quellenraum). In the same way, data acquired from sources is a new and comprehensive source.

To make it clear: no scholar can read millions of letters, no one can analyze all the League of Nations’ or UN’s resolutions, no brain can completely capture the entire content of the MGH. Yet, the machine does, and depending on our issues or methods, it summarizes information into numbers, patterns, charts, and tables. In this sense, the first step of digital history is a transforming process of sources to data-based information. But we can’t ignore the fact that during the process modelling, analysis, and interpretation grow closer together.

Every kind of modelling is a first interpretation of sources regarding the research question. To find the best methods is a decision based on assumptions as to what is going to be discovered. And finally, the interpretation is an interpretation of the applied models and methods. This brings me to the most fundamental question of a theory of historical science: Is this scientific work?

Based on source editions like Migne and MGH, and encyclopedias like Wikipedia and NDB, we know the discussions about scientificity. We invent categories like authorship, traceability, or even style and conventions to decide what is good and bad historical work. These categories are still missing for digital history, and that’s the gravest theoretical deficit.

The market for digital research in historiography is enormous. As I mentioned above, I see many potentials in some neglected fields and I see potentials to widen our view to history, historical events, or players in social networks. I skip the incomparably bigger field of digital sources like twitter, games, TV, etc. For sure, all these potentials and projects shouldn’t stand for them alone. They should be embedded in a theoretical concept which answers the question: “What does ‘Digital History’ really want?”

In my opinion digital history isn’t just a society of historians working with computational methods. I think it’s the paradigm shift away from individual considerations and away from generalizing narratives based on defined corpora to plausible and extensible data models. Discourses could change to: “Look, I found and added new data to the model and the results are now the following…” For this, we need a discussion about standards and a shift from traceability to replicability.

I’m glad to hear your comments!

Leave a Reply