Digital Humanities doesn’t work. That is what I, an initial believer and advocate, now think.
Let me explain why.
Back some fifteen years ago, as the Internet became ever more a part of our lives and as social media started to surge in popularity, I realized that my field of History, and the Humanities more generally, were in trouble.
It was obvious to me that disciplines that relied on printed texts and long-form writing were going to struggle in the years ahead. As I looked around for possible ways forward, I came across the nascent field of the Digital Humanities (DH), which at that time was primarily restricted to the Euro-American world.
At that time, there were various DH projects that were gaining attention, like the Valley of the Shadow (https://valley.newamericanhistory.org/), a website that brought together historical documents about the Civil War in America and presented them in a way that made it easy for anyone, be that person a scholar or an interested “commoner,” to explore and learn.
Seeing projects like this, I could see the potential of DH. The great contribution of the Internet was that it made it possible to easily share information, and projects like the Valley of the Shadow showed that academics could share their knowledge not only in new ways with fellow academics, but with the wider world as well.
What better way to show the continued relevance of History and the Humanities in a new age than by moving beyond the bounds of printed books and the ivory tower to bring knowledge and engagement to the community at large or even to a global community, right?
Further, at that time I had also started writing this blog, and I could see how a topic as “minor” as premodern Vietnamese history could gain wide interest as long as one discussed topics that people in general could relate and respond to.
So, I read up on DH, and around 2013, I co-taught a graduate seminar on that topic, and I also created another blog called “Digital Southeast Asia” (https://dseasia.wordpress.com/) to start sharing what I was learning and to try to get the “digital Southeast Asia” concept out there.
However, I quickly discovered some obstacles. The first was that my colleagues were very slow to see any potential in things DH, and the second was that the gradually-increasing number of people who were getting interested in DH were pursuing a different path than the public-looking Valley of the Shadow approach.
In particular, they were using DH tools to engage in ever-more obscure investigations of their already obscure research topics, and publishing their findings in journals behind paywalls.
Seeing all of this, I realized that DH would never work if we don’t have a way to communicate with people in the digital age, so I put DH aside, and turned to developing my skills at communicating knowledge through video and websites, etc.
Finally, after many years, I see that DH is starting to gain more adherents in the field of Asian Studies, but as was the case with scholars in places like the US and Europe who started to get involved in DH in the early 2010s, I find that much of the work that many people engage in is very obscure, and it is still getting “shared” in articles published behind paywalls.
Not wanting to single anyone out, I asked Grok to create some imaginary DH titles in Asian Studies, and this is what it produced:
- “Algorithmic Shadows of the Tang: Topic Modeling Poetic Structures in 8th-Century Chinese Anthologies”
- “Networked Monks: Mapping Dharma Lineages in Chosŏn Korea with Graph Theory (1392–1910)”
- “Sentiment Analysis of Meiji-Era Japanese Diaries: Quantifying Emotional Shifts in Modernization Narratives”
I’d say that this is pretty accurate, and is indicative of the type of DH scholarship that I generally see in Asian Studies these days.
I am aware that there are some places in the world that are pushing Humanities departments to embrace technology or face elimination. These are in places that are very fixated on metrics and performance. However, as sexy as A.I. and machine learning sound, if they are used to dig ever deeper into the ivory tower. . . I don’t see them ultimately being able to help anyone.
Out of curiosity, I checked some of the Google scholar citation numbers of certain people promoting DH in Asian Studies. They were shockingly low. In a world where many universities care about such metrics (such as the ones demanding the embrace of technology), DH is obviously not going to deliver.
If it is not going to bring in the metrics that professors rely on, will it bring in students? I doubt it, because I don’t think that there is a mass of young people out there who are excited by the type of work many DH scholars are engaging in.
A few years ago, I was at a small lunch with two ambassadors. While a group of us sat and listened, the two ambassadors talked on and on about. . . history.
What did they talk about? They talked about basic things like why Meiji Japan modernized before places like China and Korea.
What did they not talk about? They did not talk about things like “Sentiment Analysis of Meiji-Era Japanese Diaries: Quantifying Emotional Shifts in Modernization Narratives.”
In the previous post, I shared some recent comments by OpenAI CEO Sam Altman about the future of History and historians in the AI age. When asked about this, Altman said that there will be some job that has something to do with telling stories, but that the current job of an historian will not exist.
I agree with that, but I would also argue that much of the DH work out there these days that is getting produced in academic settings does not tell the kind of stories, or in the form of communication, that Altman was thinking of. I’m pretty sure he was thinking more about the stories the ambassadors told each other, and more public forms of communication like the Valley of the Shadow model.
I was lucky enough to have lunch today with the preeminent scholar (actually he’s the only scholar) of Ming dynasty literature in my country, and we were talking about the future of historical scholarship in the context of technological advancement. I said I was doubtful as to whether a large language model would be able to create original scholarship on premodern Vietnamese history within a decade given that the primary source materials are contained in undigitised texts in so many different languages (Cham for example) that computers cannot even process yet, and I hadn’t even touched on archaeological artifacts, oral folklore or linguistic data, all of which constitute important materials for the study of premodern Vietnam.
He said people were missing the point anyway, and that the journey of historical discovery was a goal in itself, a goal missing from answers that could be obtained with minimal effort.
Although I myself agreed with him on this, I can’t imagine that the people who count money and metrics at universities will be thinking this way when it comes to considering an academic’s performance. Bean-counters want results and the digital humanities is a way of getting those results quickly and efficiently, but reliance on technology does not necessarily advance historical scholarship in the long run. Using technology to plough through digitised texts can provide quick “wins” like the hypothetical examples you provide above, but the danger is that the historical information locked away in undigitised and hard-to-read documents and other sources will get passed over as research that takes time and effort (and the learning of scripts and languages) will be neglected in favour of these easy outcomes.
I can see why Sam Altman thinks that historians won’t exist in the future, but I don’t think it is because technology WILL soon have the ability to replace good historical scholarship, it’s because people (and especially those who count money and metrics and also people who work for or run LLM companies) BELIEVE that technology will soon have that ability. Some big questions can probably be answered by AI-assisted corpus analysis, but unless people can add new materials and perspectives from actual academic study of documents and other materials, the writing of history will end up as little more than technology-assisted regurgitation of slop based on the same old set of materials and perspectives.
Actually, I assumed that Altman’s pauses and his evasion of the question indicated that he was thinking to some extent about the bean counters (but maybe I imagined that because that is what I think?) and knew that there is no way that History in most university has any chance to produce the beans to survive against AI.
However, I don’t think DH has much hope either. Yes, administrators might think that it’s a quick fix, but when they see the Google scholar citation counts of 36, 44, 23. . . I think their patience will run out quickly.
When the whole bean counting enterprise began, the administrators said “Yes, we fully understand that the Humanities is different from the sciences. . .” As I think anyone could have predicted, that “understanding” didn’t last long. I think the same thing will happen with DH. With words like “AI” and “machine learning,” it sounds like a fix, and that will buy it some time, but it doesn’t produce beans, so. . . the bean counters will eventually get frustrated.
On another note, I think a major problem with all of the “defenses” that people come up with, like the journey of discovery is the goal, or that we teach “critical thinking” is that while certain individuals may uphold these practices/values, many do not, and institutions do not, and it’s not what most students think they are there for.
When you teach a class by reading word-for-word off PowerPoint slides or by uploading videos once and reusing them for years and don’t respond to student emails (I’ve been listening to young people complain about this recently), etc., or when your research is something totally obscure, then you have made your profession replaceable by AI. The AI doesn’t have to be powerful enough to do things that are “good,” it just needs to be better than whatever the average is out there, and by this point, I think that the averages are pretty darn low for a lot of things that happen in universities.
I started using Duolingo a while ago and as I did, I thought to myself, “Hmmm, which would I rather engage with: an app that leads me through a well-designed curriculum, or a bored and bitter prof who monotonously reads of PP slides?” After a few months, Duolingo started to annoy me, but if I really needed to achieve something, I would still be much more motivated to stick to it than to the prof reading off PP slides.
Then at more of an institutional level, we can also see ways that programs have been made replaceable.
In the US, in the past 20 or so years, there has been a big move to make transferring between universities as easy as possible. So, if you take History 101: World History at a community college where it is taught by someone with an MA in (pick any field of History) or at a “tier-1” research university where it is taught by a full professor who is a leading scholars in that field, those two courses are considered “equal” when it comes to transferring credits.
So, if we have already established that one doesn’t really need specialized expertise to teach that course, then what’s wrong with getting some people to create a good app to do the job, like Duolingo for History 101, and replacing everyone? I’m sure that the bean counters already realize this, and I could be wrong, but I’m assuming that something like this thought passed through Altman’s mind as he paused to answer and then turned the topic in a different direction.
“[U]sing DH tools to engage in ever-more obscure investigations of their already obscure research topics, and publishing their findings in journals behind paywalls” made me chuckle.
Christopher L. Caterine, in his book Leaving Academia: A Practical Guide, talks about realizing that the global audience for his academic writing on the subject of his PhD in classics, the poetry of Virgil (or something like that), was about a dozen people. I.e., academia loves to create tiny echo chambers where obscure researchers discuss their obscure research in a completely obscure manner. Caterine then became a corporate consultant.