Interview with Dr. Tamás Kiss, historian and software developer, on the effects of the digitization of the humanities and Ottoman text analysis software Rumi 1.0.
Earlier this week at a Digital Humanities conference in Budapest you claimed that people like you, despite their best intentions, are bringing about the demise of history. What did you mean by that?
I certainly didn’t mean it in any teleological sense. I was simply making a point that as a result of digitization, the humanities disciplines are transforming at an amazing speed. For instance, one of the consequences of digitization is that interpretation, which is at the core of humanities research, is being outsourced to the end user, so to say. It’s becoming more and more difficult to argue for a select few to get to the sources in far-away archives when you have the same sources available online. Even graver for the historian, when a schoolchild can have access to reliable and free tools to pre-process, translate and analyze primary sources, and then visualize the results, nobody needs Peter Burke or Quentin Skinner to interpret early modernity for them anymore. Having said that, while the resultant democratization of interpretation, critical pluralism, decline of the canon, and so on, may seem like good things, they’re in fact false illusions. People usually don’t spend their time reading thousands of pages of primary sources unless they get paid for it. Nevertheless, humanities education and academic work are losing their prestige in the process, and politically backed institutions and the for-profit sector are happy to step in to provide people with interpretations that serve interests other than scientific ones. So even if some of us are lucky enough as programmer-humanists to take part in the digital transformation of our respective fields, it’s probably worth keeping in mind that when the job is done, we will have no other option than to look for a job outside of academia, either at an IT company or somewhere in the gray zone between business and politics. This is of course an exaggeration — the “job” will not be done in the near future. However, the new technologies and methodologies that we, on the production side of the digital humanities, work with, already outperform traditional modes of research multiple times. Add in e-learning and AI, and there go all the teaching and research jobs with them. The software I’ve just finished yields in a few minutes the data that a researcher would need to work for over a year to compile. Of course, you’ll always need your history high-school teacher or celebrity historian like Yuval Noah Harari. But as a result of mass-scale digitization, our profession is more likely to be practiced in the future by programmer-historians outside academia.
You have already done your part in unintentionally bringing down a corner of academia with your text analysis software for Ottoman Turkish documents. What makes Rumi 1.0 so special that you went as far as to say that it opens new frontiers for researchers?
Although there are several out-of-the-box text mining and text analysis apps out there, none of them is applicable to Ottoman Turkish documents. In general, the difficulties that heavily inflected agglutinative languages, such as Ottoman Turkish, raise in computational text analysis, cannot be solved without language-specific applications. Programmatically finding a word’s root, telling the computer what it should consider as a word, or character-encoding become very complicated when you want to apply them to Ottoman Turkish. Computational text analysis relies on statistical methods which can be performed very easily with the help of some of the most common apps out there. But the statistical phase of the work requires input data, and the existent tools simply don’t extract such data from Ottoman Turkish documents. Rumi 1.0 does, and by that it gives new perspectives to researchers who so far have had to resort to analyzing and comparing to each other Ottoman texts manually. Let’s say, for instance, that we want to pinpoint parts of a source that are in the pen of an author other than the one the work is generally attributed to. I’ve got this software to help me, and I’ll be finished in 15 minutes. You — someone who’s got a pen and a piece of paper — will have to work arduously for months to do the same job and get the same results.
That’s great, but why should we care?
I don’t think you should! The truth is that my software will be useful for maybe 100 scholars in Ottoman studies worldwide. Maybe more… But while the vast majority of Ottoman archival materials is still waiting to be looked at, other historical fields are already using immense computational power to process their data. If studying a polity that had so much influence on the world around it as the Ottoman Empire falls behind the general infrastructural trends applied to other fields, we inevitably risk a lopsided representation of the past. For instance, the Venice Time Machine project already envisions applying AI to discover new connections between historical data yielded by Venetian historical sources. If you compare that research capacity to the output of analog Ottoman studies, it’s not hard to see that we’re headed toward a situation where, in Veneto-Ottoman relations, the Ottoman perspective will be dwarfed by the overwhelming richness of Venetian data.
So, is it that the equilibrium of historical interpretations could become unbalanced without the use of these kinds of tools?
Yes. However, one of our most important sources on anything Ottoman is already Venetian sources, especially intelligence dispatches.
Am I right to think that you’re overall pessimistic about the future of the humanities?
No, I’m simply at the forefront of the transformation of our disciplines, and I’m aware of the present trends. What’s happening right now is that the practitioners of history and literature are looking for ways to get integrated into a new, digital economy, and justify their existence in an ontological situation radically different from what we’ve known so far. Digitization will unavoidably put an end to the humanities as we know them today. However, at the same time, we may perhaps have a say in how our disciplines go down and get reborn in a new, digital, format.