Any report with 25 authors must be taken seriously. Or perhaps promises to be fairly dull, since 25 people are unlikely to agree on much that is really innovative. So it wasn’t with huge expectations that I started reading this paper, The challenges and prospects of the intersection of humanities and data science: A white paper from The Alan Turing Institute, however worthy the subject.

Certainly, the seven recommendations are somewhat bland. For example, recommendation 6 is:

We acknowledge the need to upskill humanities researchers in quantitative and computational methods

Which sounds admirable, until you read the qualification that follows immediately after it:

If they wish to.

Here is the problem in a nutshell: what if they don’t wish to? I think the authors of this paper have identified perhaps unwittingly a major difficulty, perhaps the biggest real challenge, of digital humanities. Too many humanities researchers lack vital digital skills.

Of course trying to define the term “digital humanities” gets us into difficulties. The authors point to a site by Jason Heppler that lists 817 separate definitions of the term. I’ve mentioned varying definitions of this term before, but vague terminology does not explain this particular issue.

Some anecdotal experience is relevant here. Some years ago, when I was living in Oxford, I was fortunate to be invited to join an informal lunch for anyone involved in digital humanities (which in practice was almost entirely based around the University). There was no formal agenda, simply an opportunity for academics in the humanities with IT people who had specialised in the challenges of digital solutions for the humanities. The participants included James Cummings, one of the authors of the Turing report. At those lunches, many academics realised that their requirements for software were not so exceptional, but shared with many other humanities projects – for example, to build a concordance. So there was an immediate benefit, since researchers discovered there was no need to reinvent the wheel.

But more than that, the discussion became a two-way process, so that the humanities academic would learn more about what digital tools could or could not achieve in a cost-effective way. Some humanities researchers have unrealistic ambitions, and it is far better to have a pragmatic chat before going any further.  Remember the fundamental principle in computing that there is almost always a digital solution to a problem, but you might not like the cost, time, and risk involved. As in other areas of IT, an effective solution can often be found by informal sharing of ideas by experts in different areas.

I didn’t know it at the time, but the term today for such an IT person is the research society engineer. James Baker wrote an excellent post about this role, and Melissa Terris (and several others) published an article about the use of data sets for humanities research.

As I see it, the nub of the problem with digital humanities is “if they wish to”. Unlike researchers in the sciences, who are expected to have sufficient awareness of what is or is not achievable programmatically, humanities researchers all too often lack any kind of computing awareness. Hence the need to a partnership with sympathetic IT professionals who can listen and make suggestions. Unfortunately, leaving it as an option for the humanities researcher to learn about quantitative and computational methods is pretty much guaranteed to produce poor results. I don’t want to suggest coercion, but for better or worse, humanities is today a digital discipline and researchers need digital skills. Perhaps such skills should be a component of any postgraduate course in the humanities. Not so much “if they wish to”, but because today, they don’t really have a choice.