Historians are becoming computer science customers – postscript

By Adam Crymble

Historians are becoming the clients of computer science departments. It’s a quiet conspiracy to steal our research funding and push out employment opportunities for our junior historians whose skills just don’t fit the needs of these projects. In our final seminar of the 2014/15 season, we heard from two interdisciplinary projects that included a mixture of historians, archivists, computer scientists, and natural language processing experts. If you want to find out about the projects in detail, check out their websites, or watch the talks online:

• ChartEx: Charter Excavator
  • [Columbia University, University of Brighton, University of Leiden, University of Toronto, University of York, University of Washington]
• Traces Through Time
  • [The National Archives, Institute of Historical Research, University of Brighton, University of Leiden]

Both projects depended on natural language processing – the rapidly improving ability of computers to structure and break down digitised texts in ways that are not entirely unlike how humans perceive their own language: full of types of content with various meanings that fit together in flexible but not entirely random patterns, rather than an unrelated series of letters, numbers, and punctuation. It’s safe to say that neither project could have gone ahead without natural language processing. I’m not sure we can say the same if there had not been historians on board. The results may have looked different (probably even worse), but instead of historians being a crucial piece in the puzzle, I think we’d see that without the historians the projects wouldn’t have happened because nobody would have proposed them. In this case, what we see is effectively historians (or archivists) hiring computer scientists to solve their problems. Or to put that another way: both of these projects were computer science projects with applications for historians.

Don’t get me wrong, the results are fantastic, and I think both of these projects are great as are all the team members involved. I also think that both historical studies and computer sciences are benefitting from these collaborations, and I hope we see more of them. This post is not a reaction to what I saw yesterday at the seminar or anyone involved in it, but is instead a reflection on some of the issues that hearing about these types of projects raised for me as a member of the audience.

I am concerned about the power dynamic this model is creating in the sector, particularly because of who is picking up the tab. At the moment, despite the fact that these are computer science projects, these types of projects are being initiated by historians – they have the problems that need solving. That means historians are finding the partners, pitching the challenges in interesting ways, and then turning to the funding bodies that they know best: in the UK that means the Arts and Humanities Research Council (AHRC) or the Economic and Social Research Council (ESRC). It would be utterly unfathomable for a historian to apply for a grant to the natural home of computer scientists, the Engineering and Physical Sciences Research Council (EPSRC), but that’s exactly where we need to go.

The fact that these computer science projects are of benefit to historians is completely irrellevant. If an engineer’s research has direct relevance to medicine (such as research into the invertebral disc in the human spine with the aim to ‘inform implant design‘ – a clear medical application), it’s still funded by the EPSRC, rather than the medical council. That’s because it’s an engineering project that happens to have applications for other fields. But within the digital humanities, it’s almost exclusively the AHRC and ESRC that are footing these bills and supporting this work.

That’s a problem because according to a report by the Department for Business Innovation & Skills (BIS) in the UK, the AHRC and ESRC are the poor siblings of higher education funding accounting for only a combined 9.4% of the pot – less than any other research council gets on its own. That in itself is a powerful statement about the government’s belief of the value of arts, humanities, and social sciences. But the problem becomes even worse if that meagre budget must filter its way into the pockets of computer science departments instead of into the hands of our next generation of talented humanities scholars.

What we’re seeing then is historians paying the upkeep and development of future  computer scientists, while their own graduates scramble for a handful of poverty-waged Junior Research Fellowships at Oxbridge, or take on equally meagre wages as hourly paid lecturers while they finish the ‘book’ that they’ve been promised will be the solution to all of their employment problems (it won’t be, by the way. Don’t believe that). That’s not to say junior scholars in the sciences don’t struggle to get by, but from my experience through a spouse in academic engineering, students in engineering cannot fathom the idea of doing work for free (writing a book) because they have been raised to believe that their work has monetary value, and the relative abundance of funding that is used to build teams of scholars under the mentorship of a senior scholar supports that belief. In contrast, humanities scholars are taught that their work must be a labour of love, and any money that comes their way is incidental to their progress as a scholar.

That has in part led us towards this client-supplier model, in which historians are forced to hire computer scientists (who have obvious value and costs) to solve their problems. While historical graduates (who have no obvious value) cannot be fit into the grant because there isn’t really all that much they’re needed for on these computer science initiatives. What we need to see is a shift away from the client-supplier model, and towards one that is mutually supported, in which the EPSRC or the commercial tech sector supports the computer science work, and the AHRC or ESRC supports the humanities contribution to those initiatives. That requires a rethink of the boundaries between our research councils, but it also means that the junior historians we want to hire need to make sure they’ve got skills that make them employable on these types of projects. Sure, you’ve got a PhD. But what can you do on my data mining project?

Adam Crymble is a convenor of the Digital History seminar at the IHR and a lecturer of digital history at the University of Hertfordshire.