JS: Internationally, Digital Humanities has developed strongly in recent years but New Zealand hasn’t kept up, in part because of a lack of policy documents we can point to as demonstration of its importance. International development is so strong, however, that NZ is now following along – activity is growing because we can no longer afford to ignore the challenge. Some of this lack of progress is linked to the “crisis in humanities” – the government’s focus on funding STEM education (Science, Technology, Engineering and Math) to the detriment of humanities and social sciences. I would like to see humanities integrated into national strategies for investment in eResearch along with the STEM groups.
JS: From an investment perspective, some big international projects in Digital Humanities and cyber-infrastructure have failed, which is perhaps an indication that big bang investments are not the right approach. I think a better approach for NZ would be to manage risk through integrated investment strategy for humanities alongside the STEM programmes. That said, in a multi-discipline environment it’s unlikely that a single approach will work equally well for all of us.
JS: It’s hard to know what all the requirements might be for a future Digital Humanities eResearch infrastructure. Perhaps in humanities we need to lead with a theme rather than with a technology. Traditional requirements’ setting processes are perhaps too narrow because the set of users in humanities with the same requirements might be too small to justify the investment.
JS: A really big barrier for humanities researchers participating in eResearch is a relative lack of technical skills. Humanities post-grads may have amazing ideas for digital work, but no concept of the limitations of technology and the resources it might take to achieve their aims. They see what’s possible with Google, but don’t grasp the levels of investment and technology inputs required to make that happen. They need much more than government funding – they need the technical help to design and execute their ideas.
JS: Humanities, like many domains, has domain-specific quality markers for digital work. For example, in Digital Humanities we have our own form of XML that allows us to mark-up texts in a much more granular way. Building a scholarly interface for one of Homer’s epics, for example, is a major challenge that requires the ability to ascribe many levels of citation and interpretation to individual words and phrases – some passages will have centuries of critique and commentary to go with them. Topic and language modelling is another example; as a scientific tool it has been developed for a non-fiction world … but how can we apply it to poetry?
JS: All this said, we can set out a high level set of infrastructure requirements for empowering Digital Humanities:
– We need servers and software tools on demand;
– We need our blogs and commentaries archived and searchable;
– We need toolsets and apps (such as the MEANDRE workbench) that allow analysis;
– We need help to learn how to use these tools and to think and plan in technology terms that are realistic;
– We need to establish a publishing workflow that publishes our work to a space we can control and that manages the archive and search process / DOI capability, and
– We need a level of standardisation and accreditation of our algorithms so we can show our work is repeatable science.
JS: One site that is available is www.academicAMI.org. This is basically a sandpit environment for learning how to use web apps and toolsets. I see Digital Humanities researchers using this kind of sandbox to learn how to begin their work, so long as there’s a technical expert available at some stage to help them evolve it to a production level. The site’s Amazon Machine Images (AMIs) are regularly used in Digital Humanities workshops in a variety of countries, which suggests it’s a model that would be used if implemented at a national level.
JS: Access to New Zealand’s digital archives in Humanities is primarily web interface only. DigitalNZ offer us a world-leading metadata aggregation service, and the National Library and Ministry for Culture and Heritage offer excellent digital services too, but there are no APIs or ability to download archived data (beyond metadata) to local desktops in order to manipulate it. If the connectivity and tools existed to manipulate without download, then that might be even better for many researchers but we’re not there yet, and it’s probably best to keep requirements for both online and offline analysis in mind. The MEANDRE workbench would allow us to work with archives remotely, if we could negotiate access and connection. I understand that DIA (Dept. of Internal Affairs) and the National Archives are developing strategies to offer research services directly. A key blind-spot though is that, when asked about eResearch, humanities researchers think automatically of Wellington and the National Archives before they think of NeSI. If there’s one thing that really stands out with requirements for humanities eResearch, it’s the need for a tight coupling between NeSI, universities, and central government services.
JS: I just started as an Advisor to an Otago University project that aims to digitise the papers of Samuel Marsden and make them available as digital objects online. The key question the project team has is “what would be required for this to be considered a high quality, reputable digital output?” – an indication that we’re still defining what quality means in the Digital Humanities. Rather than being viewed (entirely) as a lack of maturity), we need to accept that we’re dealing with a non-trivial problem: porting 2000 years of humanities research, craft, and conversation into 21st century digital infrastructure and formats. The work is anything from exciting to daunting to impossible depending on who you talk to; to me it’s exciting, but more than that, essential.
JS: Perhaps one way of approaching the issue of quality would be to ask what we can expect humanities students to be capable of in 2020. If we have a mature Digital Humanities infrastructure in 2020 students would be equipped (in terms of tools, theories, methods, infrastructure) to produce or assess Digital Humanities objects and publications when studying at undergraduate level. The University of Canterbury’s Digital Humanities Programme – New Zealand’s first – is currently only offered at Honours level. This means our Humanities students get to post-graduate stage with great ideas for Digital Humanities projects, but are not capable of understanding the technology constraints or resources required to achieve their goals. There are many ways to combine humanities research with digital design and development – such as data modelling one of Homer’s epics to understand both the technical and humanistic issues involved – but the government needs to ‘bake it’ into our core curriculum and incentivize teachers and researchers to develop our capability.
JS: Humanities, and tertiary education in general, needs to prepare for the digital generation that is currently emerging from our technology enabled primary and secondary schools and enrolling in universities that still rely primarily on white boards, PowerPoint and Learning Management Systems. By 2020, humanities students will already have a feel for technology when they arrive at university, because 10 year olds now already have ubiquitous visualisation and storage surrounding them. We’re moving down a path where students will learn how to code software along with reading, writing and arithmetic. Combined with the work being done by software vendors, online services, cultural heritage institutions and digital humanists to digitize the world’s knowledge, this offers fantastic opportunities.
JS: New Zealand needs to remember, though, that we’re living in an increasingly globalized world. That offers excellent opportunities – and I for one am optimistic about the opportunities digital technologies offer to engage with the wider world – but we can’t rely on the likes of Google, or even the Internet Archive or the United Nations, to preserve and develop our culture. Google Books have digitized massive amounts of knowledge, and a lot of it is accessible through the HathiTrust Library, but copyright issues mean the vast majority of it isn’t available to New Zealand scholars. A huge amount of New Zealand content is very unlikely to be digitized by these companies, either, simply because there’s a small market for it – and if we aren’t going to look after our digital heritage, we can hardly expect commercial entities to. Our cultural heritage institutions do an excellent job on presumably meagre budgets, but universities and researchers need to step up too, and the government needs to understand the massive hole in our country’s digital strategy.
JS: The myth to bust is that Digital Humanities is just about using tools; we actually encourage students and researchers to get their hands dirty with code. A classic example for gaining even a rudimentary knowledge of how digital culture is engineered is an analysis of Facebook. Facebook is just a cultural object, but to critique it you need to understand how it works. If you don’t understand the engineering and commercial model that drives Facebook, you can’t offer an informed critique of its cultural impact. In a slightly different vein, we should remember that development of digital tools and products is expensive, and computer scientists and vendors have different priorities to humanists. The international Digital Humanities community learnt a long time ago that – although collaboration across to engineering disciplines is a crucial part of the mix – unless we learn how to do this ourselves, digital products that work for the humanities community just won’t be built. It might sound strange to people used to assuming humanities scholars spend a lot of time pondering the large issues of life, but digital humanists value concepts like DIY, building, experimentation, and rapid development. That’s the way we add value to the research community, but we need resources to do our jobs just like anyone else.
JS: Another challenge is in developing usable social media datasets that are offline, historic, private and reliable. Unlike going to the library or archive and looking at historical culture in newspapers, magazines or photographs, we can’t go back and get data from Facebook and Twitter – their data is proprietary and their users control the content. If we don’t find an approach or tools to address this, then we are looking at a cultural black hole appearing in our published societal record. Some progress is being made – Twitter is now being captured in the Library of Congress – but the volume of data is already proving hard to handle.
JS: The ‘big data’ issue presents significant engineering as well as humanistic problems, and offers an excellent example of a problem domain where inter-disciplinary work will be important. A good way to explain the problem is to note that if you wanted to review and analyse the presidency of Lyndon B. Johnson you would need to work your way through 40,000 official White House memos. If you wanted to conduct the same review of the presidency of William J. Clinton you would be faced with 40,000,000 emails. That means that future humanists are going to need the skills to analyse very large datasets (and that we’ll need the infrastructure to support and provide access to those datasets). This isn’t ‘clean’ data, either. Humanities data is messy and often outstrips the capabilities of current natural language processing and machine learning tools. Our very large humanities datasets provide interesting use cases for cutting-edge data and text analysis, not to mention OCR, data modelling, and semantic web technologies.
JS: When it comes to eResearch, the humanities and social sciences are on a very steep learning curve, and we are not well positioned in NZ from a policy perspective. I believe we need to create greater urgency in government and institutions, with the aim of putting Digital Humanities on the same page as the STEM subjects in terms of policy and strategy. We could perhaps also remember that the traditional levels of underinvestment in humanities eResearch, combined with the fact that it is likely to be cheaper to fund than STEM, mean the value-add will be significant. We’re not aiming to squeeze the last 20% out of the funding $ in Digital Humanities, we’re looking for the first or perhaps second 20%.
JS: From a sovereignty perspective, humanities is informed by social, political, historical and cultural critique. This means humanities scholars are likely to be very uncomfortable with cultural data being stored offshore. We like to have local infrastructure. In fact, if presented with a choice of having access to Google and having locally stored data, I suspect many humanities scholars would forego Google. On balance, when faced with the enormous gains in quality and efficiency you get from the likes of Google, a compromise could well be reached, but it’s important the entire humanities community is involved in the conversation. Humanities typically takes a 2000 year view of the world, while technology shifts every 6 months. What many see as humanities’ lack of engagement is actually a “watchfulness”. We care about the very long term, where 100 years is only just “steady state” when considering data. Assuming our future scholars are well educated about the affordances as well as constraints of our technologies, I think that’s a valuable perspective to bring to the table. This is true in terms of Creative Commons and GPL licensing, open access and open source, too, which our community is strongly in favour of.
JS: There’s no doubt that central government has a huge role to play. As a discipline, we in Digital Humanities need govt. backing and govt. direction. We also need government to help us set KPIs and quality standards – the investment case and rationale for Digital Humanities, including the prospective returns to NZ, need greater focus if we are to progress. Naturally, there needs to be investment in capability and education as well as physical infrastructure. There’s no point having one and not the other. In the humanities, given the early stage we’re at, it actually makes more sense to invest in education and capability as much as technical infrastructure.
JS: A key challenge for us is establishing the economic justification for Digital Humanities investment at a national level. We need to consider the links or drivers between the humanities and culture, to jobs, economic prosperity and social outcomes. We know that Arts graduates are well sought after for sophisticated management and junior consulting roles. Perhaps better understanding the gap in this area of the workforce and economy may provide pointers for greater emphasis on advancing our intellectual capability in Digital Humanities.