Digital Humanities: Is it Research or is it Service?

The notion of “digital humanities” has come to be used for several different things at the same time, which makes it inevitably vague. An unwelcome consequence is that when you ask for “digital humanities”, you cannot be sure what you get.

Here is my attempt at clarifying the notion of “digital humanities”. When talking about “digital humanities” or “digital-humanities research”, one or more of four different meanings can be intended:

  1. Digital humanities as a research service.
  2. Digital humanities as a research method.
  3. Research on digital services for the humanities.
  4. Research on digital methods in the humanities.

The different meanings of “Digital Humanities” explained

  1. Research service: Digital humanities as a service means the provision and, eventually, development of software tools, data processing, online presentation etc. for research in the humanities. Typical technologies are relational databases, XML-Editors and -processing tools and digital publication technology.

    Here, the technology is not part of the research process but a supporting infrastructure for doing research (a bit like the test tubes, notebooks and other helpful things in a laboratory). The research itself will proceed largely by “conventional” means, say, hermeneutic analysis.

  2. Research method: Digital humanities as a methodology means the application of technology, say, computer linguistics, artificial intelligence and the like for analyzing and investigating the research object. Typical technologies in this category include network analysis and distant reading.

    Here, digital humanities is a part of the research process itself. Its technologies are employed as methods of analysis and crucially shape the potential results of that research. However, the kind of research done is research on topics of the humanities, employing the technological methods as a means. It is not research on digital-humanities technology. The discussion and justification of the technology used will only be a side-topic of the research papers published.

    In some areas of the humanities, for example sociology and in particular economics, the application of technical means (i.e. mathematical or computational methods) of analysis have a long standing tradition and existed before “digital humanities” became a buzzword.

  3. Research on digital services: Research on service technology is research on the particular technologies that are used as a service (as in point 1) for research in the humanities.

    Typical examples would be the development and discussion of tag sets or XML schemata for encoding texts or data in the humanities, and the development of standards. Being a kind of research activity, the theoretical discussion of service technologies and the writing of research papers make sense in this area.

    Strictly speaking, though, this activity is not humanities any more but applied computer science related to the humanities.

  4. Research on digital methods: Research on digital methods of analysis is research on the digital methods which are part of the research itself (as in point 2).

    For example, research on algorithms for automatic or semi-automatic lemmatization of text-corpora would fall into this category. Again, it makes sense to have theoretical discussion and research papers in this area.

    This, in contrast to the previous category, could be considered as humanities, because the development of and critical reflection of research methods that (inevitably) shape the understanding of the research object, should be considered as an essential part of the science studying this very object.

One can – for various good reasons – argue that this distinction is not sharp and that the borderlines between the different categories are unavoidably blurred. However, that the transition between the different categories is continuous does not mean that the distinction does not exist. The distinction as such is, hopefully, clear enough.

I would not call a research project in the humanities that also makes use of digital methods a “digital humanities” project but rather a research programme that has a digital or digital humanities part or component. And it is only this particular component that (in virtue of the second point, above), I call “digital humanities”.

Research or Service?

In virtue of the distinction above, it seems clear to me that only the second, third and fourth category can be called “research” in a strict sense, while the first category cannot. For research is an activity that is directed at gaining knowledge about or a better understanding of a particular object of research. Typically, this involves the publication of papers (in peer-reviewed journals) or books, the discussion with fellow researchers and the like.

Service, on the contrary, is not about gaining knowledge about or insight into reality, but it is about making things work. And while the discussion with peers is certainly valuable, its goal is not to publish research papers, but to solve practical problems and to build and sustain reliable digital systems (say, a database or web-application) and services (like data preparation). In the following I will mostly concentrate on the distinction between the first and second category and I will, for the sake of simplicity, distinguish only between digital humanities consisting of services (or means of research, although this expression can more easily be misunderstood) and digital humanities as an umbrella name for digital research methods in the sense of the second meaning above.

When and why positivism disputes are important

The distinction is important, because digital humanities as a service and digital humanities as research require quite different criteria and discussions for their evaluation. While for “digital humanities” as a service (type 1 above) it is almost a no-brainer that they should be employed as much as possible and affordable (e.g. if you compile a catalogue or dictionary it should definitely be published online and the materials should most certainly be managed in a research database rather than a rack or bookshelf), it is in the realm of digital humanities as a research method where such questions must be asked as whether employing digital methods is appropriate or whether digital paradigms are really superior to other, more conventional research paradigms.

It is only in the latter field, where positivism disputes are meaningful and also need to be carried through. Because an excessive and, in particular, an exclusive use of digital-humanities research methods all too easily slips into a kind of positivistic fetishism of (digital) methods, where a) the (digital) methods become a criterion of relevance and b) the phenomena are ignored.

This fetishism of method can additionally be fueled by funding agencies that – in their laudable urge to promote digitalization – favor the use of digital means without distinction. Now, while it hardly makes sense any more to fund a humanities project that does not have a digital publication plan or makes no use of digital technology to manage its data, it can still be perfectly reasonable to fund a humanities research project that is based on an entirely non-digital methodology.

For many if not most research questions in the humanities, there are no digital methods available that could rival conventional methodology, though in some but not all of these cases they may be a reasonable addition to the research plan. But when funding agencies indicate that they favor the use of digital research methods, there is an imminent danger that research plans will be accommodated to what digital methods can deliver, which means the use of digital methods becomes a criterion of relevance (characteristic a) of my definition of positivistic fetishism of method above).

All too easily digital research methods are given a credit of trust that quickly evaporates upon closer inspection. (See my criticism of the digital dictionary of German language https://dhmuc.hypotheses.org/2598 for a case-study.)

Next to the imperfection of methods that rely on automatization in the humanities, there is an other, more fundamental reason why many if not most humanities projects will now an in the future rely crucially on non-digital methods, which consists in the fact that the humanities are defined by hermeneutics, where I understand hermeneutics to be the empathetic comprehension of human action as well as the products of the human mind („Hermerneutik ist einfühlende Nachvollzug menschlichen Handelns und der Schaffensprodukte des menschlichen Geistes“). Hermeneutics in this sense is not merely a method but to a large degree the very purpose of the humanities. But if that is true than their is a strict limit up to which the humanities can (or should) be permeated by digital methods.

Digital Services are more than an addition but also shape the way research is done

This said, the opposite extreme should also be avoided. The opposite extreme in this context is the believe that „conventional“ humanities projects that require digital humanities as a service only, will not be affected in any other way than by the addition of a digital component.

Quite the contrary, in order to reap the full benefits of digital services adjustments to the workflows, project management, allocation of resources, mindset and goals may become unavoidable. In other words digitization may require changes or at least adjustments of the „corporate culture“ of a research project.

Otherwise its effects might not go very far beyond those of a retro-digitization of an already finished project. The opportunities that digitization offers will remain underused and, in the worst case, there will be additional costs, because the results of a largely non-digital workflow need to be transformed with consideable effort into proper input for a digital processing pipeline.

The cultural change going along with a thoroughly planned and consistently realized digitization can affect core areas of a humanities project. The following is an incomplete list of changes that are to be expected:

  • Digital publication affects the way works are „written“ in the humanities. For example, if you publish a dictionary in form of a searchable online-database rather than a printed series of volumes, it is advisable to cast the dictionary entries into a more stylized and regular form (a „datalike“-style) than one would do in a printed dictionary where a relatively more „textlike“-style is preferable.

    This makes it easier to exploit the capabilities that database-technology offers better. A downside may be that certain details and nuances expressed in the narrative text may get lost. However, if the advantages for the user outweigh those losses, it is worth while. Another way of putting this, is to say that when publishing digitally, the very product will be different, not only its outer form.

    This is not another instance of positivisic fetishism of method, because the methods used to assemble the articels remain largely the same. For a comparison, think about the advantages of the written word over the spoken word. This also also comes at a price: Some information and many of the nuances contained in the spoken work may get lost. Still, humanists have almost exclusively relied on the written word.

  • In the long run, digitization must be realized on a zero-cost level. It is all too easy to claim that digitization generates extra costs and therefore requires extra funding. While it is reasonable to ask for extra funding for the experimentation with digital methods and for the introduction of digital methods to a project, it must be understood that resources are limited and that in the long run, digitization must be financed by shifting resources.

    For example, if you run a humanities project with a team of four humanists doing the writing, then the transformation to a digital humanities basis might result in a project where only three people will do the writing and one developer will take care of the tool-development, data encoding or online-publication. This may have the consequence that less output can be produced in the same time. But this can be justified, if the output produced can be put to better use because of its digital form.

    Also, if the projects themselves are burdened with the digitization costs, this can be an incentive to keep the digitization costs low by investing in learning and building up of digital competencies among the humanists in the project. (See the next point.)

  • Digitization requires a new mindset by the humanists working within an DH-project. In particular, it requires acquiring of new competencies from the humanists working in an DH-project as well. A very common case is that where humanist scholars need to get accustomed to working with different tools than the standard office-software they already know.

    For, office-Software still follows a paradigm where printed paper is the end-product of the work. While this allows the designers of office software to employ the highly intuitive what-you-see-is-what-you-get principle, it becomes extremely hard to enforce the production of clean and semantically explicit data with these tools – unless a further (costly) conversion step is taken that typically includes both, script-writing and tedious handwork.

  • Digitization requires more management and active coordination of a team of researchers and workers with different backgrounds. Digital humanities projects usually rely on division of labour in a stronger sense than traditional humanities projects, where one could assume all co-workers to have more or less the same academic background and mindset. However, once there are server administrators, software developers and humanists working within one project, the management ond coordination requirements strongly increase. However, if it works well, this can also be a rewarding experience.

This are but a few important points. The gist of it is that, if taken seriously, digitization is not just an addition to humanities-research-projects but a cultural change in the way humanities research projects are done. If this is understood, then digital humanities as a service will have much deeper influence on the research agenda and the research process than might superficially be assumed. None the less, it is a supporting role, not research itself.

Semantics and Career: A question of acknowledgement

There is reason to insist on the distinction between digital service and digital research methods, because in the digital-humanities community there seems to be a strong urge for gaining better recognition of digital-humanities activities by claiming that these should be acknowledged as scientific research – irrespective of what type of digital-humanities activity is actually involved. As much as I sympathize with this urge – digital services are indeed invaluable for doing science and may even be intellectually more challenging from time to time than the very research that these services support – blurring the lines between research activities that are directed at finding the truth and gaining knowledge on the one hand and digital services on the other hand can turn out to be detrimental to the very purpose of science itself (as indicated above).

Digital humanists engaged in the service branch should rather strive to seek recognition for their work on the basis of its vital importance for research processes based on the division of labour and the intrinsic technical challenges which, as said, often enough are not smaller than the challenges of the research itself. Now, if this recognition is not only a superficial one, it should be connected with career opportunities approximately as good as those in the primary research activities itself. Admittedly, this is the hard part of the endeavour. I understand if people are inclined to cut this short by relabeling digital-humanities services as a research activity, but, for the reasons stated, I do not approve of it.


Dieser Eintrag wurde veröffentlicht in Artikel von Eckhart Arnold. Setze ein Lesezeichen zum Permalink.

Über Eckhart Arnold

Eckhart Arnold ist Leiter des Referats für IT und Digital Humanities der Bayerischen Akademie der Wissenschaften. Er hat Philosophie und Politikwissenschaften studiert und unter anderem über Computersimulationen altruistischen Verhaltens und wissenschaftstheoretische Fragen der Validierung von Simulationen geforscht und publiziert.

Schreibe einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.

Diese Website verwendet Akismet, um Spam zu reduzieren. Erfahre mehr darüber, wie deine Kommentardaten verarbeitet werden.