Should a digital humanist be able to code?

authored by Frank Lynam at 16/04/2012 18:08:11

I came upon a thought-provoking article written a few years ago by Matthew Kirschenbaum that considers whether the digital humanities scholar should be able to understand and write computer code. The author is an associate professor of English at the University of Maryland and so is deeply embedded in and has the authority to speak about the topic of humanities scholarship. The article recalls that Kirschenbaum took up coding as a youngster when his parents first brought the Apple IIe into the family home. He takes the position that in order for the student of humanities to fully engage with the creations of the digital world they need to be cognizant of the structured thinking that goes into their production. In saying this, his argument is not to suggest that humanities departments start training and spiting out computer programmers.

For Kirschenbaum the act of writing computer programs demands a certain type of structured thinking that is analogous to the discipline imposed on the poet of the Shakespearean sonnet or the author of the detective novel. These two content creators are each constrained by the rules of their chosen discourse. The Shakespearean sonnet writer needs to maintain a very particular rhyming structure and the detective novelist must include in her plot a crime, a victim, a hero, a villain and so forth. The programmer in the same way needs to conform to the conventions of her programing language.

Let us for a moment consider the skillset demanded of the scholar of the Shakespearean sonnet or that of the detective genre. The student of the Shakespearean sonnet needs to fully familiarise themselves with the conventions of that form and the detective novel student must do the same for their type. Might we then infer a similar relationship between the student of the digital form and the method used by the digital producer?

I have long mused over this question and was glad when I came upon Kirschenbaum’s article as it made me pause and explicitly set down exactly what it is that I thought was demanded of the digital scholar in the context of coding literacy. My opinion had previously been that in order to be a true digital humanities scholar one needed to be able to create digital content but also to be able to produce digital tools. I had rationalised the latter on the basis that the alternative model in which digital tools were produced by technologists was ultimately limited by the tool-creator’s understanding of the subject matter, which could in many instances be highly limited. If you decide that the best way of circumventing this problem is for the digital humanist to become the producer of their own digital methodologies, then it would appear to be self-evident that the digital humanist needs to be proficient in the language of these creations, i.e. code.

Kirschenbaum’s article seems to agree with this thesis but it goes further in the sense that it suggests that even if you are not producing digital tools yourself, you would still need to be at least familiar with the conventions of code in order to be able to understand the ways in which digital content is created. Kirschenbaum highlights correctly that code is constructed from and a manifestation of an ontology or a model. The praxis of writing code (of what is allowed and what is not), therefore, ultimately decides the limits of what is possible within the confines of these models. Therefore, in order for the digital humanist to fully understand these underlying frameworks (and this must ultimately be seen as the goal of any scholar), he or she must be code-literate.


I think the phrase 'code is constructed from and a manifestation of an ontology' is unclear. If you mean a programmer needs to be aware of the theoretical limitations of computers in terms of Turing machines etc., then this might be setting the bar a bit high. In practical terms, the limitations are time and money. The argument that a digital humanist needs to 'understand these underlying frameworks' is not clear. I can understand and analyse a poem without being able to write one myself - at least a good one.

I think a better argument is that code creation will lead scholarship to a place where traditional methods cannot go. Without the prospect of completely new areas of discovery, most humanists (and probably a large proportion of digital humanists) will never see coding as fundamental to their academic prospects.

(Henry, 16/04/2012 21:25:18)

I just read a blog post by Geoffrey Rockwell where he discusses what it means to be a digital humanist. In it he offers a mixed inter-disciplinary/disciplinary model for the field in which the digital humanities exist as a common ground of discussion between all the other disciplines (computing, humanities, etc.) and humanities computing is a discipline in itself that has a specialised skill-set associated with it and whose agenda is to develop tools for the digital humanities space. I like this and my ‘all DHers have to be coders or at least procedurally literate’ would be modified in this model to be ‘all specialists operating within humanities computing have to be coders or at least procedurally literate’.

Take your example of the poem. You say that you can understand and analyse the poem without being able to write one yourself. I would argue that you can’t do this. To understand and analyse the poem you really need to be able to understand how the poem was constructed in the first place. Otherwise, you are just looking at a black box and can’t say anything about the processes involved in its construction. I think you’re taking for granted a lot of knowledge that has sunk into your subconscious that you use when reading a poem. Most people don’t gain access to that sort of knowledge in the context of digital content. They tend to never see the code, the process.

My point about ontologies was that the constraints of coding syntax limit the sorts of things or models that code can produce. In order to be able to understand the resulting phenomena of code you would be much better equipped if you understood these constraints.

I do agree with you though that being able to code does open up new areas. This may not be its only benefit though.

(Frank Lynam, 16/04/2012 22:53:13)