Paolo Monella Post-doc scholarship in Digital Humanities Accademia dei Lincei, Rome 2012

Conference website (with schedule, full-text papers, slides and online comments).

The (unedited) notes I took at the conference.

Opponents:

• Willard McCarty, King's College London / University of Western Sydney, The residue of uniqueness;
• (Patrik Svensson, Umeå universitet, Sverige – not present)

Willard McCarty's answer to the question was: "Agenda before infrastructure". He argued that the Digital Humanities needs to follow a "humane project", not different from the general agenda of the Humanities. It has to find out what is unique of humans as opposed to animals and machines, and to study and foster that "residue of uniqueness".

A question like "What is the agenda of the Digital Humanities" is certainly very broad to answer – maybe too broad – and possible McCarty's approach of recalling the general aim of the Humanities was the only possible one, in front of such a general formulation.

However, a rather paradoxical implication of McCarty's argument is that the goal of the computer-assisted study of human artifacts is to study whatever is left of human creativity when everything that is computable is taken out (the human "residue" after a comparison with what also machines can do).

In other words, one could argue where the peculiar hermeneutical contribution of the Digital Humanities may lie, if any exists. This is, in fact, the side where the following discussion verged.

Opponents:

• Susan Schreibman, Trinity College Dublin, Ireland, Digital Humanities: Centres and Peripheries;
• Domenico Fiormonte, Università Roma Tre, Italia, Towards a Cultural Critique of the Digital Humanities.

The two talks of this controversy, in my opinion, both proved, one willingly, the other uwillingly, the same point: that the Digital Humanities today are strongly centred on the Anglo-american community, its agenda, its prospective.

Susan Schreibman, the editor of the admirably vast Companion to Digital Humanities, has discussed the sections of the Companion, by acknowledging that in its first edition the contributors outside the USA and the UK were too few and by announcing that a new edition is underway, where a more diverse pool of contributors will be involved. In other words, the Companion has been taken as a microcosm representative of the macrocosm of all possible approaches to the Digital Humanities.

Domenico Fiormonte started off by declaring that he studies now both Digital Humanities and Sociology of the Digital Humanities. And indeed the latter was the topic of his talk, one of the most interesting of the Conference. After arguing – loud and clear – that the "methodological question" is what differentiates the Anglo-american approach from the rest of the community, he demonstrated with bare facts that "Non Anglo-american Digital Humanities are almost invisible", that international organisations and their boards are overwhelmingly dominated by scholars from USA, Canada and UK, and that this has a deep impact in the definition and the control of standards used to encode cultural artifacts. Very interestingly, he made the example of the Unicode representation of writing, being more "Western-centric" than one would think, if compared to some Indian writings.

The somewhat lame discussion that followed such potentially heating arguments demonstrates, in my view, that this debate risks to boil down to a protest of the European scholars for being kept off the Anglo-american-led table of the Digital Humanities. The worst risk is that the whole controversy, if tackled only under this angle, will vanish when Europe will join the table.

A rather more interesting issue emerged very clearly from Fiormonte's discussion on the Unicode encoding of Sanskrit writing: oversimplified modelling by Western scholars will inevitably tend to be "Western-centric", and therefore exclusive against the cultures not participating actively in the modelling process. As a consequence, sophisticated thinking – and complex modelling – on cultural phenomena is inclusive in itself. Which implies that we have an ethical – other than a scholarly – duty to theorise in a complex and sophisticated manner.

Opponents:

• Jeremy Huggett, University of Glasgow, United Kingdom, Core or Periphery? Digital Humanities from an Archaeological Perspective;
• Jan Christoph Meister (replacing Fotis Jannidis), Digital Humanities is us, or on the unbearable lightness of a shared methodology

Jeremy Huggett, a digital archeologist, discussed the relationship between Archeology, Digital Archeology and Digital Humanities, and demonstrated that Digital Humanities today are "text-centric" and divided into sub-disciplines that do not always acknowdege each other.

Interestingly, as soon as the right questions are posed, as Manfred Thaller has done in this Conference, the glossy and irenic surface of the Digital Humanities discourse reveals invisible boundaries and latent fractures.

Jan Christoph Meister tackled the issue from a more theoretical perspective. He interestingly discussed the concepts of "discipline" and of "digital". Eventually, he concluded that not all humanistic questions can (or are worth) being tackled digitally today and that disciplines are a construct of the XIX Century doomed to vanishing, also "through the use of a new conceptual lingua franca: digital conceptualisation".

In this case, the discussion following was even more interesting than the talks themselves, as it circled (rather than tackling directly) one basic issue: what is computable in the Humanities? Is it worth it having a computer process it automatically, rather than letting humans scrutinise it traditionally?

Discussion on the publication of the Conference materials

The publication of the Conference materials, discussed at the end of the first day of the Conference, is worth a note apart. The recent volume Debates in the Digital Humanities was very successfull in sparkling a broad discussion in the form of conference talks related to the volume content, blog posts, comments to those posts and twittage.

The model chosen by Manfred Thaller to achieve a similar goal is slightly more centralised, but equally interesting. The starting point here is the conference itsel, and the pivot of the whole process is the Conference website, that even stated the Twitter hashtag, #CologneDialogues.

I myself saw a very efficient young webmaster asking all speakers during the Conference for their revised papers and their slides. These were thus uploaded to the Conference website in real time. The website is expected to shortly host all full-text papers, presentation slides and eventually even the video recording of all talks. Immediately after the conference, invited speakers and the audience will be asked to post their comments to the talks to the website, and after a period of virtual discussion online the proceedings will be printed – certainly before the next Dialogue (December 2012).

Opponents:

• Espen S. Ore, Universitetet i Oslo, Norge, Document Markup – Why? How?;
• Desmond Schmidt, Queensland University of Technology, Australia, The Role of Markup in the Digital Humanities.

While the first day of the Conference dealt with the disciplinary definition of the Digital Humanities, the second one was meant to address more specific issues, as it was entitled "Making the Digital Humanities work: Tools, infrastructures, technology and conceptual work".

Controversy 4, the first of the second day, addressed an issue dear to the theoretical reflection of many Italian digital humanitists and close to the research lines of the Centro Linceo Interdisciplinare "B. Segre": the digital modelling of text.

Espen Ore stated that even word separation in rune stone inscriptions, the obelisks of ancient Alexandrian philologists and modern typographical features like identation can be considered a form of markup. He thus argued in favour of the central role of markup in any form of text encoding – not only digital.

However, his discussion later focussed on a less interesting topic: XML/TEI inline vs. stand-off markup. Inline markup, he showed, is fully interoperable, while stand-off markup makes overlapping hierarchies easier to deal with and can be exported to a relational database, but does not provide any standard for data interchange.

It seems to me that the inline vs. stand-off markup issue, albeit quite banal in itself, exposes the limits of XML/TEI. The latter is meant to provide both a tool for "rich" markup and a highly interoperable standard. However, Espen Ore's case study of the Ibsen Letters Project showed that a really rich markup demands for external markup and relational databases, which in their turn hinders full interoperability. Ore himself stated that a possible solution is to use stand-off markup inside the project and then "Project the data into a suitable hierarchy for XML (TEI) export". In other words, there is a risk that the really "rich" markup is dealt within the project and are lost when exporting data in XML/TEI – the latter degraded to a bare exchange format. Do real implementations of XML/TEI imply a tension between "rich" markup and interoperability? If so, the primary goal of XML/TEI would be missed.

To me, the most interesting part of Ore's talk was his report of a decades-long work on an actual XML/TEI project, the Ibsen Letters. Among the interesting practicalities that Ore reported, I was struck by "fossile markup". The project started very early, while TEI P3 was the recommended standard. As the TEI guidelines evolved, older files encoded with P3 remained unchanged and kept trace of P3 markup no longer implemented in P4 and P5: tags no longer used, but not harmful – "fossile", in Ore's definition.

This apparently trivial practicality offered Willard McCarty and Tito Orlandi, during the subsequent discussion, a chance to stress the interpretive and subjective aspect of all markup. In fact, we should resist the templation to think that XML/TEI P5, or whatever the most recent form of digital philological markup is, is "the end of the history" of philology. McCarty provokingly argued that "Markup is only (maybe) useful for undisputable features, like line numbers, but not for interpretive features, as the latter are subjective". Tito Orlandi replied that all interpretations of a text (including essays of literary criticism) are a form of markup. While reading a literary essay, we are aware that we need to understand its rationale. In the same way, when we use markup modelled by a previous digital humanist we have to put some effort in understanding the rationale under their markup.

That a "markup criticism" will be increasingly part of the skills of future (digital) philologists is an obvious reflection, but it sounds revolutionary to those who would like to use technology (XML/TEI in this case) to eschew innerly humanistic issues – like that of the subjective nature of interpretation – rather than to tackle them better.

Desmond Schmidt, the second speaker of this controversy, carried on a very bright analysis of six major problems of XML/TEI embedded markup (excessive tag number, usability, overlapping hierarchies, interlinking, textual variation, interoperability). After a very convincing pars destruens, he proposed the design of a possible solution, involving automating computing of textual variation (a model called MVD, or Multi-version document) through Nmerge, and separating "markup set" files (including "Stand-off properties") from text files.

Schmidt's model is particularly interesting for those interested in textual variation, a key feature of multi-testimonial textual variation poorly represented in XML/TEI markup. Also, his MVD model allows great flexibility in mixing "markup sets" in order to overcome the typical XML/TEI "overlapping" issue and to leave researchers free to devise the most complex forms of markup even for multi-version texts. Also, it correctly separates markup and "text", thus improving interoperability.

A problem, however, lies right here, and has perfectly been pointed out by Dino Buzzetti during the discussion: what is "text"? Schmidt's data model separates files including "markup sets" from files including "text". To put it in Buzzetti's words: "My worry is that we do not have a clear idea of what text is. How do you relate features that relate to the 'physical' layer of text encoding with properties that refer to content?".

In slide 52, Schmidt asks: "Are there any principles in the methodology of dealing with markup and texts that aren’t linked to a specific technology?". One of those principles, notably not addressed by XML/TEI nor, seemingly, by Schmidt's model, is a clear distinction between "document" and "text". The "document" level, for instance, may be represented by the peculiar graphical representation of a text in a medieval manuscript (including abbreviations, punctuation, page space arrangement etc.), while the "text" level coincides with the content of the literary work that that graphical system of signs aims to represent. Without such a clear modelling distinction, that should be implemented by the data model by separating "document" files and "text" files, no serious digital critical edition of multi-testimonial textual traditions is conceivable.

Opponents:

• Sheila Anderson, King's College London, Taking the Long View: From e-Science Humanities to Humanities Digital Ecosystems;
• Joris J. van Zundert, Huygens ING, Koninklijke Nederlandse Academie van Wetenschappen, Den Haag, Nederlands, If you build it, will we come? Large scale digital infrastructures as a dead end for digital humanities.

This controversy, like the second one, dealt much about the "Sociology" of the Digital Humanities scholarly community, namely about the best model for scholarly collaboration. The tension seems to be between

1. large organisations (like DARIAH and CLARIN) meant to attract funds, coordinate efforts and reach out for 'traditional' Humanities scholars (as argued by Sheila Anderson), and
2. a more 'horizontal' collaboration model involving practitioners independently working on smaller-scale interoperable, modular and open source projects, as exemplified by the ThatCAMP meetings and the InterEdition network (in favour of which Joris van Zundert, an exponent of Interedition, argued very heartily).

During the animated discussion that followed Anderson scored two main points in favour of larger organisations like DARIAH: she demonstrated their capacity to attract funds and to reach out for 'traditional' scholars. Willard McCarty, however, soon warned against the risk to rely on a "funding argument" (not arguing that something should be funded because it's good, but that it's good because it's getting funding).

For the rest, van Zundert was very effective in demonstrating that these organisations drain much of the available funding, leaving independent researchers and isolated institutions in a difficult position, and that projects like InterEdition are achieving impressive results through their 'horizontal' collaboration model, while avoiding the risk to build large centralised infrastructures that end up being – in his own words — "huge highways that nobody uses".

Opponents:

• Helen R. Tibbo, University of North Carolina at Chapel Hill, Placing the Horse before the Cart: Conceptual and Technical Dimensions of Digital Curation;
• Henry M. Gladney, Saratoga, California, Long-Term Digital Preservation: a Digital Humanities Topic?

The two opponents of this controversy have embodied very well the distant approaches that may be involved by the terms "curation" (as a professional librarian's cultural work) and "preservation" (meant, in the most restricted sense, as a practical, technological issue).

Helen Tibbo stated that in the future all curation will be digital, that "Digital curation" may be considered a discipline of its own and that technology is one of the requirements for Digital curation, but only the last of a long list of "conceptual" skills.

Henry Gladney could not have tackled the issue from a more distant approach. He illustrated a digital model for a new generation of files (TDO, Trustworthy Digital Objects), engineered to resist falsification or any kind of information loss.

Gladney had stated that "The issue of digital preservation has already been solved from a Computer Sciences perspective". This may be true, but the following discussion showed very easily and from many angles that long-term preservation is not only a technological issue, but eminently a social and cultural one.

Opponents:

• (Edward A. Fox, Virginia Tech, Blacksburg, Virginia – not present);
• Hans-Christoph Hobohm, Fachhochschule Potsdam, Deutschland, Do digital libraries generate knowledge?.

Hans-Christoph Hobohm, in the closing talk of the Conference, said that digital libraries are today mostly 'static'. They deliver information to a 'user' in the terms of Shannon-Weaver's linear model of communication. Hobohm's question is: "How can we make the information stored in (digital) libraries become knowledge?". His answer lies in the concept of "blended library": a phisical place where users interact with digital objects by using their body. This model takes into account the fact that cognitive processes happen through our body.

I think that it may be said that this Conference was probably better devised by Manfred Thaller than realised by its speakers.

Some central, methodological issues (like the relationship between Humanities and Computer Science, or the idea of text lying behind our digital modelling) have been rather alluded to than fully tackled, while much of the discussion has dealt with the organisation of the Digital Humanities scholarly community.

This is understandable, as the Digital Humanities are undergoing a phase of "institutionalisation" that makes the disciplinary and organisational issues urgent. However, it may be said that the theoretical complexity of many of Thaller's questions has sometimes been eschewed.

Also, it was very difficult for speakers to organise their talks "dialogically", that is to fit their arguments into the very sophisticated dialogic structure of the Conference.

Nevertheless, many talks included stimulating reflections. Among the others, one could mention the interesting discussions on inclusiveness by Fiormonte, on disciplinarity by Meister, on textual variance and markup by Schmidt, on 'horizontal' collaboration by van Zundert.

As it is often the case, the discussions were often at least as interesting than the talks. But in this case this did not happen by accident, but as a result of the "dialogic" organisation of the meeting by Manfred Thaller, who allowed much time for debate and provided participants with a further occasion to discuss the topics by commenting on the papers online.