Local and International Scalability in DH

In my last post I wrote about the unconference session on Multilingualism in DH that took place during DHSI. This week Martin Grandjean published some really interesting, though as he calls them “unhappy”, results regarding rates of acceptance according to language in DH2014. His findings show that even though the call for papers for DH2014 was published in 23 languages, submissions were done in only 6 languages other than English. Even more shocking is that acceptance rate was about 50% lower for non-English language proposals than it was for English language ones.

The questions related to this that I launched after the unconference session were whether exchanges among different linguistic DH communities are taking place? And whether we are speaking distinct DH dialects depending on audience and sources? Grandjean’s conclusions, meant to be provocative, touch some of the issues I intuited before and I’d like to comment on them. One of his conclusions reads:

5. Ce n’est pas un problème de langue, mais de réseau de cooptation.

Dans certains colloques, il apparaît de manière évidente que le choix des keynote speakers et d’une partie des panelistes découle plus du réseau personnel des membres du comité que de choix scientifiques. En l’occurrence, la composition très anglophone du comité expliquerait pourquoi les chercheurs retenus le sont aussi en grande partie, puisque faisant partie d’un cercle plus ou moins large autour des décideurs.

DH2014, as the biggest and most important international conference in the field, carries a larger responsibility to be inclusive of practitioners around the world; while smaller, local or linguistically focused ones, though open to other communities, usually concentrate on promoting the field in their language and their context. Two very different objectives to have in a conference. Still, it would seem that the pattern is not exclusive to DH2014. Multilingual representation in the recent 2EHD in Mexico City, as can be seen in the chart below, is not that different to Grandjean’s results. Idiomas1y2EncuentrosEven when the call for papers was available in 6 languages and GO::DH contributed to the organization, only 16% of the presentation were done in a language other than Spanish—up from only 4% in the first EHD [1]. Furthermore, only English and Portuguese were represented. This parallelism suggests precisely that though efforts are made to attract a stronger multilingual presence to DH conferences in widely distinct contexts and scales, it hardly happens; at least not to the extent that one would wish. Happily, from the 2EHD data, it also appears to be increasing. We could then actually say that different conversations around the global DH’s don’t cross, but do touch briefly at times.

This leads me to the second issue. For Grandjean:

6. Quand on est bon, on choisit l’anglais.

Si le chercheur sait ou croit savoir que sa communication est de bonne qualité, il pourrait souhaiter la rendre accessible au plus grand nombre, raison pour laquelle les papiers soumis en anglais (par des non-anglophones) seraient en moyenne de meilleure tenue que les autres. Par ailleurs, le fait d’être capable de communiquer en anglais est, pour un chercheur non-anglophone, un indicateur de ses compétences et de son “internationalisme”.

This is partly what I referred to when I talked about audiences and sources. What I think the similarity in results suggests is that we definitely have different conversations in distinct DH contexts, and that language choice is dependent on that. At least from the surface it seems like a good thing, that the field can be scalable in that way: projects have a certain relevance in a given context and a different kind of relevance in another. I’m going to venture and say it is pretty safe to assume that DH2014 presentations in English by non-native speakers have been (or will be) presented in their original language in other local conferences.

The implication of this is that in its original language a particular project or research has its outlets. In that sense, yes, the purpose of taking projects to DH2014 is to internationalize them, to let others outside of their particular local contexts know about them. Indeed, this is proof that most of us have adopted English as the lingua franca in DH, but it seems to me that it can also be suggestive of how language specific conversations are taking place elsewhere. Ultimately, it is indicative of the importance of local conferences to foster certain conversations, and the relevance of international ones to promote other ones.

This still poses another problem, is the audience attending these “translated” presentations diverse or are they the same who have listened to/read them in “the original”? If non-English speakers are putting the effort to internationalize their work, it is because they (we) care about having a diverse audience and different conversations about their projects, much more than having the call for papers available in tens of languages (which, though immensely valuable, should not be an end in itself). Again, the unresolved issue is the matter of sources, conversation exchanges, and scholarship diversity.

Great efforts put towards multilingualism in DH, like the GO::DH “I whisper_______” pins coming to DH2014 are aimed at inviting everybody to come and speak in their own language at international conferences, and foster collegiality and helpfulness among colleagues. There is, still another way of contributing to this, going and listening to the diverse, though small, non-English presentations (there will be someone who can whisper in English too, I’m sure) and generate exchanges and conversations. In my opinion it will be this kind of scalability of the field in different contexts, that will help remedy the “unhappy” results found by Grandjean regarding acceptance rates and multilingual representativity.

[1] More 2EHD metrics are coming soon at the RedHD blog.

Multilingualism in DH. Notes from the DHSI2014 Unconference Session

A few days ago during DHSI2014 I organized an unconference session to discuss matters relating to multilingualism in DH. If you’ve been following my work in the last few months, the issue of knitting an exchange network of DH scholarship in various languages has been on my mind. The most important initiative I’d led until now was during DayofDH 2014 #trDH (short for translate DH) which gathered the efforts of about a dozen of colleagues from Mexico, Spain, USA, and Italy to translate important, foundational (some of them definitional) texts written originally in Spanish. At the time most of the translations were done into English, with a few exceptions. Pressing questions emerging out of #trDH were whether translations into English really ensured a wider audience for these texts, whether it was a concession to a field widely dominated by the English language, and perhaps more significant whether it would reach colleagues who–like many of the #trDH participants–speak English as a second language. Simply put, whether #trDH was promoting a more global scholarship in the long term or not.

It was with those questions in mind that I proposed the DHSI unconference session. I was hoping to get input from colleagues in other DH linguistic communities, as I sometimes am heavily biased to the Spanish speaking one. As a launching pad I proposed two broad lines of discussion: 1) Issues of standards and tools that require various kinds of (cultural & linguistic) translation/adaptation: TEI, RDF, etc. and 2) Issues of networking and facilitating exchanges with scholarship produced in many different languages mainly through translation.

Although the majority of the 50 minutes were spent on the second issue, we all agreed that  work arounds were possible with tools and standards as well as highly productive, even when constraints and limitations are frequently faced. However, the prospect of translating those seemed hugely daunting if not outright unrealistic. The key then, one of the participants added, was using them reflectively and critically in order to incorporate diverse ways of knowledge construction, a win-win in practice and theory. Similarly, someone else asked to be reflective of the kinds of collaborations (international, perhaps) that ensure that standards are adaptable (and fruitfully adapted), compatible, and, ultimately, enriched from their various uses. This certainly requires its own unconference session and more, so I will leave it there.

In matters of translation, some of the biggest concerns we all shared were whether language choice was determined by who our intended audience is, and the dangers of fragmented conversations because of this. In other words, do we speak distinct DH “dialects” depending on audience? Depending on our sources? Are we having different conversations that don’t necessarily cross in different DH communities because of this? It would seem so. Then, assuming that translation can be a way to remedy this and really has the potential to open up channels of communication around distinct linguistic DH communities, the issue that must be tackled is viability and sustainability considering how hard and resource expensive translation is.

Some of us acknowledged a position of “language mediator”–a personal concern of mine– and what it means to translate not just in terms of language, but especially and, in my opinion, dangerously, in terms of context and what one can be exposed to depending on where we’re based, who we meet, who we read, etc. in contrast to others. Another issue with this position, we discussed, is the individual dimension these endeavors usually take, and thus their instability and difficulty to become more institutionalized and active practices in our organizations. Steering away from the individual figure, we moved onto the community’s responsibility/commitment to do it. On the one hand, it seems like a great idea to make the translation of DH texts in various languages a community endeavor. Good examples of this are again #trDH, and the GO::DH essay contest. Similarly, as Alex Gil mentioned, also at DHSI2014, at GO::DH “we open language to the community, where a translation of the website or any forum post depends on the community itself.” These practices really become a community building exercise and show the priceless willingness and availability of many. Nevertheless, I worry this kind of practice might continue to reinforce networks already in place, instead of reaching out to those outside of them. A token of this is how today, Kevin Baumer on behalf of the DH2014 organizing committee reported vía email that there will be about 600 participants in the conference, while a couple of weeks ago the GO::DH list stood at 250 (I’m sure that’s grown since June 6th). In other words, not everybody is/can/must be invested in this.

Reflecting on the hard work translation means and how impossible it sometimes seems to be able to translate our work and others’, the discussion moved on to the work of professional translators, taking the matter away from the academic community itself. Though lots of academics do translate occasionally, this goes without saying, for most of us, translation is not our primary activity. In that sense, we should be fair to the people whose work is to translate and make a living out of it and perhaps adopt a really institutional practice to translation in a model akin to EU, UN, etc. in order, as Alex also put it to “place the burden on the hegemon.” Of course, the issue then would be to hire translators and raise the funds to pay them fairly–which would also need to happen from the top down whether within our organizations or our journals, as we couldn’t expect translation expenses to be footed by authors.

In conclusion, two understatements that must be kept in mind: 1) There are still more questions hanging in the air than (even provisional) answers. We have work to do here. 2) The issue of fostering a multilingual global Digital Humanities community is no easy task. Again: We have work to do here.

But also: through discussions such as this, it is increasingly becoming more evident that, first of all and as it should be expected, particular expertises are needed. Likewise, on top of community endeavors and willingness, we should be moving towards a more systematic approach in order to learn about the broader field and expand among other things, one of the pillars of scholarship: our set of sources, references and bibliographies.

DH is hard work. It’s insightful. That’s why we like it

On sunday night, after reading multiple reactions to Stephen Marche’s Literature is not Data: Against Digital Humanities I happened to be browsing my narratology basics and somewhere in Shlomith Rimmon-Kenan’s gloss over the concept of reader I found this: “At one extreme the concept is of a real reader, whether a specific individual or the collective readership of a period. At the other, it is a theoretical construct, implied or encoded in the text, representing the integration of data and the interpretative process ‘invited’ by the text” (my emphasis). I was immediately struck by the choice of words, I kept on reading and found numerous instances of this. Shlomith-Kenan’s Narrative Fiction was published in 1983, almost thirty years ago, and almost twenty years before Marche’s alleged kidnap of literature by – what he hurtfully implies are – data fascists. Regardless of time periods, since the origins of DH go farther back than the early 1980’s, it is very interesting to see that narratology, hauling its formalist, structuralist and phenomenological baggage, aimed at a systematization of literature that actually takes its components as data in which meaning is encoded – something that sounds quite familiar to many doing DH. The first question that popped into my mind was: in what way is narratological systematization different to the processes TEI, stylometrics or graph database creation require? If we look at the particulars, of course, everything will seem different: it’s not the same to talk about the heterodiegetic narrator, collocations, subaltern subject or directed edges. But the process of organizing (and by this I mean understanding) literary texts according to theoretical notions cannot be that different from one methodology to another. In principle, I don’t think it is at all different.

Before I continue I will allow myself a couple generalizations. Many of us trained in literary studies with a strong close-reading base assume the possibility of multiple readings of any given text under the condition that those readings be grounded on solidly built interpretive and contextual notions. Similarly, many of us were trained to assume the fact that nobody doing any kind of study could possibly explore and account for every minute aspect of a literary text. As complex as it might be, any study is, at least to a certain extent, an organization – a trimming – of, what Marche would term, the messiness and incompleteness, but really any topic/formal aspect of literature. When I started gearing towards digital humanities, a few years ago, it dawned on my rather quickly that the change in methodologies and research tools could not do without those same solidly built interpretive and contextual notions if I meant to say something insightful. If I meant to say something I would like to say about literature. It was also pretty clear that the issue of large scale could not be equated with an exhausting of texts, literary traditions, etc. As many different emergent practices, it seems to me that DHers have been wise to carry along at least those two fundamental assumptions.

This is where Marche’s article could use some documenting. To start, there is a deep contradiction at the bottom of his argument. The idea that mushiness and incompleteness is a quality of literature, even part of its ‘sacred essence’ it would seem – and I’m assuming also of non DH literary studies – runs throughout the article. DH, however is not only unable to account for that incompleteness but also cannot afford to be mushy “beneath the hard equations”. Marche’s view is, to say the least, uninformed and over generalizing. UPDATE: Ernesto Priego points out, Marche is not an academic and this is certainly (un)informing his approach not just to DH, but to literary studies in general, it seems to me.  As Priego himself has pointed out this morning. It turns out Marche does have a PhD in early modern drama from UofT. This makes it all the more unexplainable why he seems to have such a partial idea of the ways in which literary studies are carried out.

Marche’s assumption that “literature cannot meaningfully be treated as data” rests on a sacralization of literature that seems to ask scholars to be guards and prophets of it, not researchers. Marche’s sentence can usefully be restructured as “data in literature is treated meaningfully”. This is what we do when we read: make sense of codified signs on a page/screen. This does not just underly DH projects, but literary studies in general. Even going way back to Wolfgang Iser, this is the principle of every single act of reading as well.

Coming back to the (re)discovery that led to this, Rimmon-Kennan, in her 2002 afterword to Narrative Fiction recounts how over the two decades since her book was first published, even the strictest narratological assumptions became a means to further interpret the semantics of literary works. What I see in this analogous situation between narratology (which was also subject to wide criticism because of its systematic notions) and DH is precisely the idea that literature is data subject to/inviting interpretation being repeated. The process underlying new ways of organizing literary texts and describing them in terms of “data” in DH – or any other for that matter – is not really different to the ones happening underneath other critical schools. What Marche criticizes as “the mushiness of the words beneath the hard equations” is, in fact, the same mushiness of literary meaning being accounted for especially if discussing terms such as influence. DH needs and wants that from ‘traditional’ literary studies, and hasn’t shied away from it. That’s why it’s hard work. That’s why it’s insightful. That’s why we like it.