robin_anne_reid: (Treehouse)
robin_anne_reid ([personal profile] robin_anne_reid) wrote2013-01-03 04:16 pm

MLA 2013 Session 22: Building Bridges, Expanding Access

Notes from MLA Sessions on or related to Digital Humanities at 2013 MLA.

Disclaimer: while I've done spellchecking and basic corrections, these are very raw. There are sentence fragments; there can be some slippage from first person to third person (I type while listening!) for the speaker, and there are some terms/words I did not catch or do not know how to spell. But I've learned from past situations, if I try to polish, these notes never get pubished.

So, they're rough.

22. Expanding Access: Building Bridges within Digital Humanities
Thursday, 3 January, 12:00 noon–1:15 p.m., 205, Hynes
Presiding: Trent M. Kays, Univ. of Minnesota, Twin Cities; Lee Skallerup Bessette, Morehead State Univ.

Speakers: Marc Fortin, Queen’s Univ.; Alexander Gil, Univ. of Virginia; Brian Larson, Univ. of Minnesota, Twin Cities; Sophie Marcotte, Concordia Univ.; Ernesto Priego, London, England

Digital humanities are often seen to be a monolith, as shown in recent publications that focus almost exclusively on the United States and English-language projects. This roundtable will bring together digital humanities scholars from seemingly disparate disciplines to show how bridges can be built among languages, cultures, and geographic regions in and through digital humanities.

* * * *

I chose this section for the focus on more languages than English, and the multidisciplinary perspective. When I got here, I was thrilled to see a table in the back for bloggers and tweeters: though I still have that LJ knee jerk response that I am not a blogger! And I shall never be a tweeter! (MLA has its own hashtag though, for those of you interested--in fact, each session has its own hashtag--MLA2013 plus session number--in this case #22). But since I have my netbook and am writing up notes to post on Dreamwidth, and got here early enough that I was the first (there are only six seats, and a power bar yay), I decided to claim my space. So far two others have joined me (I'm the only woman so far). The only bad part so far is we're at the back, near open door, which means cold draft! *settles warm scarf around neck and adjusts fingerless gloves (which are proving incredibly helpful in ongoing cold hands problem*

Larson has a great handout--which includes the link to an electronic copy at a tiny url: "Expanding Access: Building Bridged" remarks and his references:

The presider noted that there would be a link to her visuals at the hashtag for the session (Storify?).

I gather that at least two of the session haven't showed up--probably one is the one from England (*shudders at cost of that ticket these days*).

On the elevator ride, I saw one of my aca-fen friends--I was going up, she was going down. I hope we can connect later (offline OR online OR both).

This is a roundtable--so far five presenters here. People coming in--five now at the blogging table (three women); fifteen or so in the audience, and about three minutes left until noon strikes --now, thirty plus, and others coming in.

SESSION NOTES: Some people did not give their names, so I'm sort of guessing in a few places.

Lee S. (readingwriting on twitter); announces #mla13

Presentation is online on Storify: search for readywriting, and find a link to this particular presentation (sources and resources). Organizer of this panel; wanted to bring group together because DH panels do tend to get sequestered along still traditional disciplinary and linguistic delineations, esp. in modern languages. There's a panel at 1:45 for digital francophones (French, and typically would not appeal to the entire group of digital humanists). Different people different academics from different disciplinary backgrounds, Big Tent DH. Linguistics, Hispanic, Francophone, comparative literature in translation. Can we get the group in the same room, and hopefully people who do that same thing in the same room. Expand the conversation--building bridges is the title--see how we do work, how we're using tools.

So I've linked to the actual proposal. One of the things who struck me is how Anglo-American the center of DH has come-=not true for people coming from Europe and other places--the center of DH has become English and American/north American. I think the best illustration of that could be the famous or infamous article "What is Digital Humanities and what is it doing in English department," appeared in ADE. It's now included in the recent publication debates on DH, and I think that DH has become houses in different departments (English, history), and historians have different perspective, but again, it has seemed to become something that is very focused on the language English and English departments.

Problems with that status quo: one of the dangers of being limited to any disciplinary department--it can limit the reach of the tools.

Visualization software: Voyant Tools. Popular tool, upload texts, multiple languages. Asked if the program allowed comparing texts n different languages--nobody had ever asked for it before. The single language per tool is not appropriate (comparative literature, translation studies). Accessibility of tools in multiple languages, and between languages, can be very powerful in understanding of literature, and translations. How do we work in DG on translations -- changes over historical periods, and geographical spaces ("world author" and how does translation impact that, and different translations, impact that).

Many of these tools are being developed in English in America with English professors in mind, and Ernesto P. writes about Globalization" (cannot make pane): it's not just about openness, but if people in developing nations cannot access the material if it's not in their language, it doesn't help. Need reliable multilingual metadata. Free/open access not solution to lack of translations. How do we translate DH?

Google Translate: this dream of the machine that can translate. L. worked for Nortel, and their dream was to develop translation software, and in the process of producing their very thin dictionary (approved usages). Can we get a machine to translate accurately--humanists need to be skeptical about this, so how do humanists and linguistics and people who specialize in different languages groups, how do we work together to make DH accessible to the most amount of people regardless of what language(s) they speak?

This question hasn't been talked about enough yet. One great example: a place that is doing just that. The CulturePlex at the University of Western Ontario (Canada). Brainchild of modern language department. Doing great stuff including the Silver Product, database management system, and all of their stuff is available bilingually, currently English and French, with goal to be available in Spanish as the other official language of Canada. Faculty and grad students working to adapt, connect, and create new tools, and the important connections with DH communities in Mexico and Latin America.

Always a danger in this: comparatists aware of limitations that are placed on us when we stray outside disciplinary boundaries--MLA Job Information List broken down by English and Foreign Language departments, traditionally how we understand disciplines is embedded in the Job Information List. Pragmatic about the realities of trying to get a job vs. utopianism of dream of working in multiple languages.

She closes on a downer note: it's important to say that these bridges being built come with a price--the bridges lead nowhere unless skills are valued by departments, hiring committees, and academia.

NEXT: Build on Lee's idea that we need to break language barriers and disciplinary barriers, and geographical boundaries. Been working in Caribbean, trying to bring some of the exciting DH work here, there--Puerto Rico, in November, helped INCA group and that ADHO organize a conference in Cuba, and we just got approved at the Alliance of the Humanities for a special group that we call Global outlook DH. It's a sort of meta group, to figure out what's going on in medium and low income countries in Latin America, Africa, Middle East, etc. and build bridges with these folks, try to establish communities of engagement and practices, so we can bring a more international crowed into the DH conference, into the DH communities which are mostly Canada, US, Europe, Australia, now Japan.

That work will depend a lot on translation, yes, and while there are great tools that combine automated work with wiki style crowdsourcing elements that also have a revisioning process so we can keep track of how translations change over time, and we need repositories to build international repositories. Tools are present, but building communities is really hard. Most of us in our disciplines are catching up with technology; the hard part is we shave communities, but we don't understand how to link communities. Need to build a truly global HUMANities (all of us, not just the United States). In order to really understand humanity, we must engage with humanists everywhere.

A call for arms: help build bridges with the low income to middle income countries--through Global Outlook, or other projects-to start building partnerships with people outside the usual areas.

Language: Open Source Tool Travel Wiki (French developers).

Building Networks: DH's Global Perspectives (start pressuring them)
Get graduate students who are working on postcolonial and comparative literature involved in the digital scholarship produced.

Some of that stuff doesn't even look like the DH we have around here.
Tendencies to reify as a sort of practice now in the profession--this is a good opportunity to do more than the Big Ten.

Puerto Rican scholar building live narratives online, an archive of folks recording their stories, in a small town in Puerto Rico. Live narrative becomes a really great resource for anthropologists, sociologists, but the humanities people don't do anything with this sort of unfamiliar projects.

Come to conferences outside the US (announces one he doesn't want to put on Twitter).
Release will be announced after this panel.

This is truly moment of opportunity for our profession: we got the theory; it's time to now build something.

(Cannot get name): Life stories, working on Quebec author who was anthropologist, something in Canada project (modernist Canadian texts in digital format) (Voice fairly low) Working with anthropologist who produced text on life narratives from indigenous communities in the northwest. Five sections, including main narrative, stories, index, including stories from (painters?). What he's found in working the last three or four years, and attempt to digitize it, and bring all this information together, when we talk about expanding access, we have to be careful about who has access? In relation to indigenous communities: I cannot make this project work unless he has the full support of community. It's also making connections between indigenous communities. The text was written in English, from translations from (cannot spell cultural names), ethnographic and literal/linguistic translations. Both are problematic. Linguistic based on colonial orthography, created to produce religious texts to convert indigenous peoples into Christians/English. Orthography shifts in the 1960s in order to politicize the language which means that the (SP?)Gicksan doesn't exist anymore not used by the community. The Giksan is used today is read to become endangered, but it's close. The access to the language its precarious; the elders in those communities are the only ones really working but save it this moment. Little work being done on language because pipeline is scheduled to go through their territory, so they are not interested in putting together colonial mentality text. Cultural translations, the stories that were recorded, are oral in nature, they're not print stories, so how do you digitize a text, and use translated documents to put up for others to read when that's not how they're supposed to be read. The Gicksahn community agrees it is necessary to do it because of the loss of their language, but there are recordings at the Canadian Museum of Civilization, hundreds of them, the text is produced from many different versions (Boaz, Sapir, Scarbo, and author of novel). The novel combined it into one story, authoritative text, against the nature of oral texts and community's wishes. I have been in contact with them, and the main character in the text is alive today, and he has agreed to do this. But after four years, he doesn't think it will ever be produced in the way he wants to make it. They're one person's stories, and it's up to that person whether that information goes up. Politicization of information and access to that information. Going to DSI this summer (?) to see what he can learn/expand on for indigenous digital archives and what can go up. Doing this work today is to consider the fact that the copyright and ownership of information that could be used by the state to well control indigenous community members.

Next: Moffat, Associate Professor of Quebec literature, at director of NT2 (Nouvelle technologies, novellas textualitis). Today I am going to talk to you about the (Gabrielle Roy project--novelist, 1909-1983. Mostly known for her first novel published in 1945 which is the first social novel in French Canadian literature. This is part of NT2 project. Design and maintenance of a site which includes academic editions of author's manuscripts, a critical analysis; bibliography and eventually a space for discussion. Gathers geographical and biographical information. Site is followed by virtual community dedicated to dissemination of knowledge about her work. One of this project's objectives is to make use of the platform for editing and publishing--contribute to greater knowledge about her work, and certain ideas of literature research, and translation studies as well. She (author?) considered English translations equal to French originals. We also hope to contribute to the development of certain areas of literary research, history and working of Quebec, and Canadian literature in French and English. The site is designed according to an open source philosophy: it is designed to be a dynamic space, administrative changes, as the discourse progresses, and as more work appears. Sharing of knowledge is not just the object of study, but designing the virtual space. The general thesis underlying the project is that development of systems that allow access to knowledge will lead to different kinds of scholarship. Access to primary object of research seems to promote of innovative reading strategies and to anchor virtual communities around work of other authors. The biggest part of work involve electronic editing of manuscripts and files of published and unpublished works Most are in archives of Canada (missed name of archive). The (offline)?community is voluminous contains twenty meters of textual records-the digititalization offers free access and brings more attention to her work. Important issue about electronic publishing of her works: canonical and unpublished texts. Many projects (name of one in French): the trend of electronic editing is becoming a recommended means of publishing --but not the same as traditional critical editing. Specific protocols need to be defined and established for this method.

In case of Roy's work, well known, the two approaches are not conflict: use electronic editing for the minor and unpublished works (stories, letters, etc.)--the issue of whether she intended them to be published or not, but she did place them into the archive, meaning them to be read. Allows comparison between published works, and the other work.

Specific case of online publication of her autobiography (she began writing four parts; completed two of them) published posthumously. The third part, unfinished, was published after. Fourth (?) about 100 pages, published electronically. Image on screen of the two manuscript drafts--handwritten in blue ink scanned onto it, and then a manuscript draft, also includes critical and explanatory notes, which provide details about people, places, intertextual references. Able to reproduce long (extracted?) works which explain certain passages. An exhaustive report of variance to see all three versions of the same text.

Index of names, people, places, references to other works.

We are trying to wipe up the electronic edition….four thousand pages of manuscripts to process, and over 15,000 variants.

Go back to aims of broader electronic and publishing project: work on the manuscripts is a part. The editing and analysis work performed within the context of this project, and the fact that electronic resources are used as the main means of presentation, will ensure a wider presentation and understanding of her work and writing process, and also Canadian and Québec literature, and pave way for major scholarship on the work, pub an unpublished, of Roy.

Some this documentation has not been used and may add to scholarship. Explore the documentation and make it available. Another challenge we are facing is the translation; bilingual version is goal in next few years.

Brian Larson: this roundtable arose in part out of a blog post (name?) raising cultural critique of DH. The "cultural and political problem of software platforms almost exclusively produced in the American environment." Wants to point out the way Anglo American researchers who have created the programs for natural language processing have worked to make work accessible through the world. Rhetoric and scientific communication, cognitive science, copyright linguistics lawyers. Uses computational linguistics but in rhetoric because computational linguistics methods are not rich enough. Not a DH (mentioned book must check out). I read them as saying DH describes the OUTPUTS, rather than INUTS and TOOLS.

Is it using digital tools applied to humanities--computational tools applied to rhetoric? I'm not quite a DH, and it's weird to be with people who are such a good vision of what DH should be from a broad sense.

I want to make the argument in contrast to (blog post): researchers have done a lot to make tools accessible to developing world. Alex's comments connected--tools are available.

First: if you're looking at using natural language processing in your research, much of the tools are open source, and the literature is open access. Natural language processing experts seem to be prepared to attempt to bridge the gaps.

Leading international association for NLP is association for computational linguistics; committed to open access (Computational Linguistics) and has been for years. ACL anthology (decades of presentations and publications): the heart of that literature is open access and widely available. Conference papers are important in light of fact that presenters must submit completed papers, not just proposals, so there's this rich literature of the kind of things that get lost. There's also some junk, but it is available free of charge. Pretty much all published in English, so colonization by English of the rest of the world, I cannot defend it against that.

So that's the literature.

Doing NLP requires software tools--open free widely available. Describe: General Architecture for Text Engineering (GATE), and another whose name I didn't catch--sheet is at back of room, with link to website.

Any of ALP tools are language dependent, part of speech taggers rely on a dictionary of works from language they are processing, and has to be trained on a corpus that has been tagged for parts of speech by linguist.

Lemmatizing: it's how we tell the fact that woman and women are two forms of the same word (singular and plural); relatively simple process in English because morphology is simple, but in Portuguese, and Arabic, with multiple verb forms, more complicated. With regard to Arabic, the tools have been deployed, but in GATE and Stanford MLP.

Gate==UK, University of Essex.

Distribute under general public license.

Largely free to nonprofit an academic uses. Runs in JAVA. Can be used easily and widely with old machines.

Development community for GATE has plug ins for non-western language (Arabic, Chinese, India, a Philippine language, didn't get the name). Also European languages.

Stanford MLP runs in java, general public license, although identifies itself as being focused as processing texts on English language, there are resources for processing Chinese and Arabic.

NLTK (Python) Natural Language tool kit

Weka (suite of machine learning tool, University in New Zealand).

Availability of NLP tools does not address all the problems (scholar) raised--training languages for programing languages are built on English's

Examples of the use of these technologies that bely complete lack of critical awareness of their applications--NLP is seeing substantial work around the world, and the researchers at the Anglo American center of the field do appear to be making efforts to make work accessible.

QUESTIONS: 25 minutes

Invited to share own work.

Computer scientist attendee--using NLP tools--names them--word net?--tool are out there, and important thing is to use correctly. If you use the tools under the assumption that tools are not perfect, you can accomplish a lot. Have to bring together different tools to work with on projects.

BL: Dangerous to assume that tools for one project can work with other projects, especially around the world, is not best approach. People cannot decide in advance what is needed.

CS: problems faced by computer sciences. Great idea of TOOL, but question is who will be using it.

Participant involved in Caribbean: mail chains and exchange articles--networks of information in Cuba that operate because of low percentage of people on the internet. How do you build this sort of commons in this situation (no T1 communication). You cannot take the access to technology if you take for granted that people all over the world have the same level of access as the Anglo American areas. Need first.

Audience member: all sorts of work being done that isn't known more broadly--lots of examples. We lack big tent structure for scholars working in DH to look at and know what others are doing--global and smaller scale. Lots of misconnections.

Lee: we're taught to narrowly specialize in disciplinary terms--cannot go to all presentations. One of the challenges DH faces is that we're supposed to process narrowly. Especially need to get programmers and humanities people together.

Sarah: University of Chicago--there are at least two major projects building bridges to Latin American communities. Speaker knows those folks International Association of Librarians.

Lee: learning curve issues regarding Digital Humanities. More stuff to read and attend.

Caribbean: Not only working with generalists, but DH feeds and information gives whole world of view.


Lee; small, geographically isolated. Has reached out.

I talked about need to get together with other small universities.

Another member: Throw up a wiki--share--collaborate -- libraries not always the best, and people don’t' always maintain links to libraries once they leave.



Challenge to librarians.

Subject guide of 20th century is in flux..

Language and multilingualism In theory, there isn't actually multilingualism.

What is the role of language requirements and language training in this? Institution required to have two languages other than English but there is no support for getting it--very little support in English programs. If I cannot get support to learn French in Canada…

Lack of immersion schools, language acquisition, not easily accessible.

Arguments on academic levels about who controls what material via discipline.

Belgrade DH, Serbia -- rain on parade--try to explain to customs people that you do DH in Serbia (second level of questions). It's not only about languages, it's not only about culture, it's about MONEY. Inequality that all this good will will not be able to change; we have tried to do work with DARIA , European project, when you have people who have huge budges, and we're doing everything on the shoestring.

LEE: you raise a very good point, and thank you for saying it. It needs to be said.

Other; need to involve capitalists.