In my previous post, I discussed how using Voyant reveals the infrastructural bias of DH tools when confronted with an Indic language such as Bengali. In what follows, I wish to outline the strategy I adopted in guiding students in an undergraduate DH class through the process of curating digital objects so as to develop practical methods of working around this structural dissonance.
Understanding Digital Curation as a Critical Method
In this age of fast moving social media and doom scrolling, digital curation is a concept that students are already fairly familiar with, even if they don’t necessarily approach the media they consume as a curated whole. In the second part of my ‘Introduction to Digital Humanities’ class, I offered students the opportunity to choose one or two of their favourite content creators, and to examine how they put together their social media pages. In particular, I asked students to pay attention to how the pages structured visual content, scaffolding the same with tags and text links. The goal was to encourage them to notice the intentionality behind the presentation of visual material, as well as to understand the role of tags in organizing and linking content. That images need to be arranged, not just next to each other, but alongside the right text so as to create meaning as a whole was one of the key takeaways from this exercise. The second was that, in any curated set, the descriptor words mattered as much if not more than the images they were describing.
Having thus prepared them, I then moved on to a visual archive they were far less familiar with, the Bichitra: Online Tagore Variorum. Since they had already experimented with the Bengali font on the Voyant exercise, they had some notion of the visual structure of Bengali, and of the challenges of digitizing texts in that language. We paused initially on the landing page to examine the mix of Bengali and English fonts, along with the tabs guiding the user through the collection.
Figure 1: Bichitra landing page.
I then drew the students’ attention to the panel on the left of the screen titled “Manuscripts.” Using this as our primary site of investigation, we navigated via the tabs to examine the myriad ways in which one could organize documents–be it by genre, by titles, by language, or the act of comparing various pages side by side. I asked students to focus on the English words and the organizing principles implied by them.
Our final stop was a sample manuscript page where students could compare the handwritten pages to the typed transcription which included proofing marks to indicate additions, deletions, and other authorial decisions. While students couldn’t read the language per se, they were able to identify the distinctions between the handwritten and typed fonts. As with Voyant, here too, they were able to experience the language, and understand the physicality of the fonts even via the digital medium.
Figure 2: Bichitra manuscript viewer, MS. RBVBMS 006; the panel on the left shows a page from Rabindranath Tagore’s handwritten manuscript of Sahaj path, while the one on the right shows the page’s typed transcription.
Creating Image Stories: Curating Digital Archives Using Omeka
While Bichitra offered students a glimpse into the possibilities of digital curation, I wanted to provide them with a free, open source, and accessible platform where they could undertake their initial experiments with creating digital exhibitions. I chose Omeka for two reasons in addition to its accessibility. The first is that in practical terms, the actual putting together of the exhibit is relatively simple, as long as one has a clear idea of the story being told by the archived objects. The second is that it introduces users to the basics of metadata, thereby revealing the infrastructural relationship between the archived object and information about the object. For this class, I restricted the assignment to using Omeka to create a digital exhibit for Betar Jagat, the biweekly Bengali periodical published by the Calcutta Radio Station between the late 1920s and mid 1980s. The class had already been working on this publication for the Voyant exercise, and were familiar with its layout and font. Working in groups, students built Omeka sites to document their initial findings from using Voyant to mine textual data from Betar Jagat. Along with that, they also chose front matter and advertisements from a few issues of the journal to create a narrative about what they saw as its marketing strategies. They isolated pages from scanned copies of the publication to create Collections (uploads organized around topics or themes), built Exhibits to highlight elements from Collections and tell a story about them, and finally added Pages to accommodate longer chunks of text to narrativize sections of the Exhibits.
Figure 3: Omeka site created by students in ‘Introduction to Digital Humanities’ class, Fall 2024.
As this sample site shows, working with actual pages from the journal allowed students a way into a text in a language they were not familiar or comfortable with. Part of their task was to look at the PDF of an entire issue in the form of thumbnails which gave them a sense of the publication’s layout, the density of text, as well as the placement of advertisements. For example, they were able to note that for the September 12, 1930 issue commemorating Betar Jagat’s first year in print, there were eight advertisements (both in English and Bengali) and two “how to place advertisements”, one in English and one in Bengali, in a 20 page journal. This suggested to them that from the very start, Betar Jagat was invested in generating revenue through advertisements. They were also able to identify at least one photograph which indicated the journal’s interest in being multimodal. Finally, seeing the pages from a distance while thinking about how to create visually interesting exhibits that would highlight aspects of these pages, encouraged students to engage with the conceptual and practical problems of digitization. While for the most part, students produced surface level analyses and narratives, it is worth noting that through this platform, they were able to productively engage with the text, and represent their findings in the form of a cogent narrative.
Interestingly, the bigger challenge for most students was understanding the nature and significance of metadata, and navigating Dublin Core. As they uploaded each item onto their Omeka site, they had to decide how best to categorise each element, ascribe creative and publishing rights, decide on how much to include in titles and how much as description, and finally how to tag each item so as to allow users to navigate the site in a controlled fashion. Alongside this, the emphasis on consistency in metadata meant that the publication’s language barrier often became a distant, at best secondary, problem.
Trying Various Angles, or How to Work with Non-Latinate Texts
When I juxtapose the two findings from this class–students’ ability to create narratives about Betar Jagat, and their challenges with metadata–I find an overall positive outcome. Omeka enables users to compartmentalise a text’s language and its ability to be represented digitally, while simultaneously reminding the user to find ways of documenting language as an integral part of the text’s data and metadata. While far from a perfect solution, a digital curation tool such as Omeka invites us to think creatively about the question of representing a text on screen. If available technology cannot “read” a page in a language that is not English, then how might we offer a potential reading through metadata and relationships between objects?
One of my goals for this class was to introduce students to the challenge of having to find work-arounds when it came to digitizing texts from the Global South. Thus while the first part of this exercise (which I discuss in my previous post linked above) acquainted them with the limited infrastructural support available to DH scholars studying non-Latinate texts, the second part invited them to find in Omeka a way to narrativize both the challenges, as well as possibilities for engaging productively with such works.
