19 November 2013

Towards a Prototype of a Digital Harmony

I've been working on a prototype of a digital facsimile "edition" of the earliest extant Little Gidding Harmony. (If you don't know what a Little Gidding Harmony is, see this FAQ I wrote a few months back.) Two pages are currently accessible here. Don't click unless you have a fast connection and are using either Chrome or Safari. As per usual, IE is not recommended.

I put "edition" in quotes above because the Harmony is not really a text one would read today the way one would read, say, Shakespeare's King Lear. Most contemporary readers are interested in the Harmonies as concrete instantiations of a particular compositional process, rather than as textual products. Not only are we interested in process over product, but the product itself doesn't face the problems that so much of textual criticism has been designed to deal with. There are not multiple, variant editions of this Harmony; it's a singular, unique object. Nor is there an audience that requires an "authoritative" edition to read linearly, from start to finish. Even theories of editing that acknowledge the fundamental instability of texts – Jerome McGann and the "textual condition," Randall McLeod's notion of "transformission" – don't really capture what's happening with the books made at Little Gidding.

In fact, one could argue that the Harmonies were designed precisely to counter this instability. Thus the cut-up method is already an editorial intervention. The women of Little Gidding collated multiple printed Bibles with scissors and paste to produce their own uniquely "harmonized" edition. If we see the Harmonies from this perspective, then the task of a contemporary editor is not to pull the text into a coherent whole, but in fact to pull apart the already harmonized text. This decomposed edition would then enable an exploration of the community's cut-up process.

So in beginning to think about an "edition" of a Harmony, I was motivated less by theories of editing and more by visualization and mapping strategies. Rather than generating multiple variant readings of a text, this edition will aim to produce multiple views of the page's landscape, and different mechanisms for manipulating these views. On the one hand, it embodies aspects of my own research on the Harmonies; on the other, it is (or intends to be, eventually) a machine for producing new knowledge.

In designing this first draft of a prototype, I have attended to:
  • openings over individual pages. This is much harder than it sounds in web-based editions. Screen realty is limited and thus valuable; to choose openings (that is, two-page spreads) over individual pages is to trade large, legible facsimile text for a more birds-eye view of the book as a whole. Importantly, you really can't have it both ways. This is a simple yet, I think, not so obvious point. Even with a variable zoom on a full-screen facsimile photograph, you can't ever escape the framing mechanism of the screen itself. As Marie Baxter said in an excellent paper she presented at the Sixteenth Century Society Conference last month, digitized historical objects are a bit like caged animals in a zoo. We think of them as authentic representations, but our perspective is wholly determined by the cage-like screen.
  • topography over text. From the perspective of the contemporary researcher, the pages of a Little Gidding Harmony are more like maps than literary texts, tracing a set of routes and relationships between different points. For this reason, I began this prototype by using the Image Markup Tool, designed at the University of Victoria, to annotate a page image, identifying each individual cut-out piece. Ultimately, I decided I was unhappy with how these annotations appeared in the web view, and scrapped most of the HTML exported by the tool itself. (Some cut-outs are also polygonal shapes, rather than rectangles, meaning that some individual excerpts require two "image annotations" – which produces a mismatch between the visualization of the cut-up and my XML-encoded transcript of the Harmony.) However, starting from this point gave me a set of coordinates, which I was then able to use as the basis for developing my views. Moving forward, I'm looking into other mapping tools. More generally, I'm interested in the question: what happens when we think of the digital edition as a map? Or when we apply GIS technologies to non-geographical image maps? What would a literary edition look like if it were made using Hypercities? I'm inspired by the speculative and conceptual work being done by projects like Z-Axis at the Maker Lab. What happens if we think about these geographical mappings in terms of relationships between elements on the topography of the page?

  • cut-up fragments over coherent paragraphs. The only view that offers legible text is the parsed XML in the textbox on the right side of the screen. I do not like this view. My marked-up transcription contains a good deal of useful information for any researcher of Little Gidding. Unfortunately, none of it is well represented in the utterly dematerialized white-on-gray text you see here. As a next step, I need to rethink how to incorporate the transcription into this "edition." For now, though, I've decided to offer a kind of counterpoint to this plain text: the cut-ups, below. These are generated on the fly from a larger image of the Harmony's page using the coordinates from the source text layers – which means that any page for which we have these coordinates can be pulled apart into the cut-up pieces that compose it. Each piece is draggable. Since I'm interested in putting the Harmonies in conversation with contemporary rhetoric about digital remix, this relationship between the digital facsimile – its layers of paper flattened on the screen – and the code that de-composes it into its constituent parts is conceptually exciting to me; for here digital media re-performs the cut-up method itself, but with a difference. That difference pinpoints the disjunct between paper and code, and is worth a blog post in itself. Conceptual framework aside, though, I'm frankly not sure what to do with this functionality yet. Currently, I'm using it as a kind of sandbox for my own research; I don't know where it's headed, or even if I'll keep it. 

As I mention in my brief description on the page itself, a good chunk of this prototype is held together with duct tape and chewing gum at this point. The code has not been streamlined but is in fact heavy and unwieldy. You'll need a decent computer and a fast connection to access it – and even then, all functionalities won't be available in IE, and probably a few other browsers. Nonetheless, I can see how it could be streamlined, in conjunction with my marked-up transcription. The next step is to begin working toward a cleaner, more robust framework, while tweaking some of the functionalities and thinking a bit more about how to incorporate the drag-and-drop cut-ups.

A big debt of gratitude is owed to Paul Dyck and Ryan Rempel for providing me with the XML of their digital edition of the King's Harmony. This allowed me to use tags that are standard at least across our two sites. It will probably create a headache down the road – but at least now it's a shared headache. Dyck and his collaborators' writings on their digital edition have also been inspiring.

Comments are always welcome.

No comments: