The middle distance

Ted Underwood’s recent posts about literary and non-literary diction between 1700-1900, and the various discussions they sparked, including Katherine Harris’s post on gender and DH archives, have had me thinking a lot about cultural poetics and the middle distance.

In 19th-century studies, most DH projects have tended to operate at two different scales: large-scale text analysis projects (associated with so-called “distant reading”) and curated editorial or archival projects, which have mostly been focused on works by or about particular individuals. Such archives extend and sometimes transform print-form editorial theory and practice, but tend to have an author as the center of gravity, around whom particular material forms (manuscripts, books, images, etc) cluster. In truly large scale text analysis, on the other hand, the linguistic text is usually separated from the material forms in which it circulated and the individual author tends to disappear, in favor of pursuing questions about diction and linguistic variance.

I’m interested in exploring how DH tools and methods could be brought to bear on the kinds of research that take place in between those poles, in the middle distance.

Cultural poetics (or, to use Bourdieu’s term, sociological poetics) is concerned with the forms literature takes in the world: the modes and methods by which texts are produced, consumed, understood, repurposed, and valued by individuals and institutions. The digitization of 19th-century texts offers us tremendous resources for exploring these topics, but there’s lots of work yet to be done in theorizing and demonstrating ways to pursue this through digital analysis.

Thinking About Genre

Genre is a property of texts and of literary works. Genre is not a feature of books or other print forms. But information about a text’s genre is encoded in various aspects of the material book, such as tables of contents, chapter or section titles, running heads, page design and layout, typographical choices, page size, paper type, etc. To understand how a genre functioned at a particular historical moment, we need to examine how it circulated in the world in its various material instances.

That’s why I’m currently working with bibliographic metadata, compiling a database of information about single-author English books of poetry published between 1840-1900. I see this as a small first step towards theorizing that middle distance and making it available for practical critical endeavors.  When completed, this database will contribute to mapping the cultural field of poetry in the Victorian period. (I’m also collecting many of these books in digitized form for future analysis — but the terms of that analysis will be guided by the historical patterns revealed in the metadata.)

As is true with any collection of data, the historical, material, and literary terms that I combined in defining the project constrain it in sometimes awkward and artificial ways. As soon as I started gathering data from various sources, the limitations of the term “book of poetry” became really obvious.

Readers in the 19th century rarely doubted whether or not they were reading a poem, because the genre of poetry is encoded through a variety of linguistic, textual, graphic, and paratextual means. But the 19th-century books (and periodicals) that contained those poems were frequently multi-generic in ways that literary studies rarely wrestles with. Posthumous collections of a writer’s works might contain poems along with essays or stories. Biographies or memoirs about an individual frequently included collections of poetry which might or might not have been previously published.  Subsequent revised editions of a published collection might include an author’s reminiscences or other material.

As much as I’m looking forward to getting to analyze this set of data, I’ve found that working through a series of decisions about “what counts” and how best to collect it has helped me better understand my theoretical and critical commitments. There’s no perfect, clean, complete set of textual data (as Lisa Rhody recently pointed out) –not only because of technical and resource limitations, but because that’s the textual condition itself: messy, complicated, and ever-changing.

In other words, I’m really glad that “single-author books of poetry” have proved to be somewhat difficult to locate. It is in the interstices of our logical and historical categories that we can begin to perceive our critical blind spots.


  1. says

    Natalie, what a great post, and it also reminds me that I reluctantly missed your MLA presentation. You move so seamlessly between these huge corpora and the materiality of textual object itself. Throw in there for good measure the sociology of the text, that D.F. McKenzie process for the creation and dissemination of books and ideas. You do this so well here. What happens when the data also needs to account for other forms of writing, such as music? Poetry was often written towards musical scores though I’m not sure which one came first in some instances. And what happens when we have to interpret the editor’s footnotes alongside the poetry itself? (In the early 19th Century, Alaric Watts was mighty famous for intruding upon the reading experience of any volume.)

    This is all to say, kudos for bringing these textual studies methodologies to bear on your digital project. And, yes, here too I think that gender is marked on the poems. If I ever get these literary annuals scanned into dirty OCR, I would love to turn them over to you for interpretation of the various genres along with the material object’s beauty.

  2. nmhouston says

    Thanks, Kathy — I don’t know much about the textual history of musical scores — although I did run across someone at DHSI a couple years ago who was working on digitization and OCR for music, which raises lots of interesting problems. Gender is definitely one of the threads I’m pursuing in my analysis — I’ll look forward to sharing some of that work in future conversations –