Houston 2014

Houston, Natalie M. “Toward a Computational Analysis of Victorian Poetics.” Victorian Studies 56, no. 3 (2014): 498–510.

• 498: frame - kinds of evidence used to make arguments
• The large-scale digitization of C19 books and periodicals now makes possible a more comprehensive exploration of the cultural history of Victorian texts and their interrelated bibliographic, visual, and linguistic codes.
• new access and new modes of analysis
• 499 qt Dan Cohen "should we be worrying that our scholarship might be anecdotally correct but comprehensively wrong?"
• cites Moretti
• to move past the rhetorical impasse of close and distant, over the past couple of years I've been suggesting this definition of digital reading: methods of literary research and interpretation that draw upon computational analysis to move beyond human limitations of vision, memory, and attention.
  • We are all already digital readers, anyway
• sociological poetics
• 500 intervention: My work restores Victorian poetry to our attention within both Victorian studies and the digital humanities by exploring what digital reading of poetic texts [incl. non-lyric poetry] can tell us about the prevalence and function of poetry within Victorian culture and about gaps and biases in our existing DH toolsets.
• 501 graphs showing relative scale of books pub about canonical poets vs all in 1875 -- knowledge gap: it also indicates something potentially interesting about reprint culture; 7% of books published in London in 1875 were by or about EB Browning but she'd been dead since 1861
• cites Bourdieu "field of cultural production as a space of struggle and dynamic change, constituted by the structural relations existing among various individual agents and cultural institutions at a particular historical moment"
• 502 [Vic poetry as Bourdieuian cultural field] Thus the structures of cultural value that surrounded Tennyson's In Memoriam (1850), Christina Rossetti's Goblin Market and Other Poems (1862), or any other volume of victorian poetry were partly created by all the other books of poetry that circulated during the period.
• her DB collects biblio metadata for single-author books but not anthologies - knowledge gap
• The process of transforming this existing bibliographic metadata from the linear format of the list into the relational format of the database allows me to explore large-scale patterns in the production and distribution of poetry that traditional catalogs currently cannot address, such as: what patterns of growth or decline in the publishing of poetry are visible over time; how the proportion of male to female poets varies by publisher, decade, or kind of poetry; and the ways in which poetry's publication was dispersed through Vic print culture.
  • using network graphs/network analysis
• 503: The cultural field of poetry during this period was also shaped by editions of the works of older poets...authors with works published by three or more different publishers in 1866: Homer, Shakespeare, Spenser, Milton, Goldsmith, Cowper, Burns, Scott, Wordsworth, [etc., etc.,]. Such quantitative measures can help us begin to understand how poets from earlier periods helped structure the Victorian literary field alongside contemporary C19 poets.
  • and can send us back to the archive of periodicals to get a more nuanced picture of the cultural forces surrounding reprint culture -- connection to my Vic periodicals project
• 504 page as interface -- How we distinguish and prioritize among the semantic, bibliographic, and graphic levels of a text always already constitutes an act of interpretation. In a multi-year project entitled The Visual Page as Interface, I've been exploring different computational approaches to analyzing the graphic codes of Victorian poetry.
• 505 Our critical attention to Victorian material textuality has mostly focused on writers whom we know to have been actively involved in publishing practices, like Charles Dickens and William Morris, and on decorated books by Morris and Oscar Wilde. I'm interested in examining the unremarked material aspects of ordinary books of poetry instead.
• 506 because the visual features of printed poetry, such as line and stanza, correspond to linguistic features of the poem, the VisualPage application can also be used to identify different poetic forms in large document collections. Blank verse and the sonnet are two examples of poetic forms easily perceptible to the human eye when thumbing through the pages of a book; we are essentially teaching the computer to perceive these forms as well.
• 508 At the simplest level, computational text analysis and traditional literary criticism are predicated on the same theoretical assumption: that there is a relationship between the words occurring in the text and the meaning of that text. This may seem so obvious as not to bear stating, until you begin to formalize it into a method. One of the clarifying effects of working computationally is the need to define all of one's terms and procedures as specifically as possible so that they can be translated into the logical structures of code. In order to read a text (with human or computational insight) one must decide what defines the text and its large and small constitutive elements, down to the level of the word or syllable. One must decide which words to pay attention to, or count. And one must decide how to define the meaning discovered, derived, created, or found in these words. Readers and literary critics already make these decisions, more or (usually) less explicitly. Digital reading offers the opportunity to put one's theories of the text to the test of practice.
• When scholars frequently describe Rossetti's language as simple, and Barrett Browning's as complex, [509] word choice is one of the features...that support such descriptions. I'm interested in using quantitative measures of poetic style to support different kinds of comparative analysis: to measure more concretely the similarities and differences we already perceive, and to begin to see new patterns across poems, books, and poetic careers.