Multispectral imaging and manuscripts

My essay, “Multispectral Imaging and Medieval Manuscripts,” appears in a volume devoted to digital humanities and medieval literature and culture, edited by Jennifer E. Boyle and Helen J. Burgess. My essay introduces a new digital approach to medieval manuscripts by summarizing a Mellon-funded project in the digital humanities at Yale University. Here’s the opening section:

Debates in the field of manuscript studies often take the form of disputes over material details that are at least theoretically quantifiable: the curve of an ascender, the shape of a damaged leaf, the color of a paraph mark. Until recently, such debates were adjudicated entirely on the basis of comparison by human eyes. Paleographical and codicological research along these lines has been transformative: it remains the edifice on which rest all other forms of knowledge about medieval texts. Yet human eyes have scientifically well-understood limitations, especially as regards the perception of color. Recently, it has become possible to supplement the human eye with a digital one.

Human eyes see in three colors, whose various proportions the brain interprets as the range of colors we perceive in the world. There are finer ways of slicing up the color spectrum, however. The eyes of some species have developed to see in more than three colors, enhancing discrimination between closely similar wavelengths of light. Computers, too, can ‘see’ in many more than three colors. Digital capture of spectral information in more than three colors is known as multispectral imaging. (Hyperspectral imaging, marginally more accurate but substantially more costly and time-consuming, refers to digital capture of spectral information continuously across the color spectrum.) A non-invasive procedure involving a professional camera and a computer, multispectral imaging reveals differences in reflected light that remain invisible to human eyes. Multispectral imaging also captures data across the infrared and ultraviolet ranges, which lie, respectively, beneath and above our visible light spectrum. In other words, computers can be programmed to ‘see’ more and more finely than human eyes.

Progress in computer science thus enables the creation of new kinds of data about medieval manuscripts (among other objects of interest). These new data promise to confirm, enrich, problematize, or even invalidate the conclusions of traditional manuscript study. This essay outlines the scope and goals of interdisciplinary research at the intersection of computer science, manuscript studies, and cultural heritage preservation. It does so through an overview of a case study: a research project in the digital humanities carried out by computer scientists, medievalists, and digital preservation experts at Yale University from 2012-2015 under the auspices of an Andrew W. Mellon Foundation Scholarly Communications and Information Technology grant. Entitled Digitally Enabled Scholarship with Medieval Manuscripts, this project had three arms, of which I will discuss one, “Creating English Literature, c. 1385-1425: Inks, Pigments, and the Textual Canon.” While at Yale I served as graduate assistant for “Creating English Literature” to Principal Investigators (PIs) Alastair Minnis and Barbara A. Shailor, who were later joined by PI Ardis Butterfield. The other two arms of Digitally Enabled Scholarship with Medieval Manuscripts are “Editions of the First and Second Recensions of Gratian’s Decretum” (PI: Anders Winroth) and “A Literary History of the English Book of Hours” (PI: Jessica Brantley).

multispectral setup

The Yale team’s technical setup for multispectral imaging of medieval manuscripts.

Counting and scanning

This weekend, I presented a paper at Digital Britain: New Approaches to the Early Middle Ages in Cambridge, MA. My gratitude to Sam Berstler, Joey McMullen, and Erica Weaver for the invitation, and for organizing the conference. My paper, “Counting and Scanning: Qualitative and Quantitative Approaches to Early English Meter,” is an essay in the philosophy of science, focusing on the use of quantitative reasoning in a specialized field of literary study. Here’s the opening frame:

This is a paper in defense of counting in literary studies. It’s also a paper about why counting needs defending in literary studies in 2016. First, I’d like to connect these topics to the conference theme.

The so-called digital turn in academic knowledge production poses a material challenge to the methodological status quo of the humanities. Digital technology makes a bold promise: the quantification of human experience. For constituents of the status quo, the promise of digital technology seems more like a threat. Many humanists worry about what is lost when the richness of culture and the arts is reduced to data. For others, digital tech enables the humanities to live up to their potential. To take an egregious example, Franco Moretti’s ‘distant reading’ does not merely claim to be a good method of reading literature; it claims to be the best method. In promoting statistical approaches to literary history, Moretti taps into philosophical attitudes that predate the digital. His aggressive empiricism is at once the attraction and the scandal of distant reading.

Thus the digital turn irritates a long-standing tension in humanistic study between qualitative and quantitative methods. I propose to explore this tension by focusing on a field of inquiry that bears an unusual relationship to it. Metrics, depending on your perspective, is either the field entrusted with explaining what makes poetry poetry, or the field charged with inferring an organizational system from linguistic patterning. Metrical scholarship combines qualitative and quantitative reasoning in a way that has become uncomfortable in literary studies.

Discomfort with the modus operandi of metrics is acute in the case of early medieval literature, whose systems of formal organization lie at a great cultural remove from those of modern literature. In reconstructing early medieval meters, the value of modern experiences of poetics, and hence the network of assumptions underlying data collection, is always open to question. About the meter of Beowulf, for example, it is currently possible to hold any one of a number of mutually contradictory theoretical views. Moreover, Old English metrical theories cluster in two incommensurable research paradigms, one continuously elaborated since the nineteenth century, the other disclosed for the first time in 2008 (see Cornelius, “The Accentual Paradigm in Early English Metrics”). Metrists disagree, and they disagree about why they disagree. For many literary scholars, meanwhile, metrics is a field that deploys the rhetoric of accumulation to make exaggerated claims on historical truth.

In light of these problems, what I’d like to do this morning is to sketch the function of counting in metrics, with reference to my research on English alliterative verse. I have two goals: to affirm the role of quantitative reasoning in literary study, and then to set a certain limit on that role. In what follows, I identify and address two philosophical challenges to counting as a critical method. One challenge comes from the right, framed in the language of neopositivism; the other comes from the left, framed in the language of poststructuralism. I contend that the neopositivist and poststructuralist positions both miss the full significance of counting as a way of getting at “the heart of a poem,” in the words of Simon Jarvis. Ultimately, I’ll argue that metrical study explodes the distinction between qualitative and quantitative approaches to literature.counting and scanning header

Stanford visit

Photo credit: Elaine Treharne

This past week, I visited Stanford University as a Text Technologies Fellow. While at Stanford I also spoke to the Workshop in Poetics, recorded a video interview for a developing digital resource for the study of prosody, and guest-lectured on early English alliterative meter in English 301B, Love and Loss in Early English, 900-1300. My gratitude to Elaine TreharneArmen DavoudianRoland GreeneNick Jenkins, and Mary Kim for these invitations and to Armen, Mary, and Daeyeong (Dan) Kim for making local arrangements. Here is a summary of my visit on Stanford’s website.

As a Text Technologies Fellow, I gave a lecture to the Stanford CMEMS Workshop entitled “The Old English Exeter Book and the Idea of a Poem.” This lecture represents new thinking at the nexus of poetic meter, manuscript form, and the history of ideas. My thanks to the attendees for helpful questions and criticisms. Here’s a modified version of the opening frame of the talk:

How do you know a poem when you see one? Consider the opening of the Battle of Maldon from the Anglo-Saxon Poetic Records, the definitive modern edition of Old English poetry. We can tell this is a poem because it has a title, it is lineated, and it is thoroughly punctuated. We can tell it is a critical edition because the textual variants are listed on the bottom of the page.

Early medieval readers did not rely on any of these modern features of textualization when encountering English poetry on the page. Consider the opening of the Old English poem we call the Wanderer. No title, no lineation, almost no punctuation; certainly no textual variants.

In this paper, I ask how Old English poets and their audiences conceptualized poems as poems. I approach this question from a codicological perspective and from a metrical perspective. I propose to explore how these different categories, material and formal, acted in tandem or in tension in delineating the idea of a poem in early medieval England. Throughout, I focus on the Exeter Book, a large tenth-century anthology of English poetry. In particular, I consider three poetic sequences in this manuscript whose form challenges modern conceptions of ‘poem’ as a unit of composition.

As a visitor to the Workshop in Poetics, I presented a work in progress entitled “Before Prosody: Early English Poetics in Practice and Theory.” This essay relates the conclusions of my first book to the emergent research paradigm known as ‘historical poetics.’ My thanks to the members of the Workshop for incisive questions and comments. Here’s an abstract:

Scholars of Victorian poetry have called for a ‘historical poetics’ that would reevaluate the received narrative of English literary history by recovering alternate ways of theorizing and experiencing poetic form. This essay takes a longer view onto the histories of English poetry from the perspective of Old English and Middle English verse. The primary purpose of the essay is to offer medieval English poetry as a case in point for historical poetics, thereby bringing a different literary archive to bear on critical conversations about the theory and practice of English versification. The contribution of this essay to the field of historical poetics will be to indicate a constitutive gap between the practice and theory of verse. Through three case studies drawn from ongoing research on the alliterative tradition, I seek to demonstrate what is distinctive about the cultural work of early English poetics. Recognition of the ways in which modern questions fail to illuminate medieval meters is the first step toward a more capacious historical poetics.

As a contributor to a developing digital resource for the study of prosody, I discussed alliteration as a poetic device and an ornament in alliterative meter; the alliterative tradition, Old to Middle English; and the state of the field of alliterative metrics.

As a guest lecturer in English 301B, I taught a class of undergraduates and graduate students about the history of the alliterative metrical system c. 900-1200, with examples drawn from the Battle of Brunanburh (c. 937), Durham (1104-1109), and Lawman’s Brut (c. 1200). We asked how this metrical system stood around 900, how it changed between then and 1200, and how modern scholars have conceptualized these metrical principles and transformations. It was exciting to be able to help the students connect meter with the primary concerns of the seminar: linguistic form, literary style, periodization, and manuscript context.

A companion website

Through a generous grant from Boston College’s Academic Technology Advisory Board, I have received funding to use the MediaKron digital toolkit to build a companion website for my undergraduate course, Middle English Alliterative Poetry. An in-progress version of the site is publicly available here. The site currently features a short guide to Middle English pronunciation, with sound clips; a short guide to Middle English alliterative meter, with bibliography; a short guide to medieval English codicology and paleography, with annotated manuscript images and bibliography; and a timeline of poems and significant historical events, with short descriptions and bibliographies. Through collaboration with my students, the site will eventually feature an introduction to each course text, with annotated manuscript images and bibliographies. Here’s the course description, which also appears on the landing page of the website:

In the fourteenth century, there were two ways of writing poetry in English. Chaucer’s rhyming, syllable-counting iambic pentameter exemplifies one tradition. This course makes a survey of the other tradition, known today as alliterative poetry. Among the poems we will read are tales of King Arthur’s court, the story of a resurrected corpse discovered in London, and a wild allegorical dream-vision starring such characters as Bribery and Truth. We ask how this poetry is formally organized, where this form of writing comes from, and why medieval English writers chose to use it. No prior knowledge of Middle English required.