Introduction: Computation and Digital Text Analysis at Melville's Marginalia Online

Abstract

Introduction:Computation and Digital Text Analysis at Melville's Marginalia Online Christopher Ohge and Steven Olsen-Smith In an erased marginal comment on Shakespeare's character Parolles, the dissembling rogue in All's Well That Ends Well, Herman Melville invoked the most unexceptionable of mathematical formulas to reflect darkly on the abiding nature of human depravity: As 2 & 2 made 4 in Noah's time, as now,so man [?figures] ever. Here we have acharacter very common in the Rail RoadCar of the [?most mighty] nineteenth century.1 (Shakespeare's Dramatic Works 2.406) Roughly forecasting the subject of Melville's tenth book, The Confidence-Man, the recovered annotation also reveals a seemingly incidental but nonetheless significant penchant on Melville's part for quantification. Quantities appear frequently in Moby-Dick and in other writings by Melville, whose fascination with numbers was demonstrated recently in the pages of this journal by Zachary Turpin. Numbers are not scarce in the author's marginalia either. "There are 75 folio volumes in that," Melville observed of Jaques's words in As You Like It (2: 284.26–27). In light of this predisposition to unite the quantifiable with the profound, the Melville's Marginalia Online (MMO) staff offer in the following essays digital text analyses of Melville's marginalia in his surviving copies of Homer, Shakespeare, and Milton. Methodologically unprecedented, and conducted in an integrative spirit, the essays combine distant with close reading, exploring hitherto unknown or under-appreciated evidence of Melville's engagement with these three major predecessors. Quantitative and qualitative digital scholarship have led to a mixture of promising results and disillusionment, with increasingly vocal proponents and detractors. Opposing conceptions of the field are apparent in Timothy Brennan's recent article, "The Digital Humanities Bust," and the [End Page 1] vigorous defense it provoked from many digital scholars (for examples, see responses by Mandell and by Bond et al.). While the fairness and accuracy of Brennan's characterizations of digital practitioners' various activities is open to question, he makes a valid point that quantitative studies could more explicitly demonstrate the significance of their findings. High-profile studies to date have mainly addressed broad swaths of social literary output defined by regions and decades––for instance, the thousands of British novels published over a given century. Digitized archives now make such approaches possible and offer an exciting and potentially transformative development in literary studies. On the other hand, the stubborn disconnect between quantity and literary meaning can hamper such approaches, presenting conceptual shortcomings even for macro-analyses performed on data of a comparatively much smaller scale. As Brennan puts it bluntly, "The significance of the appearance of the word "whale" (say, 1,700 times) is precisely this: the appearance of the word 'whale' 1,700 times." But whereas no act of counting can substitute for critically interpreting a literary work, the usefulness and significance of word frequencies––both on their own and in relation to each other––are heightened in the study of marginalia. Indeed, the approaches demonstrated in the following essays indicate that an author's record of marginalia and lifetime of reading constitute a valuable big data set in its own right, albeit a fairly small one in the context of research computing. When the author is Melville, moreover, with a substantial record of marking his books and incorporating a range of sources in his own writing, the data can reward analysis in hitherto unprecedented ways. Only a machine can quickly compute word counts of selected content Melville marked in the plays of Shakespeare's Dramatic Works, or in each book of The Odyssey, or in Paradise Lost. Only a machine can stack a series of graphs showing lexical variety in Melville's markings or instantaneously classify by positive and negative sentiments enormous amounts of words that appear most frequently in passages he marked in multiple texts. Melville himself was devoted to the art of rhetoric and explicitly lauded both the "short, quick probings" and encyclopedic forays of great writers ("Hawthorne and His Mosses" 244). By using machine-derived word counts and visualizations to explore the thematic aspects of Melville's reading, the following essays demonstrate new means for...