Versions of Melville

During his lifetime, Melville published nine book-length works of fiction, a collection of tales, and four volumes of poetry. At his death, among other documents, he left in manuscript several uncollected tales (including Billy Budd), numerous poems (including Weeds & Wildings), and a handful of reviews, including his literary manifesto “Hawthorne and His Mosses.” Most of Melville’s manuscripts appeared posthumously in print. For various reasons, almost every Melville title exists in multiple versions, either in print or manuscript, due to authorial or editorial revision. Changes to a work also occur in forms of adaptive revision, with or without authorial control, as in plays, films, opera, or even the cultural “meme.” As a body of work that exists in multiple versions, Melville’s corpus of writing is a remarkable fluid text, and therefore a considerable challenge, editorially and digitally.

MEL’s editorial project—funded by NEH and supported by Hofstra University and its Digital Research Center (DRC)—began its work by designing “model editions” of three exemplary Melville works: Moby-Dick, Battle-Pieces, and Billy Budd. By resolving the different editorial and digital problems that each work poses, these models will serve as templates for MEL’s editions of additional works. Why we chose these three fluid texts as models for digital editing has to do with the nature of print, manuscript, prose, poetry, and digital technology.


Moby-Dick. First appearing in 1851 in an American and an expurgated British first edition, Melville’s best-known work is also a significant example of a fluid text in print. In addition to the two first editions, scholarly editors create their own versions. For instance, the standard Northwestern-Newberry (NN) edition of Moby-Dick generally follows the American text but mixes into it elements from the British edition; it corrects obvious errors but also adds new wordings that represent what the NN editors believe Melville actually intended. The Longman Critical Edition of Moby-Dick—a fluid-text print edition edited by John Bryant and Haskell Springer—also follows the American text and makes corrections of obvious errors, but it does not mix versions. Instead, it provides a set of “revision narratives” that demonstrate changes—regarding style, aesthetics, and matters of sex, religion, or politics—made to Melville’s text, either by Melville himself or his editors, both historical and modern. To create Versions of Moby-Dick MEL’s editors use Juxta Editions to collate the American and British editions, highlight all variants (or textual fluidities), and display revision narratives explaining each change. Juxta Editions also enables MEL editors to provide textual notes and contextual annotations, with links to images and data in MELCat. In future development, MEL will extend the navigation of the versions of Moby-Dick to include adaptive revisions of the novel, particularly in film.

Digitally, the challenge of Moby-Dick lies in establishing protocols for transcribing and coding a prose fiction, in print only. Billy Budd poses the entirely different problem of editing a fluid text that exists fundamentally in manuscript but also in non-authorized, modern print transcriptions.

Billy Budd. Melville left his completed but unpolished manuscript of Billy Budd_unpublished at the time of his death in 1891. The manuscript is riddled with thousands of revisions, in pencil and ink, that bear striking evidence of the writer’s shifting intentions throughout at least three versions, layered over some 350 leaves, representing at least eight stages of composition. The novella first appeared in print in Raymond Weaver’s 1924 transcription (corrected in 1928). More accurate transcriptions based on more ambitious analyses of the manuscript appeared in F. Barron Freeman’s 1948 version and in Harrison Hayford and Merton M. Sealts, Jr.’s 1962 edition. A slightly modified version of the Hayford and Sealts text appears in the 2017 NN edition of _Billy Budd. These four scholarly editions base their texts on direct transcriptions or inspections of the Billy Budd manuscript, located at Harvard’s Houghton Library. Each of the three twentieth-century transcriptions vary significantly from each other, and only the 1962 edition provides a “genetic transcription” as a textual apparatus.

MEL’s Versions of Billy Budd assembles both manuscript and print versions of the work, and to meet this challenge, MEL developed TextLab. This innovative tool enables editors to mark-up revision sites on manuscript leaves, transcribe each leaf (including its revision texts), automatically code the additions and deletions, and generate a “diplomatic transcription” (a typographical simulation of each leaf or page) that is in turn a platform for displaying revision sequences and narratives for each revision site. TextLab also automatically generates MEL’s base version of the Billy Budd manuscript—essentially Melville’s final reading version with additions added and deletions deleted—which can be collated with the three twentieth-century transcriptions.

Battle-Pieces.** Melville’s first published volume of poetry, Battle-Pieces and Aspects of the War, poses other problems. Although working drafts of these Civil War poems have not survived, evidence of Melville’s post-publication revisions appears in pencil on certain poems in two of his personal copies of the book. Battle-Pieces presents the digital problem not only of coding for poetry rather than prose but also of layering handwritten revisions over text in print. Melville also published a handful of Battle-Pieces poems in magazines before the volume’s publication, and certain poems appeared as reprints in other publications in Melville’s life. The challenge of integrating magazine and anthology texts along with Melville’s penciled revisions is also addressed in the modeling of Versions of Battle-Pieces.

In addition to these matters of textual fluidity, each of MEL’s model editions includes sets of contextual annotation, which involve the work of our four MEL research groups—Art, History, Travel, and Editions—and their implementation of two other tools: MELCat and Juxta Editions. These tools, along with TextLab (and others), are explained in MEL’s Tool Kit.

In 2017, MEL will begin work on editions of Typee, Piazza Tales, Melville’s journals, and the Melville Family Correspondence.