What is a Fluid Text?

Apart from offering reliable digital editions of Melville’s writings, MEL’s signal editorial contribution is that digitization also enables us to read Melville’s works as fluid texts. A fluid text is any written work that exists in multiple versions due to authorial, editorial, or adaptive revision. These kinds of versions are critically significant in studying Melville’s revision process and the way editors and other readers have also changed his texts. All written works are fundamentally fluid texts; the problem is gaining access to the versions that reveal the revisions and editing the revisions so that readers can read them. That problem is best addressed through digital editing.

In two books—The Fluid Text: A Theory of Revision and Editing for Book and Screen (2002) and Melville Unfolding (2008)—and in various editions, chapters, and essays, John Bryant theorized on fluid-text analysis and offered protocols for editing fluid texts. (See Bibliography.) In them, he argues that our scope of literary interpretation has been limited to a reductive and static notion of the texts that constitute a written work. We typically imagine a written “work” to be a single, original, textual object. A broader perspective conceives of “work” as the energy a culture (represented by writers, editors, adaptors) puts into delivering “versions” associated with an originating title into the public sphere, over time. The study of fluid texts aims to make sense of this flow of cultural energies evident in revisions and versions.

Naturally enough, scholars, critics, and general readers alike desire a single, reliable text of, let’s say Typee, Moby-Dick, or Billy Budd, if only to have a common textual playing field for interpreting those works. The assumption has been that the author “intended” one version when in reality a literary work invariably evolves through multiple and meaningful versions, regardless of a single unifying, “final” intention. For whatever reasons, writers re-write, editors modify or expurgate, and adaptors reshape a text to fit their own interpretations. Not surprisingly, though, editors and publishers generally shape their efforts to reduce the complex evolutionary history of a fluid text in their publication of single, presumably “definitive” editions. MEL’s aim, however, is to offer an editorial alternative that enables readers to experience Melville’s evolving texts. To gain this broader fluid-text perspective, we need new ways of editing to supplement single-text editing practice, and digital technology facilitates the fluid-text editorial process.

Traditionally, scholarly editing focuses on creating a single reading text that represents the editor’s conception of a single moment of the author’s “final intentions” during the production of a written work. This moment of intentionality can vary; it might be the period of time during which a writer submits a fair copy manuscript to a publisher for further copy-editing or when the writer returns corrected page proofs, or (much later) supervises a subsequent revised edition. Through the critical editions they produce, scholarly editors play a fundamental role in supplying information about the single moment of intentionality they choose to represent in their reading text. The reliability of their editions resides in the editors’ forthright announcement of their conception of intentionality and justification for their reading text. These explanations include their display of the textual fluidities surrounding the work and discussions of their decisions to choose one variant (textual fluidity) over the others.

But given the limitations of print technology, these vital arguments are invariably placed in a Note on the Text and a Textual Apparatus, including endnotes or tables, consigned to appendices of scholarly editions, and the editions themselves are generally designed for library use only. In time, as publishers reprint the reading text of a scholarly edition for circulation beyond institutional libraries to the general public and schools, they omit the textual note, endnotes, and apparatus, so that the explanation of the original edition’s editorial process and the information concerning revision and versions effectively disappears. With that disappearance, readers are also deprived of valuable information related to the history of the work as a fluid text.

MEL’s fluid-text editions are designed not only to give fuller and more immediate access to this invaluable information and to verify and augment that textual record but also to reconfigure the data so that readers can navigate Melville’s revisions and the versions of his works as they read MEL’s reading texts.

The idea of the fluid text is nothing new to textual scholars. Modern scholarly editors have always addressed the challenges of fluid texts; witness the editing of the Bible, Shakespeare, Whitman, or Joyce. But conventional scholarly editing tends to view textual variation as a sign of errors to be corrected rather than evidence of revision. The “clean” reading text that traditional scholarly editions typically promote marginalizes variants, mixes versions, and masks revision. The Northwestern-Newberry (NN) edition of The Writings of Herman Melville (1968-2017) is a case in point. The edition’s goal is to provide a reading text that represents the editors’ conception of Melville’s final intentions before he submitted his work to editors and publishers. This “private” text is achieved by selecting one version of the work as copy text and then emending it, based on variants from other versions. By design, the NN’s mixed-version or “eclectic” editions offer an idealized Melville unconstrained by editors and audiences.

The concern from a fluid-text perspective is that editions that mix versions also mask the reality of the authorial and non-authorial revision processes that produce the versions that give us insight into the processes of creativity, publication, and adaptation. But rather than supplant the NN edition—whose scholarship has been instrumental in identifying the versions of Melville’s works—MEL’s digital, fluid-text editions showcase Melville’s revisions, through sequential versions. We use standard protocols of transcription, collation, diplomatic display, copy-text, emendation, and annotation to edit each version of a Melville work from scratch. We also settle on a minimally-modified “base version” of the work we are editing, which we use as a platform for linking to revisions in the work’s other versions. This kind of digital navigation enables readers to examine a range of intentionalities: Melville’s shifting intentions in manuscript, his final intentions or “private” version of a work released to publishers as well as the intentions of editors in shaping his texts, and adaptive versions that continue to reshape his texts in public.

A work like Billy Budd, which Melville never saw through the press, is particularly demonstrative. This fluid text exists as a complete but unpolished manuscript that contains thousands of revisions, each representing the writer’s evolving or oscillating intentions. A work like Typee which exists in manuscript, two first editions, and two revised editions, shows that Melville’s intentions shifted not only at the time of composition but also over a period of forty years. A work like Moby-Dick, which exists in two first editions also demonstrates that editorial forces—evident in the expurgations of Britain’s The Whale—represent external forces that inflect, even censor, the writer’s words. And when a writer like Melville or his individual works—like Typee, Moby-Dick, Billy Budd, and others—gain international recognition and capture the popular imagination of generations extending beyond Melville’s lifetime, we find additional versions of these works in adaptation and translation, which represent a culture’s need to revise the very texts they idolize. To authorial and editorial revision, then, the scope of MEL’s fluid-text editions extends to adaptive revisions in stage, film, opera, fine arts, music, and even the “meme.”

If we want to study not only the private and but also the social processes that impel a writer like Melville, his editors, or his international adaptors to revise Melville’s fluid text, we need fuller access to a fuller scope of versions of a work (in manuscript, print, and adaptation) and a fuller embracing of revision as an authorial, editorial, and adaptive process. The access and embrace cannot happen without archiving the versions and editing them. Digital editing offers approaches to editing that facilitate our understanding of fluid texts.

Digitization allows us to archive the versions, but subjecting readers to a “data dump” of manuscript and page images is useless without providing users the means of navigating the texts of these versions and ways to highlight their textual fluidities. One fundamental problem is that revision is an invisible phenomenon, and a revision only becomes visible when we edit it into visibility. For example, in his 1845 working-draft of Typee, Melville typically revised the inflammatory word “savage” to the culturally neutral “native” or “islander.” He and perhaps his brother Gansevoort made further, similar cultural and stylistic revisions as Typee was going into press in England in 1846, and, later that year, Melville expurgated the book for American readers. This range of revision in Typee represents the private and social forces that shaped Melville’s politics and creativity. But to “read” Melville’s savage —> native revisions, not to mention all of the rest of them, requires us to transcribe the manuscript, collate it with the print versions, identify revision sites, determine revision sequences, and compose explanatory revision narratives.

MEL designed its editing tool TextLab to facilitate the transcribing and coding necessary to address these editorial challenges. As a partner project in the University of Chicago’s CEDAR digital initiative, MEL also uses the OCHRE database platform to curate texts and images. In turn, we use Performant Software’s critical editing platform FairCopy (formerly Juxta Editions) to develop interfaces that access MEL data in the OCHRE database and enable the navigating of the versions. TextLab automates the labor-intensive TEI coding of revision sites, thus making the important and intellectually stimulating process of manuscript transcription available to scholars, critics, and students. TextLab also automatically generates a diplomatic transcription and base version of a given manuscript leaf, and it enables the creation of revision sequences and narratives. In FairCopy, editors can emend TextLab transcriptions, create textual and contextual notes, collate versions, and create further revision annotations. Both tools along with MELCat, our archival content manager, work together in an integrated platform to give readers access to the versions of a fluid text in a reliable format. Because MEL’s textual core exists in the highly atomized OCHRE database, MEL editions exist on a platform that assures continued sustainability and enhanced potential for interoperability with other Melville projects or digital sties of all kinds. For further technical details, visit our [MEL Toolkit].

Different Melville works pose different challenges in coding and navigation. MEL’s first step, then, was to create three “model” editions—Moby-Dick, Battle-Pieces, and Billy Budd—to address these challenges, which are discussed in more detail in Versions of Melville.

John Bryant (2021)