Posted on behalf of John Bryant, Director, Melville Electronic Library, and Hofstra’s Digital Research Center
On April 7, 2016, just a week before the Society for Textual Scholarship conference in Ottawa, Arthur Koestler biographer Michael Scammell disclosed in his New York Review of Books article “A Different ‘Darkness at Noon’” fascinating editorial details about the publication of Koestler’s famous Soviet show-trial novel, which appeared first in 1940 in English, and not in its original German.
What had triggered Scammell’s article was Matthias Weß el’s discovery in 2015 of the novel’s original German typescript hidden away in a Swiss publisher’s files for 75 years. Koestler had composed his novel in German as he was eluding German and then French authorities, and had his partner Daphne Hardy, an English art student, to draft a translation into English. Neither Koestler nor Hardy was particularly adept at English or the art of translation. Their English translation made it to Koestler’s London publisher Jonathan Cape, but the German original did not. Editors at Cape considered commissioning a better translation but went ahead with the inadequate Hardy-Koestler version. No matter, the novel became instantly successful and during the war and postwar eras the English translation was itself translated into over thirty other languages. Using the Hardy-Koestler English translation in 1944, Koestler (with some assistance) translated the English back into German.
Scammell goes on to compare the newly discovered German Darkness with its 1940 English translation and Koestler’s 1944 retranslation, finding the English version upon which all translations are derived to be seriously flawed—including mistranslations, dropped texts, and missed linguistic opportunities—and he concludes with an emphatic call for a new translation of Koestler’s Darkness at Noon that will do justice to its “aesthetic qualities” and “offer a richer and more nuanced account of Koestler’s complex narrative.” What is needed, he says, is something akin to the cleaning of a painting to remove the textual grime (let’s say) and restore it to its original coloration.
As I made my way to Ottawa’s STS conference to join the Committee on Scholarly Editions panel on “Role of the Scholarly Edition in the Digital Age,” I could only concur with Scammell that a new translation was needed, but equally needed, I felt, was an edition of Darkness at Noon, or more to the point a fluid-text edition. A fluid text is any work that exists in multiple versions because of authorial, editorial, or adaptive revisions. And I could tell immediately that the revision narrative Scammell related about Koestler’s novel made it a perfect instance of a fluid text. An edition of Darkness’s versions would permit other scholars to create critical and interpretive revision narratives of their own. I was prepared to ditch my prepared remarks for the panel and dwell only on Koestler and on problems derived from Scammell’s call: would not the creation of (in his view) a better translation of Koestler create (in my view) simply another version of Koestler, a further evolution of the fluid text?
Nor do I imagine that all would agree on my call for a fluid text edition. Koestler intended his original and now that we have it, that document is what German scholars should edit, and let translators do what they will. But no translator of 2016 will be able to restore the 1940 English translation to something cleaner, more nuanced, without recognizing that translations are original creations themselves, that they reflect the moment of their creation and the agenda of the translator, and that the seventy-five years of reading that have shaped the way we read Koestler in English (not to mention in 30-40 other languages) will also shape the way we retranslate Koestler again into English. A new translation of Koestler will not fix his text; it will only make it more fluid. Moreover, preserving through editorial processes the array of translated texts would give us unprecedented access to a textual field representing Cold War era trauma. And besides, translation itself is a critical act fraught with interpretive variation. The “betterness” of Scammell’s hoped for better translation is something we can only know by the focused comparison of it to all other flawed translations. A scholarly edition seems all the more needed, or better, a digital critical archive.
From a fluid-text perspective, Koestler’s work has its life first in English, then in dozens of other languages, and then in Koestler’s retranslation back to German; and it is this life of the work—evident in the totality of its versions—that needs both archiving and editing. A good editorial thought game is to imagine how such a fluid-text edition would work. The challenge of fluid-text theory lies in its embrace of revision and their related versions, and in particular its valorization of non-authorial adaptive versions, such as translation. Because Koestler’s German retranslation of his own English translation is a kind of revision verging on original composition, the Koestler materials complicate any attempt to distinguish authorial and adaptive versions definitively. Each of these two versions are authorial adaptations of an original. Theoretical concerns aside, any practically conceived fluid-text edition of Darkness at Noon would have to be digital, and obviously the edition itself would require some sort of digital archive that could contain the various versions of Koestler’s work, including the thirty or so translations, which we would also want to “back translate” to English, and the new translation Scammell calls for, not in itself the restoration he imagines, but another version on its own.
As exciting as the newly discovered Koestler original may be, and as useful as it may be as a problem for fluid-text editing, I realized, as I landed in Ottawa and had a delightful conversation about Donald Trump with my cab driver, that I could not touch upon Darkness at Noon without explaining it all and would certainly overextend my time on the panel, as I am surely doing here, by over-extending whatever space is allotted me on this blog version of my comments. So permit me this Koestler intro to my actual panel comments, which as it happens, pertains precisely to the kind of digital critical archive that I feel modern scholarly editing in the digital age might aspire to.
My first thought in addressing the panel’s topic—“The Role of the Scholarly Edition in the Digital Age”—was that it seems to quietly signal a sea change in digital editing. In previous years, and for decades, the topic would have been expressed as an inversion: “The Role of the Digital in Scholarly Editing.” That is, our earlier concern was how digital technologies might be used to facilitate editing. In my own attempts to theorize about fluid-text editions and to create practical examples of what they might be, I worked up editorial protocols for editing revision and digital technologies for navigating revisions through digitized revision sequences and revision narratives, culminating in the development of the digital editing tool called TextLab (http://mel.hofstra.edu/textlab.html) for the Melville Electronic Library (http://mel.hofstra.edu/). In short, I was enlisting digital means to solve certain problems of critical editing, in particular the editing of revision as a phenomenon. And my efforts—theoretically, practically, digitally—still continue.
But the panel’s announced topic reveals new growth in our conception of the relation of digital and scholarly editing. The idea that scholarly editing can give shape to the admittedly messy aspect of The Digital, or Digital Humanities, or (as I prefer to call it) Digital Scholarship is one that is rarely articulated and yet its articulation here should mark an important turn in the now decades-old conversation about the relation of the mere mechanisms of Digital and the higher ideas and standards under attack in traditional Scholarship. I applaud the intent of the panel’s topic to get us beyond all that. In my comments before the panel my underlying argument is that scholarly editing can indeed play a role in the digital age because both scholarly editing and digital technology are mechanisms for promoting shared modes of critical thinking, and I believe these shared modes are evident in an editorial genre I call the digital critical archive. I also argue that the principles and practices of scholarly editing are an effective template for constructing such an archive.
The scholarly edition readily lends itself as a model for the digital critical archive. A scholarly edition consists of a reading text and a textual apparatus. In digital terms, the reading text is a visualization of an editorial conception of a text that represents a version of the work that the editors have announced they wish to achieve. The textual apparatus is the database of textual variants from which the visualized reading text is derived. If we add to the reading text further visualizations, such as annotations that provide historical, social, cultural, biographical contexts, identification of quotations, and the edited texts of letters and sources, we can also see that the database of digital objects must grow beyond the standard list of textual variants to include an encyclopedia of materials related to the edited work. This digital version of the textual apparatus becomes the submerged iceberg database that supports the tip of the visualized reading text. Thus, the traditional textual apparatus is made to grow into a critical archive that contains the texts of letters, biographical materials, and sources, themselves edited for the archive as well as illustrations, photographs, images of art works, and the like. Digital technologies enable and shape this kind of transformation from apparatus to archive, but scholarly editing shapes the archive itself. Nothing I can find in current or past editorial theory would deny this potential in the textual apparatus; in fact, the variable size of a given scholarly edition’s textual and historical notes and its “related documents” attests to this possible growth. Only the limitations of print have prevented editors in the past from adding more data to their textual apparatus and back matter.
If textual editors are to take the lead in shaping digital critical archives, they need to be cognizant of the shared values of both the editorial and digital enterprises, and their principal contribution to critical thinking. Digital strives to make visible what is otherwise invisible in cultural and writing processes. Similarly, editing makes visible otherwise invisible processes of revision and textual evolution that are evident in otherwise hidden versions and variants. Together in a digital critical archive, both modes of thinking combine to give users not only access to otherwise concealed materials but also opportunities for controlled interaction with the scholarly edition itself. A digital archive is, then, the space where editions are made through collaborative efforts by teams of editors, who learn to become editors by editing.
One potential pitfall in digital development is the temptation to aestheticize the database, to make visualizations that dazzle but finally fail to engage readers in critical thinking.
The Mapping of the Republic of Letters site generates appealing and varying visualizations of the lines of correspondence, decade by decade throughout the 18th century, to convey networks of influence. But this promising critical archive does not allow users to click on a line or bubble, see the image of a letter, and read a searchable transcription of the letter. In today’s parlance the “macro” visualization gives you a distant reading of the database but it does not permit a close reading of the “micro” content. This is because the letters have yet to be edited. I mention this example not to dismiss digital technology but to argue that scholarly editing theory and practice can be used to design digital archives in ways that balance macro and micro levels of thinking.
In building the Melville Electronic Library (MEL), we needed a tool that could both build our scholarly editions as a reliable “textual core” of all versions of all Melville works and display those editions. When we began designing TextLab, as a tool for editing revisions in working draft manuscripts and for collating them with print variants—whether we knew it then or not—we developed both macro and micro features together. With TextLab, MEL editors working on Billy Budd mark-up images of manuscripts in revision and transcribe the full texts of each leaf.
On the macro level, TextLab generates a base version of the working draft [SLIDE 3] and a diplomatic transcription of each leaf [SLIDE 4]. In addition, on the micro level, editors can select a set of revision sites, create a revision sequence for them, and compose a revision narrative that explains each sequence [SLIDE 5]. TextLab’s database also contains multiple revision sequences and narratives for comparison of different options for any given revision sequence [SLIDE 6]. With TextLab, MEL editors also collate the two first editions of Moby-Dick; the tool highlights variants and allows the user to create revision sequences and narratives for each highlighted variant [SLIDE 7].
In designing MEL as a critical archive, we also wanted to give readers workspaces where they can use the digitally-edited scholarly editions to generate new scholarship. This, too, is a logical extension of the traditional textual apparatus. An abiding principal of any critical edition is that any editor can come to the apparatus and reconstruct the reading texts of variant versions of the edited work. Obviously, the print volumes of critical editions cannot supply the tool that would enable such reconstructions, but such an editorial lab is possible digitally. Of course, since the digital critical archive would necessarily house all versions of a work, such a lab is effectively obviated. Nevertheless, a comparable digital workspace, based on the textual apparatus, would allow for other editorial and critical activities: users could assemble, compare, and analyze the texts and images associated with whatever the archive collects. Again, scholarly editing shapes the archive.
Similarly, the workspace permits MEL associates to collaborate with editors on the building of MEL’s scholarly editions. By encouraging users to edit in a controlled workspace, we are teaching our craft, instantiating standards, promoting a kind of critical thinking based on the inspection of evolving material texts. Accordingly, we also recognized that we needed to develop a system of controlled vetting of contributions to the archive and editions. A digital administrative site is therefore needed for any digital critical archive to assign credit to contributors and to ensure consistency of coding, quality of display, and high editorial standards. Through the development of editorial, critical, and interpretive workspaces, we provide a field of discourse for a digital version of the kind of editorial team that has existed for centuries: yet another role editorial traditions can play in developing the textual cores of digital critical archives.
I’ve spoken here of the mutually supportive nature of digital technologies and scholarly editing process in promoting the kinds of critical thinking that can bring balance to the macro and micro levels of textual scholarship. I’d like to conclude by addressing some problems that I feel the Committee on Scholarly Editing’s white paper—a delightfully diplomatic document—raises.
In their white paper and practice, CSE has successfully articulated the standards by which it evaluates digital scholarly editions, and in one section of the document, it rightly looks for evidence of “longevity,” or in our current digital parlance, “sustainability,” in a given digital project. Nothing could be more important than the addressing of this matter. It is a truism in digital technology that a website that no one uses dies, and this fact is a source of anxiety. Scholarly editors are inured to the death of their scholarship. They assume that their hard work will be put on a shelf and rarely consulted. Their consolation is that the solid, material, published volume or set of volumes will outlive them. But a digital edition is subject to shortness of life: it can fold at the flip of a switch, or more likely the degeneration of an image, the demise of a server or some underlying software. That said, what sustains a digital scholarly edition or critical archive and gives it “longevity” are mechanisms that ensure continued use, in particular interactivity, collaboration, and community.
These important but somewhat ineffable features are not necessarily things that CSE can measure or legislate. Even so, there is abundant precedence of all three in scholarly editing. As already noted, textual editors presume that future editors and scholars will consult and want to interact with, or argue with, their published reading text and apparatus. With TextLab, users can perform this kind of critical interaction with the edition by submitting their own optional revision sequences and narratives for inclusion in the edition. If the administrators of the edition are not open to such continued augmentation, they can provide separate editorial spaces where users can experiment. In this way, MEL has also used TextLab to create individual editorial projects so that students can hone their transcription skills. And with other tools developed in consultation with MEL, such as Annotation Studio, students and scholars can learn how to create succinct and relevant contextual annotations of MEL’s editions.
Through this kind of interaction, users collaborate on editing, and through collaboration they become a community, which is surely the best antidote to the Death of Scholarship Syndrome. Without a team of collaborators, there is no archive. Happily, critical archives are generally performed at colleges and universities where people under the age of 25 congregate with young scholars in their 30s, and even retirees such as myself over the age of something or other. It is vital for any critical archive’s sustainability to attract and train younger scholars who resist the death of the edition by growing in their learning to replace their mentors and continuing to keep the archive and edition alive through interaction and collaboration, and through the community that they constitute. These values are not alien to scholar editing. Even though print editions have generally been created by teams, CSE does not necessarily have the means to judge the viability of a digital editorial project’s interactive and collaborative design or the projected longevity of its community. That is a tough nut to crack, but one that must be cracked nonetheless.
Finally, I’d like to address a problem of collation and adaption that brings me back to the problem of editing Koestler; these are problems resolvable through the mutual efforts of scholarly editors and digital scholars.
Darkness at Noon’s textual history raises questions about the nature of versions. What kind of version is his retranslation? Is it a revision of an original or an original in its own right but linked to an earlier original? Do we edit the novel in German or English; which version of each language do we edit? Naturally, we must edit the original, the translation, and the retranslation together. Complicating matters is that translation a kind of “adaptive revision” of an original work into a different language for a different audience.
Scholarly editors do not generally focus on adaptation or translation in their editorial work. As non-authorized adaptations, translation is not an expression of authorial intentionality. But Koestler’s work is an imbrication of texts, over time, whose ontology is defined by textual evolution. If texts evolve through different states in important ways that alter our understanding of the text, and if critics want to know more about how circumstances inflect the evolution of a text from the earliest revisions to the latest adaptive revision, then translation and adaptation is a legitimate subject for editorial scrutiny. But the textual scholar’s primary means for detecting evolution is collation, and currently our standard collating tools are inadequate for comparing versions that are not symmetrical, such as adaptive revisions like translation, which present a variant at every lexical position in the text.
The problem of “nonsymmetrical collation” extends beyond translation to all forms of trans-genre and inter-medial adaptation, such as a film, which may adopt some texts word for word but delete, rearrange, or visualize others. How, then, do we collate the “nonsymmetrical”: an original text to the video of its stage adaptation, to an art work representation, a piece of music, or even a cultural meme like Melville’s white whale? If a digital critical archive collects such materials as evidence of the versions of a work, digital tools are needed to assemble, collate, and annotate these adaptive (audio, video, and memic) revisions. Textual scholars understand the protocols of collating texts, and digital scholars understand algorithms for linking, comparing, highlighting, and annotating texts. These variant modes of critical thinking need only work together to realize a nonsymmetrical collation tool. In the process not only will the scholarly edition shape the critical archive, but it can also push digital thinking into new modes of and machine/human interaction.
While Michael Scammell rightly calls for a new translation of Koestler’s Darkness at Noon, I would like to call for a community of textual scholars and digital scholars to come up with tools that address the problem of nonsymmetrical collation that will permit the editing of translation and adaptive versions of a work. Assembling such a community and building storyboards are easy enough to imagine; securing funding for development is not impossible; acknowledging that adaptation is a legitimate subject for scholarly editing also seems implied in CSE’s white paper. All that remains is what Ishmael calls for: “Time, Strength, Cash, and Patience.”