The French philosopher Jacques Derrida began his career in what can now feel like a distant age, a time of ink pens, typewriters, and manuscripts of paper stacked inches thick. By 2004, the year he died, that era had ended. The age of the computer had dawned.
For the father of “deconstruction,” an approach to language that emphasizes its fluidity and uncertainty, the jump from paper to computers was a fascinating change, and one that he embraced both personally and intellectually. By the early 1990s, he was doing his own writing on his “petite Mac,” an early Apple computer, and he used it to write about the seismic shift computer writing had created in what it meant to write, draft, revise, and publish.
“With the computer, everything is rapid and so easy,” he wrote in “The Word Processor,” published as a chapter in the book Paper Machine. “You get to thinking you can go on revising forever … you no longer retain the slightest visible or objective trace of corrections … everything negative is drowned, deleted; it evaporates immediately.”
That evaporation struck Derrida as a “provocation for ‘genetic criticism,’” a branch of literary scholarship that analyzes not only the final, published versions of literary works but also draft materials, notes, and edits. Without a record of changes, what would be left for analysis but the final text? And Derrida’s own drafts, edits, deletions, all recorded on his beloved Macintosh computer — those would evaporate as well, lost to both critics and time.
So Derrida thought, anyway, and there was no reason for him to think otherwise, given the state of technology and the average understanding of a computer user in his time. But some 20 years later, a small team of researchers, including UT Austin assistant professor Thorsten Ries, have plenty of proof to the contrary.
Trained in genetic criticism as a student, Ries now specializes in the digital humanities as a faculty member in UT’s Department of Germanic Studies. He is also an expert in digital forensics, a set of techniques and practices focused on accessing, identifying, and analyzing the ghosts left on a hard drive or in a data file. When it comes to computers, in other words, he can find what once was lost.
This background earned Ries a spot on a research team working to recover deleted text fragments from the archived Macintosh of Jacques Derrida himself. Working in collaboration with the French scholars Aurèle Crasson (ENS/ITEM), Jean-Louis Lebrave (ENS/ITEM), Jérémy Pedrazzi (ENS/ITEM), and Nathalie Léger (IMEC, Caen), Ries has successfully recovered traces of Derrida’s digital writing process that — unknown to the philosopher — were preserved on his computer’s hard drive and in his text files, establishing a line of potential research and criticism that wasn’t possible before.
In the researchers’ new papers, now pre-published on a French open-access platform for scientific research documents, Ries and his collaborators demonstrate just how much digital forensics can offer the humanities — and how much material may be lost for good if old computers and other digital storage media are left to gather dust on archive shelves. The findings are an accomplishment in themselves, but Ries also hopes they’ll serve as an example of how much the field of digital forensics can offer the humanities both now and in the future, and as a wake-up call to his fellow scholars.
As Ries puts it, “this is something we humanities scholars and archive subject specialists have to do ourselves. We must become experts in historical digital forensics and born-digital preservation, or we risk losing the born-digital records and cultural heritage of our time.”
Growing up in Northern Germany as the son of an engineer, Ries had two seemingly unrelated interests. He knew he wanted to study the humanities, particularly literature, and he was fascinated by the early personal computers then in vogue.
“Despite always having humanities ambitions, I was still dabbling with very early personal computers, like the Sinclair ZX81, Commodore 64, and Amiga,” Ries says. “So I have a weird kind of early contact with material computing history that I just lived with during a formative period of my life.”
In that earlier era of computing, it was difficult to foresee how technology would revolutionize every part of our modern lives. The ZX81, first released in the early 80s, was a small black box with a simple membrane keyboard and no moving parts. To use it, you plugged the box into a television set and saved simple programs on audio cassette tapes. While some writers and other cultural producers were using early word processing systems, paper was still their primary medium. But as the sophistication of computer systems increased, so too did the number of people switching away from pen and ink altogether.
That shift posed, and continues to pose, a challenge to scholars interested in manuscript studies and archival preservation. When everything’s digital, what happens to drafts, to edits, strikethroughs, and comments in the margin?
“Many of my colleagues in genetic criticism thought for a long time, ‘this is going to end. We are dinosaurs, and we’ll just do this thing with manuscripts as long as we can and then we’ll be extinct,’” Ries says. “But I knew that there was so much more to these digital drafts. My colleagues in philology thought everything we were doing would end because all of the material would be gone, but it wasn’t. It’s not.”
Ries’s background in computing, his work in e-learning development during his student years, and his encounter with Matthew Kirschenbaum’s influential book Mechanisms: New Media and the Forensic Imagination led him to believe it was possible to find and preserve draft materials buried underneath the zeroes and ones of text files and operating systems — to find a way, he says, where there didn’t seem to be any before. But the difficulties were and are daunting. Technology degrades — the magnetized hard drives used in early computers age particularly quickly — and pieces of hardware get lost or damaged. Tech know-how gets lost, too. As we chase the next innovation, older systems are ignored or forgotten. File types become unreadable; applications and operating systems no longer run.
These challenges further compound when dealing with computers from the era of Derrida’s early Mac. For newer systems, especially those from the internet age, forensic techniques have been developed in step with technological updates. There are established tools for gaining access to contemporary hard drives and data files, not because academics are interested in them but because law enforcement needs to be able to investigate crimes committed using modern technology. The techniques developed by investigators and data analysts can be used on modern systems to track disinformation or simply to restore a locked hard drive. They can also be used to salvage literary and cultural materials, though that’s usually low on the list of priorities. But for older computers like the ZX81 or the original Mac desktops, those techniques weren’t developed at all. At the time such systems were widely used, nobody had any concept of cybercrime, much less the need for digital preservation.
It’s the combination of all these challenges that made the Derrida project so interesting to Ries and his collaborators. The team was drawn to Derrida’s digital archive because of their interest in his philosophy, his famous treatise Archive Fever, and his claims against genetic criticism, but also by the possibilities his “petite Mac” offered more broadly. If they could find lost draft material on this hard drive, it would prove the possibility of finding such draft materials on many others lingering in archives around the world.
“The Derrida project is a huge opportunity for the computational humanities,” says Ries, “in no small part because of where we started. We didn’t have tools for this kind of operating system, but we were able to develop methods to recover draft material from old Macs and recover draft fragments of the writing process from MacWrite Pro files.”
In order to understand how the researchers were able to pull those pieces of material from the Mac, it’s helpful to have a basic understanding of how the computers of that generation saved, accessed, displayed, and deleted data. When saving a file in, say, MacWrite Pro, the software Derrida often used, or in another text editor, the computer saved it in binary code — the classic zeros and ones — to a physical place on its hard drive. The operating system, such as Mac OS, kept a map of the drive that allowed users to find that place and access the file again. But when they deleted a file, it wouldn’t actually be erased, at least not immediately. It was the directory information, the map that led to the file, that disappeared. The file’s content, encoded in those binary bits, could linger on a hard drive indefinitely until overwritten.
A similar thing could happen with the deletions and edits inside that document, too, depending on what kind of software was used to create it. In Derrida’s case, Ries and the rest of the research team theorized it was possible Derrida’s hard drive and his MacWrite Pro files still contained fragments of deleted or edited text that could offer clues about how his thinking and writing had developed. With the blessing of Derrida’s heir and the archive, they set to work proving it.
First the team went to Caen, France, where Derrida’s Mac is archived at the Institute for Contemporary Publishing Archives (IMEC), and made an exact copy of the computer’s hard drive. This copy, or bitstream-preserving image, duplicated every zero and one still in the machine. Working with the copy allowed Ries to cut up pieces of code for further analysis without risking any damage to the original drive’s contents.
Ries was then able to use modern forensic tools, adapted to the idiosyncrasies of this outdated technology, to identify and isolate individual MacWrite Pro documents for further study. But he still had to find a way to sift through those files for deleted or edited materials, like panning gravel for gold.
“In order to find any deleted data still left on the drive, you have to know where the not-deleted data is, because at this point it all looks the same,” he explains. “For some kinds of software, there are tools that isolate that with one command, but those tools did not work for this old Mac system. Nobody ever wanted to know this, so they didn’t make a tool for it. In the end, I had to create a very brute-force method — I call it the ‘hole-puncher’ — to separate the deleted data from the not-deleted data.”
The brute-force method worked. The researchers were able to identify multiple deleted files and file fragments with hidden edits and deleted passages, the equivalent of a written draft with strikethroughs visible in ink. It was exactly the kind of information a literary analyst working in genetic criticism looks for.
The results are a win both for the research team and for those digital humanities scholars and archivists pushing for greater adoption of digital forensic methods. It’s also fuel for the bigger fight: Convincing more humanities scholars and archivists to learn and apply digital forensics in their fields.
“What this study shows us is there is a need for humanities scholars to develop and maintain digital forensic methods in order to access historically and culturally relevant historical material,” Ries says. “This case is so interesting because digital forensics didn’t exist when this material was written, but even for the systems we are using now and for which digital forensics is very much alive, the methods will age. If we as scholars want to preserve our digital cultural heritage, we need to preserve the methods and the knowledge about this historical technology. In the same way that we have experts for certain types of manuscripts, we need people who are able to access and understand this kind of data and who know how to preserve it so that it can be studied and analyzed.”
The alternative may well be what Derrida predicted and Ries’ colleagues once feared: The end of genetic criticism, or, less apocalyptically, the loss of troves of digital literary and historical material. Archives are already working to stave off this dark future — Ries and other scholars are working with many to develop the techniques and workflows needed to preserve aging hard drives — but the challenges will continue. Modern encryption techniques, hard- and software diversification, and obsolescence introduce new practical and ethical considerations, and the cloud computing revolution may make locating any given document, let alone tracing its evolution, that much more difficult.
But there’s plenty of reason for optimism, too. Young scholars are studying digital archives, including those of still-living authors. Archivists’ know-how is increasing as digital preservation techniques are taught in information schools. UT’s own Harry Ransom Center is a pioneer in this field, Ries says, and many other archives and institutions are paying attention, eager to learn how to keep their materials accessible for generations to come.
“Working in this kind of forensics and digital preservation is a chance to do something very relevant that is useful in the future, even necessary in the future,” Ries says. “We have no chance to avoid this.”
It’s a shifting, unpredictable landscape not unlike the one Derrida described in “The Word Processor.” Just because materials can be saved from digital vaporization doesn’t mean they all will. Think of the photos you’ve likely lost in the last 20 years, still stuck on corrupted USB thumbdrives, or an essay draft lost in a hard drive crash. Our digital lives are haunted by ghosts, and every technological evolution creates more. But thanks to the work of digital humanists and archivists, some may walk again. The drafts — Derrida’s included — will live on.