New Article: Material Cultures of the Digital

At long last, an essay I wrote pre-pandemic is now out. “Material Cultures of the Digital” can be found in the Cambridge Handbook of Material Culture Studies, a brilliant new collection of interest to bibliographers, book historians, media historians, and (I hope) digital humanists. I enjoyed many aspects of writing this piece, not least getting the chance to focus on more contemporary artifacts than I do typically, starting with a rumination on server farms located in nineteenth-century paper mills and expanding from there. I’m also particularly attached to the long inventory of digital-material artifacts midway through the essay.

Per Cambridge University Press’ open access policies for edited collections, I’m archiving a pre-print of my chapter here. This version does not reflect the edits made in response to the volume’s wonderful editors, Lu Ann De Cunzo and Catharine Dann Roeber, which improved the chapter’s structure substantially and added important connections to other chapters in the volume. The published version of this chapter, and the book, are available from CUP, while portions can be previewed through Google Books.

Material Cultures of the Digital

Introduction

When Google sought to expand its data center operations in 2009, the company bought an abandoned paper mill in Hamina, Finland. Built in 1953, the Summa Mill had been operated by Finnish pulp and paper manufacturer Stora Enso, but was closed in 2008 due to “a drop in newsprint and magazine-paper production” as “Newspapers and magazines are slowly giving way to web services along the lines of, well, Google.” According to Wired magazine, Google was interested in the mill because it “included an underground tunnel once used to pull water from the Gulf of Finland” to cool “a steam generation plant at the mill.” Google needed efficient and ecologically-friendly cooling systems, as well—not for a steam plant, but to cool the processors and other components in its massive servers. “Google’s Hamina data center is,” Wired declared, “the ideal metaphor for the digital age.”¹ We might recognize in Wired’s confident emblemizing of this factory a tidy narrative of new media in which the future necessarily subsumes the past, eliding the messier present.

While popular-press accounts of Google’s paper mill purchase tended to claim the purchase illustrated the digital replacing its preceding media, it might instead remind us of the ways material cultures imbricate across time. Google purchased a space designed to produce one kind of information medium because its design, only slightly modified, would well serve the needs of a new information medium. Through this process the tech giant collapsed “analog” and “digital” material histories into one entwined narrative. As Alan Liu contends, narratives of new media too often hinge on ideas of replacement or conversion, while “The better term is indeed ‘encounter,’ indicating a thick, unpredictable zone of contact—more borderland than border line—where (mis)understandings of new media are negotiated along twisting, partial, and contradictory vectors.” In this encounter, the data center inhabits the paper mill, and the paper mill shapes—at least partially—the data center, a “déjá vu haunting of new by old media.”²

Such encounters are everywhere in the histories of hardware and software. For instance, early internet companies like Prodigy and America Online relied on the infrastructure of the US Postal Service to distribute CD-ROMS of their software cheaply to potential users, as well as on the infrastructure of telephone lines to connect those users to the World Wide Web. More recently, Nicole Starsielski has traced the global fiberoptic network, a planet-spanning series of “winding cables the size of a garden hose” that run underground and undersea to “transport 99 percent of all transoceanic digital communications, including phone calls, text and e-mail messages, websites, digital images and video, and even some television.” Like Google’s paper mill-cum-server farm, the extensive infrastructure of the global internet often “follow…the contours of earlier networks, layered on top of earlier telegraph and telephone cables, power systems, lines of cultural immigration, and trade routes.”³ We cannot consider the internet’s history or present without also considering the objects and people who have constituted its network, or indeed the previously existing networks the internet runs along or rides atop.

A popular aphorism in the computing profession reads, “THERE IS NO CLOUD: It’s just someone else’s computer.” This phrase appears on laptop stickers and t-shirts, and on a bevy of coffee mugs found in computer science departments, tech conventions, and the offices of internet startups. This aphorism offers a wry reminder that for all the airy metaphors we use to discuss computation in the internet age, data doesn’t drift around the atmosphere. A computer is every bit as material an artifact as an abacus or printing press, as are the servers, cables, monitors, and other objects that make up the computational ecosystem. According to Tung-Hui Hu’s A Prehistory of the Cloud, “The cloud is both an idea and a physical and material object, and the more one learns about it, the more one realizes just how fragile it is.”⁴ It is incumbent on scholars of material culture across disciplines to learn about this constellation of objects we call the cloud if we are to account for this age.

Nevertheless, we often talk of the newly ubiquitous “apparatus” of the digital age—to borrow N. Katherine Hayles’ word in Writing Machines—as if they were objects made of mist rather than metal, plastic, glass, and wire.⁵ In fact, the phrase “It’s just someone else’s computer” somewhat diminishes the full materiality of the internet. What we call “the cloud” comprises a lot of computers and servers, linked by routers, switches, and overland or undersea cables: the net work of the world wide web. As of 2017, Google alone operated 15 data centers around the world, with many other smaller “cloud centers” and “caching sites” supplementing their operations. The largest of these facilities approach 1 million square feet, while reports about a new Google data center in the Netherlands claimed, “the company…contracted for the entire 62 Megawatt output of a nearby windfarm and ran 9,941 miles of computer cable within the facility.”⁶ The infrastructure for other major tech companies, such as Facebook or Amazon, is just as expansive, while a host of others employ smaller—but still extensive—server farms and data centers. While they exude, perhaps, an aura of corporate sterility off-putting to many scholars of material culture, such data centers comprise the backbone of digital culture and, though often unseen by users, leave enormous economic, environmental, and material footprints in the world.

Caption: A server room at CERN; Creative Commons image via Wikimedia Commons

We are faced with a strange and hazardous irony in the digital age: surrounded by a rapidly-proliferating domain of objects, increasingly central to our social, economic, and political lives, we fall back on a strange dichotomy of “digital” versus “analog,” eliding a the much more complex and interesting ways these two modes entwine and interact. Popular attitudes toward the digital vary widely and are often highly contradictory. Digital content is both unreliably ephemeral and stubbornly eternal; hyperlinks go dead rapidly, and thus cannot suffice as scholarly references, while embarrassing photos posted to social media in high school will surely derail presidential campaigns decades hence. Such contradictions arise in part due to the novel materiality of digital media, which is both widely distributed (i.e. pervasive) and rapidly iterative (i.e. impermanent). For scholars of material culture, the digital represents the fastest-growing domain of material culture in the late twentieth and early twenty-first centuries, both in terms of its literal materials and its broader cultural effects. Understanding and theorizing the material culture of the digital is thus one of the field’s most pressing mandates.

“Digital” was never truly a synonym for “virtual,” and each year the boundary between digital and analog materiality becomes more difficult to ascertain. 3D printing technologies manifest computational models in human spaces: everything from pop-culture tchotchkes to artificial human organs. Our digital network now connects a vast, growing, and uniquely vulnerable “internet of things” to the internet and to each other. The internet of things comprises the light bulbs, refrigerators, home security systems, thermostats, garage door openers, and other not-obviously computational devices we now can control from our devices, thus scattering the material culture of the digital across a wide swathe of twenty-first century culture writ large. Even more broadly, from our computers, tablets, and phones we summon people into action and spur a host of material objects into circulation. We click a button, for instance, and a worker begins hunting in an enormous factory for the item we have ordered, the first human being in a chain of humans who will close the commercial circuit. For scholars of material culture, then, the digital is essential to understand both as an assemblage of material objects and as a network that increasingly circumscribes the interactions of people and other material objects.

A Register of Digital Materials

To begin grappling with the materiality of the digital, we might compile a brief and necessarily incomplete inventory of computational artifacts:

• computer cases (desktop, laptop)
• monitors (CRT, LCD, OLED)
• CDs (CD-DA, CD-ROM, CD-R, CD-RW, VCD, SVCD), DVDs (DVD-ROM, DVD-R, DVD-RW, DVD+RW, DVD-RAM)
• CD and DVD drives (internal, external)
• mice (ball, roller, optical, laser, trackpad, 3D, ergonomic)
• sever
• network router closets
• circuits
• motherboards
• sound cards
• graphics cards
• RAM chips
• cables (SCSI, PS/2, 3.5mm, VGA, DVI, USB [many varieties], Firewire, Lightning, Ethernet, HDMI)
• hard disk drives (internal, external)
• flash drives
• solid-state hard drives
• punch cards
• floppy drives
• floppy disks (8”, 5¼”, 3½”)
• Zip Drives and disks
• Jazz Drives and disks
• keyboards (standard, laptop, flexible, portable, optical, mechanical, illuminated)
• printers (dot matrix, thermal, ink jet, laser, desktop, drum, line)
• laptop or monitor stands
• software packaging
• tablets
• mobile phones
• communications satellites
• chargers (corded, wireless, portable, bicycle dynamos)
• cell phone charging lockers (in airports and public spaces)
• external power supplies, speakers (internal, external, portable, smart)
• headphones
• smart appliances (light bulbs, refrigerators, coffee makers, stoves, garage doors)
• gaming consoles
• controllers (game pads, joysticks, remotes)
• watches (digital, smart)
• fitness trackers
• scales (digital, smart)
• televisions
• virtual reality glasses
• 3D printers
• laser cutters

I expect anyone who reads the list above will think of at least one artifact it misses. It largely cannot cover the vast array of objects now produced with embedded computer components: everything from automobiles to medical equipment to children’s toys to voting machines.

Our material culture is suffused in digital objects and even more, in another strata of objects designed for carrying, storing, protecting, decorating, or displaying digital-material artifacts: e.g. laptop sleeves, mobile phone cases, or computer decal stickers. Moreover, the design of many objects designed primarily for non-computational purposes have been nonetheless altered for the internet age. It would be nearly unthinkable to imagine a modern briefcase, for instance, that does not include a separate, padded compartment for a laptop computer or, increasingly, a car without a USB charging port in its front cabin. Fields such as design and fashion have been radically reoriented around the digital—and radically reorganized in practice around digital platforms—so that any account of their material culture must engage the effects of computation. Discussions of twenty-first century architecture, for example, likely must account for AutoCAD and similar programs, just as discussions of clothing design likely must include accounts of Adobe Illustrator or its kin. Around this amorphous construction of “the digital,” then, we find a penumbra of artifacts that exist because of software or hardware, or that manifest the affordances and limitations of software.

Perceiving Digital Materiality

The word “network” has become increasingly metaphorical, describing relationships among devices but also among people, but it was coined because connected systems—whether of wire, rails, roads, or other conveyance for information or people—we thought to resemble the connected ropes of a net, or the “net work.” It was, perhaps, easier to remember the material entanglements of computing when users were required to physically connect their computers to hard-wired networks using telephone and then ethernet cables, establishing a literal and visual link that could pull or even tear, severing one device from the collective. Increasingly, people connect to the internet’s network of servers wirelessly, but still through material devices such as laptops, tablets, and mobile phones that communicate with yet other devices such as wifi transmitters or cell phone towers. As our computers have disconnected from wires, the number of computer-centered or computer-driven devices in our daily lives have increased exponentially. The US Environmental Protection Agency reports that “the average American household uses about 28 electronic products such as personal computers, mobile phones, televisions and electronic readers (e-readers);” a good many of these products are, especially if produced in the past decade, in some sense “digital” (My emphasis).⁷ In becoming more portable, the digital has become for many essential to the minutia of both public and private life, even to bodily autonomy. Our mobile phones have become extensions of our memory and our digits. We are cyborgs in habit, if we are not (yet) corporally entangled with our devices.

Some of the specific sensations facilitated by digital objects are likely novel; how many people in previous generations ran their fingers along perfectly smoothed glass as regularly as a twenty-first-century smartphone user? Nonetheless, our digits manipulate matter when we interact with digital interfaces. Dennis Tenen describes the effects of our fingers on keyboards as initiating a long chain of material changes: “What originates from (1) the keyboard as the mechanical action of a switch becomes (2) an electric signal that (3) leaves electromagnetic marks in computer memory, which (4) morph into phrases of liquid crystal on-screen, leaving behind (5) letters that emanate outward as light.”⁸ Our interactions with the touch screens on phones or tablets seem to close this loop, encouraging the perception that our fingers interact directly with light-emanated letters. In reality, however, our fingers brush capacitive touch screens that send small electrical charges into our fingertips, completing circuits and sending signals to our devices about where their screens were contacted.

Paradoxically, the materiality of computation seems to become increasingly difficult to apprehend even as computation increasingly suffuses culture. The room-filling computers of the 1940s and 50s were literally and metaphorically hefty. The cumulative mass of twenty-first century computation would outweigh these machines by many orders of magnitude, but its ubiquity and even its interfaces serve to mask this reality. The graphical user interfaces (GUIs) through which we typically interact with computers deliberately obscure the machinery that creates and sustains them. Interfaces are largely organized around skeuomorphic representations of older media: we work from a desktop, we delete files through a trash can, we save files by clicking a floppy disk, or we browse the internet through windows and tabs. These skeuomorphs help communicate functionality to users, to make new and unfamiliar operations tractable, but they also serve to separate the objects we interact with from their own materiality—the simulation on the screen encourages us to look past the screen itself, along with its attendant components, and imagine something like a disembodied cloud of software rather than a warehouse full of servers.

Marlene Manoff outlines some of the dangers that follow when scholars misrecognize the digital space as immaterial. Writing primarily to librarians, she cautions them not to replace the term “collection management” with “content management,” because “When one calls collection management ‘content management,’ librarians are encouraged to think about content in the abstract, as if it existed apart from any particular physical embodiment…The term ‘content management,’” she continues, “suggests that we have somehow moved beyond mundane considerations of physical reality when, in fact, the electronic environment introduces a whole new set of questions about the material aspects of library collections.” One danger of making a mental shift from “collections” to “content,” Manoff argues, is that such a shift will encourage notions of surrogacy between digitized materials and their archival originals that can flatten the unique properties of either medium. As a result, “some librarians are rushing to identify funds that can be freed by canceling print subscriptions that are duplicated in electronic formats. Others are eager to jettison paper back-files in order to free up shelf space without much consideration of the reliability of the back-files or the digital archive.”⁹ By failing to fully account for the digital medium’s specific materiality, in other words, we place the past and future at greater risk. In my own research I have suggested that the “myth of surrogacy” stunts our historical and contemporary imaginations alike, shrinking the materials of the past to the size of a computer screen and, perhaps paradoxically, encouraging us to use digitized materials as if they had no unique properties to their medium.¹⁰

Digital Environments

As Jussi Parikka reminds us, we cannot separate the myriad objects of the digital world from the materials and systems underlying their construction, use, and eventual disposal: “Media and information technology are far from zero entropy mathematical dreams, and embedded in physical networks, afforded by hardware and hardwork – practices of mining, shipping, polishing, constructing, and then the other way round, when disgorging such machines.” Parikka insists “the materiality of information technology starts from the soil, and underground” before listing the minerals necessary for the creation of many digital artifacts:

Cobalt Lithium-ion batteries, synthetic fuels
Gallium Thin layer photovoltaics, IC, WLED
Indium Displays, thin layer photovoltaics
Tantalum Micro capacitors, medical technology
Antimony ATO, micro capacitors
Germanium Fibre optic cable, IR optical technologies
Platinum (PGM) Fuel cells, catalysts
Palladium (PGM) Catalysts, seawater desalination
Niobium Micro capacitors, ferroalloys
Neodymium Permanent magnets, laser technology¹¹

While it may seem obvious that computer hardware is material, common metaphors such as “the cloud” obscure the gritty realities of cobalt, tantalum, and niobium, and the human labor required to extract these materials from the ground. Creating, running, and maintaining our digital infrastructure takes a substantial environmental toll.

Moreover, the energy costs of our ubiquitous devices are even more substantial, and only growing. Social scientists Richard Maxwell and Toby Miller note in their book Greening the Media,

residential electricity consumption for powering ICT/CE [Information and Consumer Technology/Consumer Electronics] is also growing at unprecedented rates, accounting for about 15 percent of global residential electricity consumption by 2009. By 2011, upwards of ten billion devices needed external power supplies, including two billion TV sets, a billion personal computers, and cell phones, which reached five billion subscriptions in 2010, including 85 percent of the US public. In 2011, nearly three-quarters of the world’s population owned one, and threequarters of these accounts were held in the Global South. By 2009, about 40 percent of US homes had video-gaming consoles, which collectively consumed electricity at the same annual rate as San Diego, the ninth-largest city in the country. If media usage continues to grow at this rate, the IEA estimates that electricity consumption by electronic equipment will rise to 30 percent of global demand by 2022, and 45 percent by 2030.¹²

These predictions have largely proven accurate, save minor improvements made through the development of more energy efficient devices in the years since 2012. However, new environmental wrinkles have also emerged during this same time period. For example, the networks of people and machines mining Bitcoin, a “cryptocurrency” that exists entirely in code, are estimated to consume as much electricity as countries such as Ireland or Austria.¹³ The process of verifying electronic transactions, which is required to produce new digital currency, is often referred to as “mining.” These processes are often automated for speed, scale, and efficiency, set running on computers and servers twenty-four hours a day. It might seem a strange skeuomorph to call these operations “mining,” but just as mining for natural resources requires large expenditures of human and machine energy and can damage the natural world, technologies for extracting digital resources require substantial investment of time and resources, and take an environmental toll.

Our digital devices take a human toll as well. Mined minerals are shaped into computer components and devices in massive factories, many in the developing world, and often with poor records regarding employee health and happiness. Most prominently, in 2010 at least 18 workers attempted suicide at a complex of Foxconn factories in China, where, among other products, workers assembled Apple’s iPhone smartphones and iPad tablet computers. Accounts of conditions in these factories vary widely, but grim images of suicide nets, installed to prevent workers jumping off factory buildings to their deaths, circulated online and in broadcast media, soberly remind consumers of the human costs of building digital devices—though sales of Foxconn-made products has not seemed to slow as a result. Stories juxtaposing the hard labor of device assembly with the conveniences of digital access often deliberately recall the factory literature of the nineteenth-century and early twentieth centuries, which challenged readers to recognize the human costs of that age’s technical marvels. Frankly, this is not a reckoning well undertaken by either the general public or scholars: myself included, as I type this article on a device made by unknown workers under unknown conditions.

Concerns about material and economic exploitation attend the entire digital lifecycle. While as a category digital devices are ubiquitous in the early twentieth century, particular digital devices have very short shelf lives: new software often will not run on models only a few years old, new features convince users to upgrade, or older devices simply become unfashionable. Many companies prioritize “planned obsolescence” in their design and marketing, in order to keep consumers buying. Because older digital devices remain physical objects, however, they must go somewhere when discarded. In the best cases, components or even whole devices can be recycled or “up cycled” toward new uses: an old laptop becomes an interface for a home entertainment system, perhaps, or components from an old phone are reused in a different device. But old electronics all too often become simply waste, and often dangerous waste. Developed countries often ship “e-waste” to developing countries, where it can be stripped for parts and chemical components more cheaply and often under much more hazardous conditions for workers.

In other words, the digital or virtual environments we navigate on our screens are made of terrestrial stuff, and make lasting terrestrial changes. A stunning, comprehensive account of “the full stack”—another weighty, materialist phrase—of labor, technology, and economics required for humans to interact with a single digital device, the Amazon Echo, can be found in Kate Crawford and Vladan Joler’s essay and map, “Anatomy of an AI System: The Amazon Echo as an anatomical map of human labor, data and planetary resources.”¹⁴ This remarkable piece begins noting,

A brief command and a response is the most common form of engagement with this consumer voice-enabled AI device. But in this fleeting moment of interaction, a vast matrix of capacities is invoked: interlaced chains of resource extraction, human labor and algorithmic processing across networks of mining, logistics, distribution, prediction and optimization…each small moment of convenience—be it answering a question, turning on a light, or playing a song—requires a vast planetary network, fueled by the extraction of non-renewable materials, labor, and data.

From here, Crawford and Joler outline those extractions, connecting work mining the lithium reserves of the Salar in southwest Bolivia; to Athanasius Kircher’s 1673 invention of “the statua citofonica, or the ‘talking statue;’” to “hardware manufacturing and assembly processes in Chinese factories” and “exploited outsourced cognitive workers in developing countries labelling AI training data sets;” to the Victorian destruction of the trees that produced gutta percha for insulating telegraphic and other cables; to the shipping boats full of cargo containers that “produce 3.1% of global yearly CO2 emissions” today.

This summary unfortunately elides most of their essay’s points, but hopefully hints at the complexity of the material-social-technical systems they demonstrate as necessary for understanding a single computational device. Scholars of material culture could and should take on the responsibility of producing similar accounts of the many other devices and artifacts of the digital age. Doing so will require both the analytical capacities we typically associate with humanistic inquiry and the technical capacities that will enable scholars to understand interplays between hardware and software, devices and networks, or users and systems.

Soft(ware) Materials

Even less apparent than the materiality of hardware is the materiality of the software running on it. We too often overlook the intimate connections among hardware and software, strangely failing to link—intellectually, at least—our ubiquitous devices to the programs that run on them. But software is also a thing with a physical presence in the world. In a 2014 lecture, Matthew Kirschenbaum looked at software from fourteen perspectives to outline all the ways in which it is a “thing”: “software as asset,” “software as shrinkwrap,” or “software as epigraphy,” to name just a few of the frames he draws around a category that for libraries, archives, and museums “remains a narrow, niche, or lesser priority” for preservation.¹⁵ Software exists in particular material states and on particular media. Software can be transmitted from device to device through wireless signals, but it must ultimately be somewhere, and be something.

The field of digital forensics often relies on the material relationships among software and hardware, the inscriptions (and similar traces) made by software on its hardware. The field was developed and is typically employed to find digital evidence of crimes—shady bank transactions or incriminating emails—though its methods have been adapted by scholars for other purposes. In his 2008 book, Mechanisms: New Media and the Forensic Imagination, for example, Kirschenbaum excavates the material traces left on disks and hard drives to show how techniques adapted from book history and media archeology can locate the physical traces of digital texts. Beginning from the question, “In what…does the materiality of electronic texts consist?” Kirschenbaum argues that, like the letters inscribed on paper through print,

Electronic textuality is similarly locatable, even though we are not accustomed to thinking of it in physical terms. Bits can be measured in microns when recorded on a magnetic hard disk. They can be visualized with technologies such as magnetic force microscopy (MFM), which is a variation on the scanning tunneling microscope (STM). When a CD-ROM is burned, a laser superheats a layer of dye to create pits and lands, tiny depressions on the grooved surface of the platter. The length of these depressions is measured in microns, their width and depth in nanometers.¹⁶

Cold metal bites into rag paper to print a hand-press era book; electrons are stored in the cells of a solid state hard drive to save a Word file. As Kirschenbaum acknowledges, the materiality of something like a Word file is less immediately tangible to human senses, which makes it harder for scholars to apprehend.

Nevertheless this scholars must develop new analytical capacity to close this intellectual distance. As Alan Galey notes in his bibliographic reading of Johanna Skibsrud’s The Sentimentalists in both print and digital editions, understanding such objects will “require a synthesis between forensic methods and the humanities’ interpretive strengths.” Such a synthesis is necessary, Galey argues in his conclusion, because,

e-books, like all digital texts, require us to interpret phenomena not directly observable by the senses. We must rely on layers upon layers of digital tools and interfaces, as we have seen in the examples above. A purely empirical and forensic perspective assumes that objects speak for themselves, and yield up their evidence to the observation of human senses and the inquiry of human reason.¹⁷

The peculiar materiality of software cannot be entirely understood through direct observation, and will require researchers to develop new standards of evidence, as well as new methodologies for gathering it.

Scholars of material culture studies must cultivate such proficiencies, as the future of storage promises to write digital data onto new, even less-apprehensible material substrates. In their ongoing efforts to find more efficient, capacious, and compact media, researchers have succeeded in saving “502 terabits per square inch” in a “flat two-dimensional lattice” of chlorine atoms and vacancies. These researchers tout the potential capacity of atomic memory, noting in conclusion that “Translating the two-dimensional storage density presented here to three dimensions, would…allow the storage of the entire US Library of Congress in a cube 100 μm /[micrometers, or millionths of a meter wide/].”¹⁸ Another area of current research marries computation and biology, using DNA as a data storage medium. One group of researchers, for example, “encoded computer files totalling 739 kilobytes of hard-disk storage and with an estimated Shannon information of 5.2 3 10⁶ bits into a DNA code, synthesized this DNA, sequenced it and reconstructed the original files with 100% accuracy.” Currently DNA storage is too time- and resource-intensive for most archives, but researchers believe “if current trends continue,” then “DNA-based storage becomes practical for archives with a horizon of less than 50” years within a decade. In particular, researchers are looking to DNA as an archival medium,

The DNA-based storage medium has different properties from traditional tape-or disk-based storage. As DNA is the basis of life on Earth, methods for manipulating, storing and reading it will remain the subject of continual technological innovation. As with any storage system, a large-scale DNA archive would need stable DNA management and physical indexing of depositions. But whereas current digital schemes for archiving require active and continuing maintenance and regular transferring between storage media, the DNA-based storage medium requires no active maintenance other than a cold, dry and dark environment (such as the Global Crop Diversity Trust’s Svalbard Global Seed Vault, which has no permanent on-site staff) yet remains viable for thousands of years even by conservative estimates.¹⁹

While these media may almost impossibly microscopic, that we can encode digital data onto atoms or into DNA strangely makes the intrinsic materiality of that data plainly apparent. Data can be digital, chemical, or even biological.

Preserving and Accessing Digital Culture

One of the most potent and sobering reminders of the digital’s materiality is its rapid dilapidation. Technologists and archivists both increasingly use the phrase “digital dark age” to refer to the early era of modern computing, which becomes less accessible each year as its hardware degrades and its software grows less compatible with new hardware. These processes are often called “bit rot” or “data rot”—an intriguingly biological metaphor for the decay of digital artifacts. Google Vice President Vint Cerf has been perhaps the highest profile voice worried that “even if we accumulate vast archives of digital content, we may not actually know what it is” very far into the future.²⁰ As a result of material decay and the obsolescence cycles of the computer industry, an entire generation of cultural heritage is in critical danger of being lost to history. These endangered artifacts include software objects themselves—programs, computer games, operating systems—and objects created with them, such as the correspondence or working documents of important social, political, or artistic figures.

As it is written today, computer code operates at several removes from machine language. Programmers do not write in binary, they write in C++, Python, R, or any number of programming languages that are compiled or assembled, translated into machine code that the computer can execute directly. Machine code depends quite literally on particular hardware parameters; the set of available instructions differs for each processor, though newer processors may include instructions from their predecessors. When hardware evolves too far past the parameters for which a given piece of software was written, however, that program can no longer be translated and run as intended. Storage formats change even more rapidly, and many storage technologies physically degrade, in some cases at faster rates than pre-digital media, such as paper or vellum. As Roy Rosenzweig writes, “Print books and records decline slowly and unevenly—faded ink or a broken-off corner of a page. But digital records fail completely—a single damaged bit can render an entire document unreadable.” Even if data saved to a floppy disk is compatible with a modern program—and the floppy disk has not itself deteriorated—accessing that data would require access to a dwindling number of functional floppy disk drives, and an operating system capable of reading from the drive and disk. Rosenzweig asserts, “Well before most digital media degrade, they are likely to become unreadable because of changes in hardware (the disk or tape drives become obsolete) or software (the data are organized in a format destined for an application program that no longer works). The life expectancy of digital media may be as little as ten years, but very few hardware platforms or software programs last that long.”²¹ The challenges of preserving born-digital materials, then, comes not from the development of media, hardware, or software in isolation, but instead from simultaneous—but not necessarily coordinated—change across all three.

Preserving Virtual Worlds was a project that brought together researchers from the Rochester Institute of Technology, Stanford University, the University of Maryland, the University of Illinois at Urbana-Champaign, and Linden Lab to investigate the challenges of preserving video games and interactive fiction from the 1960s to early 2000s. In the project’s final report, they identify obsolescence as the first obstacle to their work:

The most obvious problem affecting these materials is the obsolescence of the hardware and software infrastructures necessary to allow software to run. The earliest game in our case set, Spacewar!, currently exists in its original form stored on a punched paper tape intended to be read into the memory of a PDP-1 computer. There is, to the best of our knowledge, only one functioning PDP-1 computer left in the world, at the Computer History Museum in Mountain View, California, and paper tape readers are not exactly common equipment at this time. The fate of the paper tape of Spacewar! is the fate awaiting all games without the active intervention of preservationists. A book may pass 50 years on a shelf and still be readily accessible; rapid technological change and the resulting obsolescence of the technology necessary to access software mean that a computer game will not.

Beyond this “most obvious” problem, however, the team noted the difficulty in bounding precisely what object must be preserved, because “While we tend to think of the game as a relatively discrete package of software, the reality is that a functioning game involves a web of interconnections between the game’s executable, an operating system, the hardware platform used to execute both, and potentially network hardware and software and a multiplicity of other computer systems.”²²

Scholars and archivists alike wrestle with different ideas about how best to preserve and present histories of hardware and software, and it is likely that some combination of these ideas will be required as the urgency and scope of the problem increase. One of the primary methods for born-digital preservation requires the maintenance of original hardware, on which researchers can access pertinent software, while the other relies on the emulation of older software environments on newer hardware. The Emory University Libraries, for instance, rely on emulation to present the digital “papers” of writer Salman Rushdie. Emory Libraries’ help sheet for this collection offers insight into the preservation challenges that digital technologies raise for libraries, museums, and other cultural heritage institutions:

Welcome to the Salman Rushdie digital archive. On this workstation, you will find selected digital files from Rushdie’s Macintosh Performa 5400, one of several computers and other related devices that form the born digital component of the Salman Rushdie papers in Emory University’s Manuscript, Archives, and Rare Book Library (MARBL).

The majority of the digital files date from 1992-2002, and consists of notes and drafts of Rushdie’s writings and selected correspondence. Of particular interest is a small cache of email correspondence, representing Rushdie’s first foray into this emerging form of communication in the late 1990s. Writings include drafts of Rushdie’s fiction, such as East, West (1994), The Moor’s Last Sigh (1995), and The Ground Beneath Her Feet (1999). Nonfiction writings include notes and drafts for Step Across this Line, Rushdie’s collection of essays and criticism, published in 2002. Other writings include drafts of the Midnight’s Children and The Courter play scripts, as well as drafts of letters to the editor, newspaper columns, poems, and speeches. The Performa 5400 contains a backup of an earlier computer, which Rushdie entitled “OLD MAC,” and a laptop, the “Powerbook,” which Rushdie likely used in tandem with the Performa 5400.

Emory Libraries does allow researchers to access these materials through a searchable database, but also through an emulation of Rushdie’s Macintosh Performa 5400. “In this environment,” the library’s help sheet reports, “you will be able to view Rushdie’s exact directory structure and open each file in the application in which it was created, such as MacWrite Pro or ClarisWorks.”²³ These two projects evidence ideas of emulation that are becoming increasingly important for thinking historically about digital artifacts and culture.

By contrast, other scholars seek to preserve historical computing objects in working order for future study. Founded in 2009, the Media Archeology Lab (MAL) at the University of Colorado at Boulder, describes itself as “a place for cross-disciplinary experimental research and teaching using still functioning media from the past.” Contrasting its mission to media labs focused on the latest technologies, the MAL claims to be “propelled equally by the need to both preserve and maintain access to historically important media of all kinds—from magic lanterns, projectors, and typewriters to personal computers from the 1970s through the 1990s, as well as early works of digital literature/art which were created on the hardware/software housed in the lab.”²⁴ In spaces like the MAL, students or scholars can play the 1987 adventure game King’s Quest III: To Heir is Human on an Amiga computer or experience writing in WordPerfect on an Apple IIe. Such resources are an enormous boon to those concerned with the social and material histories of the digital age, but are also an enormous undertaking to create and, especially, maintain as parts and expertise in older computers alike disappear. Whether through emulation or preservation, it has become increasingly clear that the early history of computing—a phase, I would argue, that continues to this day—will require substantially more engagement from scholars if we hope to preserve its texts and artifacts for future students, scholars, or the public.

Digital-Material Circuits

While the digital realm certainly is material, its materiality is in many ways distinct from those objects we identify as “analog.” The objects we encounter on a screen often mimic a physical form quite distinct from the form they actually inhabit. As Kirschenbaum argues, “a digital environment is an abstract projection supported and sustained by its capacity to propagate the illusion (or call it a working model) of immaterial behavior: identification without ambiguity, transmission without loss, repetition without originality.”²⁵ It is the domain of images displayed by computer screens that we mark as “virtual reality”; these are models of the real or fantasy worlds. Increasingly, however, the digital and analog worlds bleed into each other. When the mobile game Pokémon Go was released in July 2016, it sent players scrambling around their towns and cities in search of the cartoon creatures, which, though an “augmented reality” application, could be found on real streets, in real buildings, and even, sometimes problematically, at real historical sites and monuments. In 2016, for instance, the Holocaust Museum in Washington, D.C. requested that Pokémon Go players stop catching the cartoon creatures in the memorial space. The game’s developers automated the creation of “Pokestops” to coincide with landmarks as identified on digital mapping platforms, but without considering that some landmarks might be inappropriate locations for such play.

While this implementation of augmented reality was rightfully decried—and soon corrected—there are arguments in favor of interfaces that blend digital and analog environments. Even Pokémon Go was praised for urging players toward fresh air and exercise, and some scholars and cultural heritage institutions have turned to augmented reality to engage the public. The Museum of London launched the app Streetmuseum in 2010 (currently unavailable). The Histories of the National Mall website, developed by the Roy Rosenzweig Center for History and New Media at George Mason University, allows visitors to this Washington, D.C. spot to explore historical maps, photos, stories, and other materials related to the National Mall as they explore the physical space itself.²⁶ The Smithsonian’s Museum of Natural History’s Skin and Bones app attempts to breathe new life into one of its oldest exhibit halls, allowing visitors to overlay muscles, skin, and other anatomical features onto the skeletons in the exhibit hall, and even breathe virtual life into the creatures on display.²⁷ These applications bring virtual reality into explicit dialogue with the “real world,” layering the digital and analog.

In the past decade or so, the digital world has manifest in the analog in another significant way, through the growth of technologies such as 3D printing that express computer models in materials such as plastic, resin, ceramic, metal, and even tissue. The practical and theoretical applications of 3D printing are expanding rapidly, and the technology can be found everywhere form engineering firms, to hospitals, to libraries. The broadest application of 3D printing has been for creating physical models relatively quickly and cheaply, allowing engineers (or other researchers) to create a digital schematic of a design and then generate a physical representation of it relatively quickly and affordably. These models can be used to test structural properties of designs before production, compare different design options, or simply to demonstrate for clients.

As with augmented reality, museums and other cultural heritage institutions experiment with 3D printing as a way to bring students, researchers, and the general public into contact with artifacts they might not otherwise be able to interact with. New York’s Metropolitan Museum of Art was one of the first museums to begin making 3D models of its artifacts available for download and reproduction, as have national museums such as the Smithsonian²⁸ and British Museum, alongside a host of smaller institutions and individual research projects. The University of North Texas’ 3Dhotbed is “a collaborative project that seeks to enhance book history instruction by providing access to affordable teaching tools and related materials for pedagogical purposes.”²⁹. The project has created 3D models of hard-to-find artifacts from the hand-press era that teachers of book history, bibliography, and related disciplines can use to demonstrate objects’ use to students. The project’s current models include a two-part type mould and matrix used to demonstrate to students how individual pieces of moveable type were created, a facsimile of a Renaissance-era woodcut image, and a facsimile of a Chinese xylographic woodblock. These are materials not likely to be available for most classroom instructors, made accessible through 3D models and printing.

As the possibilities expand for material substrates used in 3D printing, so to have its applications. Practically, 3D printers can be used to create hard-to-find replacement parts for rare or outdated machines. Though there are relatively few replacement parts that justify the time and cost of creating a 3D model and printing it, but groups such as NASA are experimenting with ways the technology might assist with repairs on highly customized tools in out of the way places, such as outer space.³⁰ Experiments also continue in 3D printing another kind of rare “spare part,” as biologists and doctors work toward 3D printing viable human organs for use in transplants. While it might sound like science fiction, this technology has produced near-viable organs. A 2014 review article in Nature describes the major approaches to “3D bioprinting,” including biomimicry, autonomous self-assembly, and mini-tissues, as well as overviewing the primary technological approaches to this challenge. In this instance, the material culture of the digital becomes the biological culture of the digital, a new corporal entanglement between computers and ourselves.³¹

In addition, the past decade has seen a growing body of research and practice around ideas of “data physicalization” or “haptic data,” both movements that seek to challenge the domination of sight in expressions or analysis of data, and both movements that bring computerized data viscerally into the material world. In contrast to data visualization, “A data physicalization (or simply physicalization) is a physical artifact whose geometry or material properties encode data.” This paper goes on to imagine a museum exhibit in which visitors pick up stones of various temperatures to experience changes in Earth’s climate over time, as well as describing existing data physicalization, such as “a wooden three-dimensional model of Mexico City where height encodes population density.”³². Such physicalizations blend data science with the art installation, and close the circuit of data collected from physical or built environments, aggregated and analyzed computationally, and then expressed—transformed—in material form.

These kinds of experiments bring information out of screens and into human spaces, often putting people into literal contact with data representations. Haptic data experiments do not simply encode data through material properties, but invite people to experience data using senses beyond (though perhaps including) sight. The Vibrant Lives project, for instance, describes an approach that “look[s] to somatic and contemporary dance practices for new design strategies that engage users in affective, felt relationships with personal technologies and personal data.”³³ In installations sponsored by this project, participants experience the data streaming from their mobile phones as vibrations felt through devices worn on their clothing, or they both heard and felt the history of forced eugenic sterilization in California. In a blog post about the latter event, Jacqueline Wernimont describes participants “leaning in to feel a history of sterilization.” She notes that “The haptics are being shared with a thin, red metal wire that the participants have to touch lightly in order to not dampen the signal for others,” which is part of the Vibrant Lives team’s “effort to bring care for the experiences of others into the performance”³⁴

None of the above manifestations of the digital within the material quite touch an even broader set of ideas and practices gathered under the heading of “wearable technology.” These include the kinds of commercial trackers (e.g. the Fitbit) critiqued by projects such as Vibrant Lives, but also a range of experimental interfaces born from maker culture, most excitingly in dialogue with feminist or other ethical frameworks. Kim Brillante Knight surveys a range of feminist interventions in wearable technology, including this description of Kathleen McDermott’s work:

Kathleen McDermott’s Urban Armor collection is a series of wearable garments in which we glimpse the cyborg’s radical possibility. Of particular note is the “Personal Space Dress”; its skirt expands when activated by a proximity sensor in order to protect the wearer from unwanted contact (Taylor 2014). The choice of a feminine garment—a dress—calls attention to the way women in particular are subject to harassment in public spaces.

Knight describes her own research project Fashioning Circuits, which works in undergraduate classrooms “to articulate a counterpublic to dominant computing publics by creating a space in which women and underrepresented people of color feel comfortable learning and have the freedom to pursue projects that reflect their priorities and needs.”³⁵ They do this through a range of theoretical readings and discussions alongside making their own wearable technology that works against the dominant models of commercial devices.

The kinds of pedagogical, research, and artistic experiments described in this section certainly create new material artifacts for description and study by scholars. In such work we identify another way in which the abstruse digital realm is brought into meaningful, physical contact with people. For scholars of material culture, physicalizations, haptic data, and wearable technology are doubly resonant, as both objects of potential study and as modeling potential modes of analysis or engagement. As in the catalog of digital devices, this section cannot fully document the many ways in which the digital reemergences into the analog world. With several decades of digitization past, we are now coming full circuit, and such manifestations of the digital are only due to increase.

Conclusion

In Writing Machines, Hayles argues that the need for “media-specific analysis” became more important, rather than less, upon the introduction of the computer:

As the vibrant new field of electronic textuality flexes its muscle, it is becoming overwhelmingly clear that we can no longer afford to ignore the material basis of literary production. Materiality of the artifact can no longer be positioned as a subspecialty within literary studies; it must be central, for without it we have little hope of forging a robust and nuanced account of how literature is changing under the impact of information technologies.³⁶

For Hayles, this moment of transition or remediation makes media apparent where it was once functionally transparent. The digital makes clear the unique material properties of the media that came before, though—as this essay has discussed–the new medium can be difficult to perceive so clearly. Scholars of material culture must conscienciously address this self-effacing property of new media: to recognize that it mediates no more or less than its predecessors, and to attend deliberately to its affordances and limitations.

There is no cloud. What does exist is a vast plethora of digital devices, and a rapidly growing ecosystem of materials related to the creation, use, and maintenance of those devices. It would be a mistake to conflate digital culture with twenty-first century writ large, but it would be equally misguided to cordon digital studies from other forms of cultural analysis: material culture studies perhaps most of all. Hardware, software, and even data have material properties that warrant investigation and robust theorization.