Virtual Roundtable on “Compression”

The mass of objects lead quiet lives awaiting activation. On shelves or in boxes, as papers or digital files, storage furnishes an ever-present ...

The mass of objects lead quiet lives awaiting activation. On shelves or in boxes, as papers or digital files, storage furnishes an ever-present background. Investigating this background shifts our focus to overlooked material properties: to the hardness of corners, to pliability and tensile strength. It forces us to ask how objects have been rearranged or redesigned for differing spaces, real and virtual, according to varied demands of efficiency, accessibility, and mobility; according to how we think future users should search, and what they should find. For storage does not merely hold objects in waiting. It transforms and even generates them.

The following essays emerge from a conference exploring the transformation of stored objects through the lens of “compression.” Beginning with one of the oldest alleged forms of compression—the alphabet—Stephanie Ann Frampton leads us through a history of consonants and vowels lost, discovered, and invented, complicating any simple notion of alphabets as efficient containers for sounds. Nor have sounds alone posed dilemmas for alphabetic storage: as Evan Hepler-Smith demonstrates, by condensing molecules into alphabetical formulae, chemical information systems flattened our understanding of molecules themselves, with potentially deadly consequences. Finally, Craig Robertson exposes the centrality of compressed storage for the imagination of modern political economy. Examining a once-revolutionary symbol of workplace efficiency—the filing cabinet—he shows how the practice of storing paper “on its edge” emerged alongside that other product of “vertical” capitalism, the skyscraper.

Sounds, chemicals, papers: these case studies reveal how compression as an ideal of storage has fundamentally reshaped everyday objects that our social, economic, and political structures take for granted. This insight today seems ever more vital. As storage devices grow in capacity, promising unlimited data for grasping reality, we cannot forget that storage, too, produces its own fictions.

—Hansun Hsiung, guest editor


• Stephanie Ann Frampton: Our Lossy Alphabet?

• Evan Hepler-Smith: A Leak in the Chemical Archive

• Craig Robertson: File Under “V” for Vertical


Stephanie Ann Frampton


One example of compression lies on the screen in front of you. Writing has been invented independently as few as three times in the history of the world: in ancient Sumer, in ancient China, and in medieval Mesoamerica. Sumerian cuneiform was likely the inspiration for Egyptian hieroglyphs. Those, in turn, inspired the early Semitic script from which all European and several Asian scripts ultimately derive.

The earliest evidence for our branch of the scriptural family tree, dating to around 1800 BCE, was discovered less than 20 years ago in the Egyptian dessert west of the Nile, at a site called Wadi el-Hol, the “Valley of Terror.”

Wadi el-Hol alphabetic inscription #1 (Photograph by Bruce Zuckerman and Marilyn Lundberg, West Semitic Research)

Carved into the soft limestone walls of caves above an ancient trade route, the two short inscriptions capture a moment when speakers of a Semitic language, far from home, adapted and adopted foreign characters to transcribe their own tongue. Probably using what is now known as the “acrophonic” principle (Greek: ἄκρος uppermost + φωνή sound), they borrowed several pictographic characters of Egyptian hieratic script (a simplified form of hieroglyphics) but used them to represent the initial sounds of the Semitic names for the objects illustrated, thereby creating a phonetic code.

Smatterings of writing in later versions of this script are found across the ancient Levant, from the Sinai Peninsula through northern Syria, concentrating in the Orontes River estuary, at Byblos—the city that would give us the Greek words for book and Bible (βιβλία)—and at the Phoenician port of Al Mina dating from the 10th–8th centuries BCE. It is there and then that many scholars think that writing was taken across to the western Mediterranean, first by Euboean Greek traders active at Al Mina from around 800 BCE. Through them, other Greeks, Etruscans, Latins, and a great number of other peoples of ancient southern Europe acquired writing.

Of the six or seven thousand world languages, fewer than a third are written. Of that third, nearly a quarter are typically written in the Latin alphabet, one of only 50 global writing systems. The alphabet today, in its most common guise—namely, this one—is thus regularly used to transcribe some five hundred languages and, if the Coca-Cola Company is to be believed, is recognizable by nearly every person on the globe: a fact the ancient Phoenicians, Greeks, and Romans never could have anticipated.

Can we attribute this easy mobility to the alphabet’s efficiency in compressing sounds into symbols? Rather than the 2000–3000 characters one must know to read a Chinese newspaper fluently or the 500–1000 signs used in Sumerian and Akkadian cuneiform, only 20–30 alphabetic characters are needed to record an increasing number of languages. Yet our alphabets’ history reveals just how slippery and complex “compression” is in relation to the messages that it allegedly compresses. For strictly speaking, alphabets expand rather than compress. Each alphabetic character records far less information individually than the more graphically complex characters of syllabic and logographic systems from which they arose, like hieroglyphics and cuneiform.


“Exekias made me”: a vase signed in two versions of the archaic Greek alphabet (Villa Giulia 50599)

Consider the case of vowels. Just as they are absent from most Hebrew written today, marks of vowels were not usually necessary for transcribing ancient Semitic languages. Nor were they absolutely necessary for transcribing ancient Greek; Linear B encoded vowels along with consonants in its syllabary. When Greek speakers adapted characters from the West Semitic script to systematically represent vowels, they in fact expanded the inventory of types of sounds encoded in their writing.

To be sure, expansions—such as the inclusion of separate vowels—have proven helpful in transcribing a vast range of languages, including our own. But instead of “compression,” we might at most say that the alphabet affords instances of parsimonious expansion: to grow while changing as little as possible.

Moreover, these growths and changes are by no means linear. Often, they occur through a convoluted series of substitutions and erasures. For its vowels, the Greek alphabet appropriated several West Semitic characters representing various consonantal sounds, discarding their original value in the process: gone were a glottal stop, a voiced pharyngeal (something like a rough “r” pronounced at the back of the throat), and the glides /y/ and /w/. In Etruscan, /k/ replaced /g/ as the sound encoded by the third character of the alphabet, bringing us the C we have today in place of Greek gamma. When 3rd-century BCE Romans decided that they needed a sign for /g/ after all, they created one by adding a small diacritical mark to C, creating G. This G, they placed in the seventh position, replacing the Greek zeta. Z was added back some centuries later, but at the end of the alphabet, where we find it today.


2015 revised IPA

At the limit, the International Phonetic Alphabet attempts to document all of the possible sounds in human languages with an inventory of more than a hundred signs. But the superfluity of the IPA, useful only for scholars and pedants, reveals a core lesson: that talk about compression and its efficiency is not about any inherent relation between technologies and the information they aim to encode, but about human conventions. Writing will always be “lossy” relative to spoken language: a term borrowed from data science, as “compression” itself is, to name the discarding of information considered inessential or unnecessary.

Alphabets never actually document sound. They are not a script for performance, nor a faithful recording of speech. The legibility of alphabets, like other scripts, relies not on their phonetic accuracy relative to a given language, but on the agreements made between their users across space and time. Pliny the Elder calls writing one of civilization’s greatest “tacit agreements” (along with beard-trimming and time-keeping). More recently, in his wonderful romp through the history of the English alphabet, Michael Rosen calls the rules of writing “more like treaties between consenting groups.” Whatever is lost in the transcription of sounds into marks on a page, it is certainly in our power to find that much more.


Jump to: Stephanie Ann Frampton, Evan Hepler-Smith, Craig Robertson


Evan Hepler-Smith


On January 9, 2014, just outside Charleston, West Virginia, 7,500 gallons of an oily liquid leaked out of a faulty chemical storage tank belonging to Freedom Industries. A half mile downstream, a treatment plant drew drinking water from the Elk River. A curious smell, like licorice with a hint of peppermint, began to emanate from the river, and, soon, from taps across nine counties. Over the following week, hundreds of area residents sought treatment for nausea and rashes, about a dozen were hospitalized, and 300,000 were left without running water.

Charleston residents and public health officials were keen to learn what this substance, a coal-processing chemical called MCHM, might do to them. The chemical’s full name, 4-methylcyclohexanemethanol, was a bit ominous, but it augured scientifically precise answers. This systematic name was an ontological promise: a pledge that a chemical—an oily pool of licorice-smelling, rash-inducing matter—was really a molecule—an assemblage of atoms and bonds. Yet those who managed to navigate the thickets of chemical nomenclature to find a data sheet or inventory record for this compound were left frustrated. It turned out that the oily liquid MCHM and the database entry MCHM were not quite the same thing.

La chimie crée son objet, wrote the chemist Marcellin Berthelot: chemistry creates its own objects of inquiry. Scattered among the millions of such objects known to modern science are the drugs, plastics, fuels, solvents, adhesives, soaps, lubricants, dyes, pesticides, flavorings, scents, toxic wastes, and sundry other products of the chemical and pharmaceutical industries. At the base of this edifice of science and industry lies a century-and-a-quarter-old foundation of chemical information. First in print, then on digital media, chemists have kept track of myriad novel and potentially lucrative synthetic chemicals by rendering them as molecular units that could be easily stored and retrieved. Information systems create their own objects, too.

Late 19th-century chemists devised rules that took each compound’s structural formula—a genre of molecular diagram that had become an indispensable tool for chemical research—and transformed it piece by piece into a systematic chemical name. Such names could be illegible, unpronounceable, and indistinguishable, but they had the signal virtue of putting chemicals in alphabetical order according to their molecular structure. Since then, directly or indirectly, nearly anyone seeking to study chemicals, harness their virtues, or control their dangers has relied on a printed handbook or computer database filled with such names.

“Is that …?” “No, no, it’s …” A cartoon submitted to the editor of Helvetica Chimica Acta satirized the indecipherability of systematic chemical names. Journals required the use of such names to enable the compilation of molecule-by-molecule subject indexes.

But there was a catch. As an astute observer noted when the scheme was first proposed, such systematic names “were really names of formulas rather than names of substances.” In 1892 this slippage in reference did not seem too troublesome. Nobody had much of an idea what the atoms and bonds represented in structural formulas even were, physically; there was little danger of mistaking a notional molecular formula for an observable chemical substance.

Over the decades that followed, as information about thousands and then millions of chemicals was cataloged in this way, chemistry changed. Chemical theorists developed an elaborate conception of the three-dimensional microstructure of molecules; other chemists grew accustomed to conflating their doodled structural formulas with these elaborate physical models. Information technology changed, too. The labor of forming and decoding systematic names became the province of machine algorithms. The difference between chemicals and molecules, so clear a century before, became harder to discern.

The chemical that spilled into the Elk River took on its dual identity within such information systems. MCHM entered the storehouses of chemical knowledge as one of a half-dozen compounds mentioned in passing in a 1908 article by a pair of British chemists. The chemical’s systematic name was duly registered in the reference literature. Like most of the chemicals in these repositories, it garnered little attention, but those few chemists who did study MCHM treated it not as one chemical but two. Known as cis- and trans- isomers, these two compounds had distinct three-dimensional shapes and different properties; however, as pictured below, they shared the same two-dimensional structural formula that gave MCHM its name. By default, most information about MCHM continued to be stored under its original name, flattening these two physically distinct chemicals into one molecule. It was this flattened molecular object that the Nalco Chemical Company patented in 1990 for use in coal processing, setting it on its way toward the Elk River.

Top: the two-dimensional structural formula and systematic name for MCHM, illustrating the unit-by-unit correspondence between name and formula. Bottom: depictions of the three-dimensional structures of cis- and trans-MCHM; the methyl and methanol groups are on the same side (cis) and opposite sides (trans) of the ring.

Sometimes, such a pair of isomers can be a chemical Jekyll and Hyde, one benign, the other toxic, as was famously the case in the epidemic of birth defects caused by one isomer of the drug thalidomide. Fortunately, MCHM has not (yet) been found to be especially hazardous at the concentrations it reached in the water supply, according to toxicological studies conducted in the aftermath of the spill. These studies brought something else to light: only trans-MCHM has the licorice odor that pervaded the rivers and taps of central West Virginia in January 2014. The noses of Charleston residents could distinguish the two chemicals; the databases of chemists and regulators did not. The isomers of MCHM are “the same and not the same”: the same when sitting anonymously in chemical information systems, not the same under the focused attention of chemists or in the water of West Virginians.

A one-inch hole in a chemical storage tank thus revealed a gap in a global system for storing chemical information. It might seem perverse to look to databases for the strengths and shortcomings of chemical storage, particularly where there is a lively story to be told concerning coal mining, tainted water, and the negligence of a company named Freedom Industries. But arcane as nomenclature and information systems might seem, their reach is broad and their influence immense. If MCHM had turned out to be harmful, the distinction between its cis- and trans- isomer versions could have been as consequential for the chemical’s dangers as it is for the chemical’s smell. Like storage tanks, databases have holes.


Jump to: Stephanie Ann Frampton, Evan Hepler-Smith, Craig Robertson


Craig Robertson


If you work in an office, maybe it has a file cabinet; maybe you still use it. These days, however, I’m more likely to see file cabinets discarded on the sidewalk. Ask most people for a “file,” and they’ll think of an icon on a desktop, rather than paper retrieved from a bulky metal structure. As we inch forward in the 21st century, the file cabinet’s days seem numbered. But when it first became pervasive a century ago, the file cabinet was understood as something truly radical—something that revolutionized information storage technology. It was one of the innovations that altered the organization of the office, and through that, models of corporate capitalism.

Invented in the United States in the 1890s, the file cabinet changed the way that offices stored and accessed paper and, therefore, information. It did this through one innovation: by providing the mechanisms to store large amounts of unbound paper on its long edge, or “vertically,” as it came to be called, in distinction to the existing method of “horizontal” filing that stacked papers in trays or drawers.

That vertical storage was celebrated as a novel idea may seem counterintuitive today. But the easiest position in which to rest paper is its face, not its side. Though people had long stacked paper in boxes or rolled it in pigeonholes, storing paper on its edge became fully possible only after a series of technologies were added to a file drawer, including some adapted from the recently invented library card catalog. A critical first step was the development of a “follower block” or “compressor”: a piece of wood or metal placed at the rear of a cabinet, that ideally functioned to keep papers in place and prevent “sagging or slumping in drawers.”1 These anthropomorphized papers didn’t just lie in wait; they stood to attention.

The vertical storage of paper engendered other technologies to organize the space of the file cabinet into discrete locations. Folders emerged to surround and protect papers. Tabs were then added to flag a specific location. Yet even with such changes, “information at one’s fingertips” did not materialize so easily. Paper tabs absorbed moisture. Their celluloid replacements scratched handlers as the tabs bent, curled, or cracked with use.

When it first became pervasive a century ago, the file cabinet was understood as something truly radical.

Still, by 1912, a key federal commission on efficiency would recommend that all government departments store papers on their edge in vertical file cabinets.2 According to the file cabinet’s celebrants, placing paper “vertically” not only economized space, but also time. Drawers stacked high freed up floor space; paper on its edge, in folders with tabs, allowed for faster search and retrieval. Manufacturers frequently used the phrase “at a glance” to further emphasize their claim that storage and retrieval would be instant or automatic. Behind this understanding of efficiency was the assertion of what could usefully be termed “granular certainty.” The proper path to efficiency lay in breaking objects down into smaller parts that could be located within a visible spatial order.

Paper storage is only the most tangible manifestation of corporate capitalism’s vertical bias. The file cabinet emerged in the era of vertical integration, along with hierarchical management and the metaphor of the corporate ladder. The file cabinet took up residence in another new structure: the skyscraper. Both these structures stacked discrete horizontal units (floors and offices; drawers containing folders and papers). Rather than a heap—a pile that led to disorganization—the vertical, enacting faith in repetition and granular certainty, forced activity and objects into a space organized to facilitate flows of storage and retrieval.

The comparison between the skyscraper and the file cabinet didn’t go unnoticed. In the first half of the 20th century, advertisements and catalogs often paired images of file cabinets and skyscrapers. Michigan-based Shaw-Walker marketed its file cabinets with the trademark “Built Like a Skyscraper.” To build a file cabinet “like a skyscraper” was to construct a steel skeleton that supported weight in the same way that allowed a skyscraper to transfer a building’s load-bearing function away from masonry walls. This support was important if heavy file drawers were going to fulfill the instant retrieval promised in ad copy. Decades later, C. Wright Mills would also compare the file cabinet and the skyscraper, labeling the latter “the enormous file” in his mid-century critique of white-collar work.3

By Mills’s time, the now familiar model of the “network” was emerging to challenge the value of verticality as an organizational strategy in capitalism.

The rise and fall of the file cabinet offers valuable insights into the historical shifts of capitalism’s logics. At a time when paper-based information had become a critical source of legitimacy for the corporate decision-making process, the file cabinet’s verticality articulated a new vision of how both the space and time of paper could be rendered more efficient. This same “verticality” came to structure the business imagination of the early 20th century at large. As a disciple of modern business (and cofounder of Shaw-Walker), L. C. Walker, put it in 1930, “Every time a piece of paper stops, a dollar is resting.”4


Jump to: Stephanie Ann Frampton, Evan Hepler-Smith, Craig Robertson icon

  1. E. R. Hudders, Indexing and Filing: A Manual of Standard Practice (Ronald Press, 1919), p. 59.
  2. “Memorandum of Conclusions,” Transmitting Reports on the Commission on Economy and Efficiency, H. R. Rep No. 62-670, at 519 (1912).
  3. C. Wright Mills, White Collar: The American Middle Class (Oxford University Press, 1951).
  4. L. C. Walker, The Office and Tomorrow’s Business: Modern Points of View on Office Organization and Administration (Century Company, 1930), p. 47.
Featured image: Park Street Bridge, Alameda, CA, December 20, 2012 (detail). Photograph by photoresonate / Flickr