The Declaration of Independence. The Book of Psalms. Hamlet. Speeches by Helen Keller.
These were some of the first works made available to people who were blind in the 1930s, when the Library of Congress, in partnership with the American Foundation for the Blind, began sponsoring the recording of hundreds of titles per year—on new, 12-inch records that played for a whopping 18 minutes per side (longer than the 11-minute commercial standard). Stacks of copyright-exempt records and the specialized machines needed to play them were distributed, postage-free, through a network of libraries to borrowers holding a valid medical certification indicating that they were visually impaired. (Book publishers wanted to ensure that sighted people wouldn’t gain illegal access to recordings of popular titles.)
Other early standouts produced years before the commercial audiobook industry hit its stride included recordings of “celebrity” authors such as Thomas Mann, Maya Angelou, and Harry Truman reading from their own works.
But blind people, like any readers, could hardly live on a diet of classics alone. Within a few decades of the program’s start, amid the countercultural revolutions of the 1960s and ’70s, there was a growing demand for recordings of works that—whether politically radical or sexually explicit—the Library of Congress deemed inappropriate or unworthy of recording. In other words, what about the juicy stuff?
Intrigued by this question and others, Steinhardt media, culture, and communication assistant professor Mara Mills has been exploring how such programs fit into the larger history of how people have consumed media throughout the ages. Her curiosity for such an undertaking was piqued when a colleague asked her to write an encyclopedia entry about Talking Books years ago—a small assignment that ultimately inspired a current NSF-funded archiving effort to trace the evolution of technologies for tactile reading and “reading by ear.” With a $239,138 NSF Scholars Award, she aims to produce a history of “reading formats” beyond print (namely audiobooks and text-to-tone, text-to-vibration, text-to-speech, and text-to-braille reading machines)—and to connect present-day debates about the definitions of reading to earlier discussions about the role of technology in blindness rehabilitation. What she’s begun to uncover is a winding tale of innovation and frustration—and one in which blind activists dissatisfied with what government programs had to offer sometimes took matters into their own hands.
In addition to digitizing the collection of about a hundred surviving recordings from the American Foundation for the Blind’s archives, she and her team have also been focusing on cassette recordings made by the Womyn’s Braille Press—a Minneapolis-based collective of blind women who worked in the 1970s and '80s to make audio recordings of books on feminism and queer theory that the Library of Congress deemed too risqué for blind readers.
That includes works like Alison Bechdel’s graphic novel Dykes to Watch Out For, which the author gave Mills special permission to reproduce. (Her digitization project for the NSF must otherwise obey copyright law.) Because these tapes produced “underground” weren’t a part of the Library of Congress system, Mills reflects, they represent a labor of love by “volunteers who put hours of work into reading each page aloud, and in the case of a graphic novel describing every image—and by Womyn’s Braille Press coordinators who facilitated their own network of tape exchange.”
This little-known effort is just one gem unearthed in Mills’ expansive—and urgent—project. Aided by two student research assistants, she is collecting and digitizing audio materials—from historic Talking Books to reading machine instruction tapes—that risk being lost or destroyed as many dedicated schools and libraries for blind people shut their doors. From the 1920s to the 1970s, most blind students attended specialized schools, a trend that dramatically reversed following federal legislation requiring “mainstream” public schools to provide any adaptive equipment or special accommodations for students with disabilities.
After just one semester in a two-year research period, the quest has already led the team not just to the archives of major institutions like the New York Public Library and the National Federation of the Blind, but also to the garages and basements of activist groups and even individual collectors who happened to salvage valuable recordings or machines. They may yet visit a prison.
And their discoveries about the variety of assistive reading technologies developed decades ago raise striking questions about the persistent challenges people who are blind encounter as they navigate the world of print today. “Partly what we’re doing is trying to tell an alternate story about how people have read, and what counts as reading in the 20th century,” says Mills.
Looming in every conversation on the topic is the industry’s equivalent of Coca-Cola—the Kurzweil Reading Machine, invented in the 1970s by Ray Kurzweil, now director of engineering for Google. The contraption, then about the size of a washing machine, was marketed as the first optical character recognition (OCR) reader that could translate print of different fonts into synthetic speech: You simply placed a page on the scanner’s glass and the machine would “read” it aloud—albeit in an unnatural-sounding computer-generated voice.
Though the early Kurzweil readers—most purchased by libraries and other large institutions—were bulky and expensive (about $22,000 a piece in the early 1980s), text-to-speech captured the public imagination and soon became the most widespread method used by blind people to read. Stevie Wonder bought one of the first Kurzweil units, and Walter Kronkite used one to sign off from his evening news broadcast in January 1976. OCR scanning technology, now ubiquitous in applications for blind and sighted users alike, was harnessed in the 1970s to develop the electronic LexisNexis database of journalistic and legal documents. And today, many blind readers continue to use later iterations of Kurzweil software for laptops or even smartphones.
But for people who are blind this has hardly been the only—or even, one could argue, the best—option for reading: Braille dates to the 1820s, and as early as 1912 engineers began developing a variety of scanners, often portable, that converted printed matter into vibrations or musical tones. Then came the first Talking Book records—and later their descendants on cassette tape and CD. And in the years after WWII, innovation in reading technology flourished as the government funded research to assist veterans blinded in combat.
One prevailing theme in the history of “reading by ear” is the need for speed, something the Library of Congress program reckoned with almost from the start. Early Talking Books featured theatrical narration and even sound effects, but by the 1940s, Mills says, the AFB was inundated with “letters from blind people saying, ‘we don’t want interpretive or dramatic readings anymore—we just want the narrator to be a vehicle, with clear enunciation.’” Plainer recordings could more easily stand up to early pitch-shifting technology used to avoid the telltale “chipmunk” sound that comes from playing a record at high speed.
Mills’s research assistant Shafeka Hashash, a joint bachelors-masters student in politics and Middle Eastern and Islamic studies who is herself blind, says that today, now that we can speed audio up digitally, it’s possible to listen to an entire 200-page book in a little over an hour—though blind readers continue to clamor for ever-faster playback of things like emails and text messages. And though there's still no substitute, Hashash says, for “reading yourself”—reading braille, in her case, as opposed to being read to—she adds that “reading by audio definitely has its advantages and once you get used to it you really can be very efficient.”
And when the words are going by so quickly, it doesn’t much matter who’s reading them. After the early star-studded era when politicians and even Broadway stars as well as authors lent their talents to Talking Books, the program came to rely on ordinary volunteers. But Mills has also learned of prisons—in Massachusetts and California—that once boasted in-house recording studios where inmates could record textbooks and other works not available to blind readers through the Library of Congress system. “It was thought that it would be good for the rehabilitation of prisoners’ minds and souls to help blind people,” Mills says. “It was a fascinating experiment in the 1960s: One program even had a whole newsletter about the process and encouraged correspondence between the blind students and the prisoners.”
So far, she hasn’t unearthed any of those recordings. “But we’re trying to find out if any of the participants—on either side—are still alive to interview,” Mills says.
For every new lead Mills chases, there are delays and even the occasional dead end: In the digital age, many libraries have tossed items in older formats like cassette tape or LP, and surviving materials that haven’t been requested in decades can take time to locate. And when Mills does find something of potential value, there’s the tricky business of making sense of it: Many of the tapes and other materials she’s found are labeled only in braille, so she relies on Hashash to transcribe the labels and any important braille correspondence she finds. On the other hand, a surprising number of finding aids even in the archives of blind institutions are available only in print, not in braille, making solo research by a blind researcher next to impossible. “The research for this project has been really unusual and really fun, but also really complicated,” Mills reflects.
Preserving their choice finds for posterity falls to research assistant Patricia Flores—a master's student in media, culture, and communication and former collections management coordinator for the Metropolitan Museum of Art—who can digitize cassette tapes and LPs at Bobst’s Digital Studio (which offers this service FREE to all students). Reel-to-reel tapes and other unusual formats must be sent to be digitized elsewhere. Another hurdle will be figuring out how to make the final product—a book by Mills that will necessarily include multimedia components—accessible to all.
Lately the trio is most excited about a trove of braille, tapes, and reading machines taking up space in the Chicago garage of Harvey Lauer, who worked as a technology transfer specialist for the Hines VA hospital for over 30 years, testing and teaching veterans to use various reading machines. Now in his 80s, Lauer, who is blind, is one of the world’s experts on reading machines—especially those, beginning with the Optophone in 1912—that translated print into musical tones or vibrations. Lauer himself still uses one such device that dates to the 1960s, although most became obsolete with the rise of Kurzweil-type text-to-speech readers in the 1970s.
Hashash, who’s been interviewing Lauer for the project, says it’s telling that Lauer continues to find value in the older technology. Simply put, there’s a lot OCR can’t do. “Penguin Books is a good example,” Hashash says. “Each book has a picture of a Penguin on its cover, but a blind person would never know that—it would be entirely skipped over by text-to-speech.”
The tactile readers that Lauer favored didn’t just “see” letters; they attempted to render everything on the page into sound, making it possible for an experienced user to, say, read a graph.
In these text-to-speech heavy days, it’s technology we should and are finally beginning to revisit, Hashash says, pointing to Math Tracks, an online graphing calculator that uses musical tones to outline the shapes of parabolas and the like. “The big downfall of Kurzweil is their inability to do more complex things other than books—things like science,” she says. Text-to-speech technologies often fail to recognize the elements—like Greek letters—of standard equations and formulas.
And even the newest assistive technologies for people who are blind—like smartphone apps that can describe the objects captured in a photo—have yet to address some common everyday headaches. Consider the case of a blind woman who’d like to sift through the private contents of her handbag—without a sighted person present to help. “Something to read absolutely everything that’s on an old receipt, and the label of a lotion bottle, and the expiration dates on a bottle of medicine—this has not yet been fully developed,” Hashash says.
The answer to the riddle might be a multipronged approach—drawing on some combination of text-to-speech, braille, and even elements of the rare or forgotten machines Mills’s history is bringing back to light.
“This is a story about successes, of course, but also about missed opportunities,” she says. “There are ways that people might have read that aren’t happening now, and we’re trying to figure out why.”