Audio brought to you by curio.io, a Lapham’s Quarterly partner
“Might a writer dictate a fleeting dream in the middle of the night and upon waking find not only that it has been written, but rejoice in his freedom from the pen—that instrument that struggles with and chills expression?”
So mused the nineteenth-century French inventor Edouard-Léon Scott de Martinville, who believed the technology he had developed might enable this conversion of speech to text.
Scott has largely lurked in the footnotes of history. Credit for the discovery of sound recording, according to American conventional wisdom, has gone to inventor Thomas Edison. Tinkering in his laboratory in Menlo Park, New Jersey, in 1877, Edison found he could register noises as grooves on tinfoil-wrapped cylinders that could be played back multiple times, until the tinfoil tore. In this version of history, all our modern audio technologies descend from Edison’s phonograph, at the very top of the recording family tree.
Two decades before Edison, however, Scott claimed to have captured sounds out of the air using his own invention, which he called the phonautograph. His machine was made up of a large drum (later versions would use a horn much like the phonograph’s). Sounds directed into the drum vibrated against a membrane attached to a needle, which etched the undulating lines onto soot-blackened paper wrapped around a hand-cranked cylinder.
These strips of paper were Scott’s recordings, or “phonautograms,” but he never developed a way to play them back. He didn’t think it possible—but the prospect didn’t really interest this man of letters. Instead Scott spent years studying his squiggly lines, hoping they represented a “natural stenography” people might learn to read by sight. No one would hear the sounds he recorded until 2008, when Scott’s own voice would speak directly to listeners more than a hundred years after his death.
Imagine yourself in Paris around the year 1880, shortly after the invention of Edison’s phonograph. His invention is being presented before an audience of learned men in Paris, perhaps at the Academy of Sciences, where one researcher turns to the other and says, “That’s all very well in practice, but it will never work in theory.”
It’s a joke David Giovannoni, a sound historian and one of the central figures in bringing Scott’s recordings back to life, likes to tell. Its appeal lies in its depiction of the affectionate American stereotype of the cerebral Frenchman. “I didn’t mean to insult the French Secretary for the Academy of Sciences,” says Giovannoni, “who found it very funny the first time I told it, in her office last year.”
The story gets at a truth of this era, however. Scott believed that his wavy lines held not just information about volume but about the shapes of words themselves, the content of speech, and we now know he was right. But the theories that dominated European scientific thought in the late nineteenth century suggested Scott’s machine couldn’t work as he claimed it did. Absent the ability to play back his recordings as proof, Scott could not disprove them in the decades before Edison’s invention.
Scott died of an aneurysm within a year of learning of Edison’s phonograph in virtual anonymity, literally buried in an unmarked grave. His family could not afford a headstone. In his will, Scott asked his children to make sure he and his invention were not forgotten.
Giovannoni and his colleagues were out celebrating the completion of their latest restoration project, an album of “indecent” recordings from the archives of the Edison National Historic Site, when they first thought of seeking out Scott’s phonautograms. The group—which includes Rich Martin and Meagan Hennessey, who co-own Archeophone Records, and Patrick Feaster—is now known as the collective First Sounds. At the time they discussed the idea of seeking out the world’s oldest audio recordings.
Feaster had read about Scott’s invention, which he confesses he initially found uninteresting, “a technical precursor for the phonograph” that seemed mostly to have been used for scientific experiments. His assessment changed as he read more about Scott’s ambitions for his machine. Another researcher sent him copies of Scott’s patent paperwork for the phonautograph: “handwritten documents, Xerox copies of Xerox copies of Xerox copies.” Still, they were clear enough to show that the patent included an example of a phonautogram, which he mentioned to the group.
On a trip to Paris soon afterward, Giovannoni consulted Scott’s patents, which he found contained two phonautographs. He made high-resolution scans of both, but they turned out to be disappointing. One was unreadable. The other proved to be “a nice long phonautogram of a tuning fork,” says Feaster, “which is not very exciting. Although it’s from 1859, so it’s a very old recording of a tuning fork.”
Meanwhile they had started gathering all the documentation they could find on Scott, including his rare 1878 book The Problem of Self-Writing Speech. In that volume, Scott mentioned depositing phonautograms at the French National Academy of Sciences.
Giovannoni requested a set of low-resolution scans from the academy. When they arrived, he could tell immediately they had found something more promising: what looked like spoken passages and songs. He also saw evidence of a tuning fork, running simultaneously under the spoken words, which excited the team. The problem with recording on a hand-cranked cylinder is distortion caused by its uneven speed. Because a tuning fork vibrates at a steady frequency, running one simultaneously with a cylinder recording could provide a sort of time code they could use to realign the recording.
Giovannoni was wheeling a flatbed scanner into the Academy of Sciences within a week. Feaster began with a phonautogram labeled with the remarkable date of 1860 and the French folk song “Au Clair de la Lune.” Feaster pulled an all-nighter, cleaning up the scans in an audio editing program, where they appeared much like the sound wave shapes familiar to users of the voice memo function on their smartphones. Zooming in, he manually adjusted the sound waves, using the tuning fork as a reference, to smooth out the fluctuations in speed.
“The real moment when I knew that this was going to work,” he recalls, “was when I finally got to the second different note in ‘Au Clair de la Lune.’” He hums the opening of the piece to demonstrate, several beats on the same note before the tune rises. “That first series of notes could just be coincidence, but when it goes up where it’s supposed to, you know it’s worked,” he says.
Feaster sent the recording to Giovannoni in the early hours of the morning. Within weeks, their reconstitution of the oldest recording of the human voice was on the front page of the New York Times.
Later work on other recordings led them to figure out they had initially gotten the note of the tuning fork wrong, so that the recording came out higher pitched, sounding like the voice of a young girl. It seemed credible, Feaster says, because “the voice of the young girl sounded a lot more like what we would expect from a sound recording. It sounds like a lively performance. When it’s played back at the correct speed, which is half the original speed, it’s lugubrious,” less like a performance than the science experiment it really was.
Scott drew out each syllable “very, very deliberately, because he was wanting to see if he could make a record he could decipher visually,” Feaster says. The sounds are scratchy—intelligible with the help of the accompanying text but otherwise difficult to interpret.
To Laurent de Martinville, the voice of his great-grandfather on the recording sounded familiar, like a cross between two of his uncles. De Martinville was contacted by the BBC after news of First Sounds’ reconstruction made international news.
De Martinville had grown up believing his relative had created a successful sound-recording device without ever receiving his due recognition. His father had a phonautograph in his office that had interested him as a child.
Almost a decade after the reconstruction vindicated his great-grandfather, de Martinville finally stood face-to-face with David Edison Sloane, the great-grandson of the man who had gotten all the credit—the first time members of these families had ever met.
Sloane and de Martinville met in West Orange, New Jersey, at The Thomas Edison National Historical Park. The complex of buildings included Edison’s laboratory for the later part of his life. His Menlo Park workshop, where he developed the phonograph, was moved to Dearborn, Michigan, by Henry Ford. It is now part of a collection of historic buildings called Greenfield Village.
De Martinville had flown to New Jersey from France for a commemoration of Scott’s two hundredth birthday, organized by Giovannonni. De Martinville has perfectly white hair and wears stylish glasses. Sloane has a mustache and wears a brown fedora, looking every part the academic historian of Mark Twain that he is when not representing the Edison legacy.
Sloane’s famous relative was kept a secret when he was growing up. “We as children were cautioned to not talk about our relationship to Thomas Edison ever in public,” he recalled. “The kidnapping of the Lindbergh baby was still fresh in people’s minds in the 1940s. We were just plain folks,” he said, living in a middle-class neighborhood, but his parents “were very afraid for our safety in relation to the hazards of being connected with a presumably very rich and famous man.”
For de Martinville, the long-awaited meeting represented the important establishment of a connection between the two families, something his grandfather and his grandfather’s sister had failed to achieve during Thomas Edison’s lifetime. Both wrote letters to Edison. In one his grandfather asked if Edison would speak at the ceremony in honor of his father. De Martinville summarizes the short note: “Illustrious master, you have achieved all the success that man could possibly achieve in his life…could you be kind enough to say a few words for my father who passed away in great poverty and totally unknown?”
He never received a reply.
It turns out there is heated debate in academic circles about whether Edison’s phonograph, which consists of a metal funnel which amplifies the sound played from cylinder mounted atop a wooden base—a lot like early record players except for the shape of the recordings—owes a debt to the French inventor. While an actual phonautograph could be found in a university laboratory in Hoboken, New Jersey, and another at MIT at the time Edison invented the phonograph, some question whether Edison could have missed references to Scott’s invention in the scientific press.
“I think it’s really unfortunate,” says Giovannoni. “Most people who want to argue about the course of events that led to the invention of the phonograph or the telephone have a dog in the race. They’re very eager to latch on to any evidence that makes their person look good. And very quick to attack any argument or any evidence that calls their accomplishments into question.”
He and Feaster believe an examination of Edison’s lab notes leaves little room for doubt that Edison did not have direct knowledge of de Martinville’s phonautograph in advance of the phonograph. He did clearly know of a buggy telephone it had inspired, developed by German schoolteacher Philipp Reis.
Edison appears to have arrived at his recording method based on a very different—and, it turns out, wrong—conception of how sound traveled. Believing that sound consisted of a series of impulses of varying strength, he saw his recordings as a series of dots, not recognizing the continuous up-and-down path between the indentations he saw in his tinfoil recordings—a path that, viewed from the side, would have looked a lot like Scott’s phonautograms.
The audio historians—who have since deciphered other Scott recordings, including lines from Shakespeare’s Othello and Jean Racine’s Phèdre—do not have entirely satisfying answers to questions about who first invented sound recording. For Giovannoni, the importance is correcting the historical record of technological development. For Feaster, it serves as a lesson on the assumptions we make about the progress of human invention.
“We tend to assume that correct inventions that work are based on correct understandings about how the world works, correct theories,” says Feaster. “But what I’ve found from looking at the history of sound recording is that your idea doesn’t have to be right. You don’t have to understand why your invention works or how it works.” Instead, he says, “you just have to have a certain type of creativity that looks for the gaps in the type of knowledge that people have—the facts that haven’t yet been fully resolved—and play around with them, try things out.”
It’s a generalizable point but less emotionally resonant than the one discussed by the two great-grandsons, sitting together in Edison’s impressive laboratory in West Orange. The high-ceilinged room, built to impress investors, has elaborate wooden carvings and several tiers of balconies encircling the center. Both are uninterested in the idea of a competition and are moved instead by the idea that both of their ancestors belonged to a wave of scientific progress with many participants.
Sloane explains, “You have to remember you can only invent something that the scientific climate makes possible to invent, which means there have to be in your control the appropriate mechanisms, the appropriate chemical compositions, of the appropriate opportunities.”
“Science is very cumulative,” adds de Martinville, “general conditions of knowledge progress everywhere.” He brings up Albert Einstein, born the year Scott passed away. “Had Einstein been born fifteen years earlier, fifty years earlier, I’m not sure he would have made this extraordinary contribution to science.”
Sloane is nodding.
“The other information was not there that he needed to make his theories,” he says, “and he was thought quite crazy at the beginning himself.” He concludes, “I think it’s better to share the glory and share the relationship and share the excitement of being in a community of inventors.”