While Stephen Hawking’s ashes were being interred during a memorial service at Westminster Cathedral in 2018, the European Space Agency (ESA) beamed into space an original piece of music by the Greek composer Vangelis that featured Hawking’s voice. The Cebreros station outside of Madrid broadcast the transmission toward the nearest black hole to Earth.
In a press release, ESA’s director of science, Gunther Hasinger, is quoted as saying, “It is fascinating and at the same time moving to imagine that Stephen Hawking’s voice together with the music by Vangelis will reach the black hole in about 3500 years, where it will be frozen in by the event horizon.” The move was meant to honor Hawking’s primary contribution to astrophysics, his theory that thermal radiation is spontaneously emitted by black holes due to the steady conversion of quantum vacuum fluctuations into pairs of particles, one of which escapes at infinity while the other is trapped inside the black hole. Hawking radiation, and the related issue of whether information that falls into a black hole is lost or is somehow recoverable from the radiation, was a profound concept that still engenders controversy among theoretical physicists. Interestingly, Hasinger’s comments take for granted that the voice made of information, from the voice synthesizer Hawking had used to communicate for more than 30 years, was simply “Stephen Hawking’s voice.”
Upon his death, many notices in the press made more of Hawking’s voice than of Hawking radiation, suggesting he was famous in public for a persona that merged both. Reuters reported that Hawking’s synthesized voice “was his tool and his trademark,” describing it as a “robotic drawl that somehow enhanced the profound impact of the cosmological secrets he revealed.” Even an obituary for Hawking in Nature by longtime colleague Martin Rees referred to “the androidal accent that became his trademark” and also speculated that Hawking’s field of cosmology was part of what made his life story resonate with a worldwide public — that “the concept of an imprisoned mind roaming the cosmos grabbed people’s imagination.” Certainly the publishers of Hawking’s popular account of the universe, “A Brief History of Time,” must have thought so, as they chose to feature a photograph of a demure-looking Hawking sitting in his wheelchair in front of a night sky full of stars on the front cover of the first U.S. edition of the book. Even if only figuratively, Hawking was one of the world’s most famous living cyborgs, moving and speaking by electromechanical means because his flesh-and-blood body could not, traveling the universe in his mind.
Hawking, who had amyotrophic lateral sclerosis (ALS), lost his ability to speak after a trip to Switzerland in 1985 when pneumonia forced doctors to put him on a ventilator and, after being transferred back to the UK, performed a tracheotomy to help him breathe. After recovering from the near-fatal pneumonia, Hawking initially used a spelling card to communicate, indicating letters with a lift of his eyebrows, an interim solution with obvious constraints.
Hawking’s first voice synthesizer used an iteration of “Perfect Paul,” a synthesized voice created by MIT research scientist Dennis Klatt, and was run on an Apple II computer with modifications that made the system more mobile. It allowed Hawking to communicate at a rate of 15 spoken words per minute using a hand clicker to select words from a separate software program that displayed them on a screen. ALS is a degenerative nerve disease that affects muscle control, so Hawking’s need for an apparatus that allowed him to write was arguably much more critical than the voice synthesis component, but voice synthesis allowed Hawking an additional modality of communication.
Hawking’s public “voice” is certainly created differently than one from a human body’s vocal system, but even this voice generated from electricity, signal-processing algorithms, circuitry, software code, and the distant echo of the body of Dennis Klatt — Perfect Paul was synthesized from Klatt’s recordings of himself — became unique in its specific instantiation for Stephen Hawking, even in how it sounded. The synthesized voice became “his” through his body’s intimate interactions with and through it. Although the voice itself, as an electronic technology, lacked most of the expressive capabilities of a body’s, Hawking was known to use its affordances and constraints as expressions of his personality.
He gave his lectures by sending saved text to the speech synthesizer one sentence at a time, but he stated that he did “try out the lecture, and polish it, before I give it,” a statement implying that he was practicing for a performance that included revisions to the way that the lecture sounded to him as much as to its content. Known for his playfulness, Hawking appeared as himself in or recorded dialogue for numerous television shows, including “The Big Bang Theory” and “Star Trek: The Next Generation.” He recorded the dialogue for his own depiction in animated series “The Simpsons” for three separate episodes and was also made into a Simpsons action figure. He told reporters in 2014, “My ideal role would be a baddie in a James Bond film. I think the wheelchair and the computer voice would fit the part.”
Hawking’s synthesized voice was his “trademark,” as obituaries described it, though not literally trademarked by Hawking (it was not his intellectual property). On a web page that Hawking maintained during his lifetime, he credited David Mason of Cambridge Adaptive Communications of putting together the original Speech Plus system for him. “This company manufactures and supplies a variety of products to help people with communication problems express themselves,” wrote Hawking. As early as 1988, Speech Plus had offered Hawking a new synthesizer, a CallText 5010, with improved text-to-speech capabilities, but Hawking insisted that the company provide him with one that used the same voice as his previous unit. Of his version of Perfect Paul, Hawking stated, “I use a separate synthesizer, made by Speech Plus. It is the best I have heard, though it gives me an accent that has been described variously as Scandinavian, American, or Scottish.” He was known to joke to American audiences that it gave him an American accent.
Hawking famously refused “upgrades” to his communication apparatus, having come to accept the specific configuration of hardware and software, and the sound that it produced, as his own. As the technology that he used to speak became more obsolete, it became easier for Hawking to control other people’s use of the synthesized voice that had become so intimately associated with him. Hawking approved of a cut of the biographical film “The Theory of Everything” in 2014 before he let producers use his own synthesizer to rerecord dialogue. The identification of this specific synthesized voice with Hawking went both ways. Not only would many people hear it today and call it “Stephen Hawking’s voice” even after his death, but Hawking himself also insisted on keeping the very first synthesized voice he ever used for the remaining 32 years of his life, even as voice synthesis technologies improved.
Eventually, the 1980s-era hardware of Hawking’s synthesizer was failing, and other factors, from Hawking’s continuing nerve degeneration to the incompatibility of his specialized software and input devices with the latest computer systems, made it imperative to upgrade. Several people — graduate students, programmers, and a research team from Intel — were involved in getting a new system working to Hawking’s satisfaction.
One thing that Hawking insisted on was keeping the version of Perfect Paul that emerged from the specific hardware that he had started using in 1986. The programmer Peter Benie worked on a software emulator of the original microprocessor and digital signal processor of the CallText 5010, without the benefit of the original hardware schematics or software source code. The only thing he had was the machine code. The team struggled to find a copy of the original voice, especially as the company that was once Speech Plus had been sold or acquired four times since. A backup tape from 1986 was finally located by contacting Eric Dorsey, an engineer who had worked on the CallText 5010 and had fielded questions from the press about the synthesizer when Hawking visited California for a three-week lecture tour in 1988. When the team tracked down Dorsey in 2014, he was working for TiVo and was shocked to learn that Hawking was still using the CallText 5010. The new emulator was finally presented for Hawking’s approval in January 2018, only two months before he died. He never used the new system in public due to his failing health.
Since first adopting the CallText5010 with its version of Perfect Paul in the mid-1980s, Hawking had many opportunities to upgrade components of his communication system, but he refused because the voice produced by the original setup had become his. Only Klatt’s friends and colleagues heard the echo of someone else in Hawking’s speech. And they would have heard it all over the place in the late 1980s. Versions of Perfect Paul were a common default for text-to-speech systems, including many “audiotext” phone information systems, for a couple of decades. Even though Klatt’s system included other voices like “Beautiful Betty” and “Kit the Kid” (based on recorded samples from Klatt’s wife and daughter, respectively), Perfect Paul was perceived as the easiest to understand, although it had its limitations. The New York Times described Klatt’s synthesized voices as “usually understandable but that sound as if they have a foreign accent,” higher praise than another system, which the paper described somewhat offensively as sounding “like a scratchy recording of a person with a lisp.”
As hardware and voice synthesis software improved over time, the echo of Dennis Klatt grew fainter and the association between Hawking and his specific instantiation of Perfect Paul more solid. “I keep it because I have not heard a voice I like better and because I have identified with it,” Hawking explained in 2006. The desire to have a voice that was his own outweighed any benefit from upgrading his systems. In 2013, journalist Lucy Hawking, the daughter of Stephen, interviewed Klatt’s daughter, Dr. Laura Fine, on BBC Radio. Lucy was 15 when her father lost the ability to speak. Laura was 18 when her father died of cancer. Reflecting 25 years later, Lucy said, “[It] means that my father is actually speaking with your father’s voice.” Laura replied, “[My father] would be so, so thrilled.… I had never really thought before how my dad’s voice lives on.”
Hawking’s attachment to Perfect Paul shows how intimate our subjective experience of our own voice is, as well as the relationship that it affords us with the social world, which might reflect our society’s biases. Hawking’s use of Perfect Paul conveyed his sense of humor, his intellect, and also his physical disabilities. Even though it was the standard voice on equipment used by many people, it sounded unique as used by Hawking as an individual, to the point that he wouldn’t upgrade his equipment because it would have meant changing the sound of his voice. In this example of cyborg collaboration, the machine became an extension of Hawking’s body, but one as prone to aging and breakdown as our biological bodies are. Far from transhumanist dreams, Hawking’s example shows the experience of being human is a thoroughly embodied one.
Sarah A. Bell is a writer and professor who studies the impacts of information technologies on society. She is the author of “Vox ex Machina: A Cultural History of Talking Machines,” from which this article is adapted. An open access edition of the book is available for download here.
This article Stephen Hawking’s eternal voice is featured on Big Think.
The post “Stephen Hawking’s eternal voice” by Sarah A. Bell was published on 01/04/2025 by bigthink.com