In 2020 as the pandemic began to spread, it became increasingly clear that infected people sneezing and coughing were a prime reason for contagion. Yet super-spreader scenarios in a Korean call center and a church in Seattle, among many others, pointed to speech—talking, singing, and breathing proximity—even among those who were asymptomatic as a major factor for spread as well.
At the time, there were few highly cited studies that assessed the peculiarities of speech on transmission or isolated which speech patterns and mechanisms were most pernicious, according to Caleb Everett, professor in the Department of Anthropology and senior associate dean for academic affairs at the University of Miami College of Arts and Sciences.
“Those scenarios in Korea and Seattle heightened the gap in our knowledge to ask: ‘What is it with speaking, what are the aspects of speech that are actually causing this?’” said Everett. “What we saw in this pandemic was mainly reasonable speculation.”
He referenced an article in The Atlantic magazine published at the height of the pandemic that suggested that in addition to mask-wearing, talking less and whispering might minimize the spread.
“Even at the time, I thought we don’t really have good evidence for whispering,” Everett recalled. “And it wasn’t that the methods were bad, it’s just that they had very low temporal resolution,” referring to the ability to make multiple measurements within a limited timeframe.
Colleagues at the University of California San Diego medical school contacted him to glean what an anthropologist who has studied the interaction of language with nonlinguistic cognition, culture, and the environment for decades might divulge about speech production and patterns. They added a few chemists and formed a research team.
“We’re trying to isolate what actual things are happening in the vocal tract to produce the aerosol particles,” he explained. “So that for a future pandemic, we might have some idea which sounds are actually producing, or which movements are actually more responsible for the aerosol productions—that may provide some health guidance.”
To find new ways that aerosols are produced, the team is using an electrical particle impactor, a device primarily used in atmospheric science to test aerosols coming off ocean spray. This allows the team to sample speech at 10 times the rate of previous studies.
“This makes a big difference when it comes to isolating what’s happening in the vocal tract,” said Everett, noting that many sounds that we make in speech are as short as 50 milliseconds or 100 milliseconds.
The team is collecting aerosol and acoustic data simultaneously to better analyze the role of a range of factors: how loud the voice is, individual sound types, the amount of airflow produced, and the role of the speaker’s body size, among others.
Additionally, the utilization of a pneumotachograph—used extensively in biomedicine—has served to measure the air flow that comes out of someone’s mouth.
To date, the team has tested only a small sample base of healthy people, yet indications are that whispering actually produced more aerosols because it generates really high velocity air coming from the lungs and past the vocal cords.
“There clearly seems to be some variation in terms of the kinds of sounds or at least things going on in the vocal tract and there are variations within speakers,” he noted. “Some speakers seem to be super emitters, where they produce a lot of aerosols, and we don’t know why exactly, we just know that there are some.”
Everett emphasized that, while promising, the results are very preliminary. The team needs to greatly expand its sampling to produce more conclusive results.
“If we can isolate—and we think our method can—whether the aerosols are coming from deep within the lungs with certain sound types or whether they’re just coming from the mouth, that’s relevant,” he explained, noting that “the ones coming from the lungs have a greater likelihood of carrying the virus or whatever the virus may be, possibly another airborne disease.”
Everett noted that English is among the 20 percent of languages that produce an aspirated sound such as the first in “pill,” pronounced with an accompanying forceful expulsion of air. The team is interested in this type of sound.
“There are thousands of different sounds in languages,” he pointed out. “It’s possible that certain sounds are more likely to produce aerosols, we don’t know that. But the potential import is there.”
His component of the research includes sifting through the sounds of the International Phonetic Alphabet and testing it with the pneumotachograph device.
The research team’s first paper provided a proof of concept showing the temporal and physical resolution, indicating that the method used can isolate the role of individual sound types in producing aerosols.
“Our hope is that it’s a way forward and that other researchers can get the idea and develop it further,” Everett said.
Additionally, Everett’s newest book explores the interface of language and cognition and culls from the wealth of new knowledge of the past decade or two to show in much greater detail how the physical environment influences the way we think.
Due out this fall, “Myriad of Tongues: How Languages Reveal Differences in How We Think” has already been named a top 10 science book by Publisher's Weekly.
