Humans’ ability to speak has been attributed to their unique anatomical features, specifically with respect to the vocal tract—although this is debated—and brain capacity. A study published online today (August 11) in Science posits that, somewhat counterintuitively, an anatomical simplification in the human larynx may have been an important step in the evolution of complex speech. Based on a comparative analysis among modern primates, experiments, and mathematical modeling, the authors propose that the loss of a laryngeal membrane may have helped stabilize vocalization in human ancestors, facilitating extensive information sharing through spoken language.
“Language evolution is one of the most enigmatic” events in human evolutionary history, says coauthor Takeshi Nishimura, a primate researcher at Kyoto University in Japan. Other transition events, such as bipedal locomotion and increased brain size, can be studied through the fossil record, he says, “but the language was never fossilized.”
Comparisons among extant nonhuman primates are therefore “one of the most powerful approach[es] for exploring language evolution,” adds Nishimura. That’s where he and his colleagues started when they set out to reconstruct the evolutionary steps that led humans to speak and sing, focusing specifically on changes that the larynx has undergone throughout primate evolution.
The comparative task was made easier by the fact that not far from Nishimura’s lab is the Japan Monkey Centre, which is home to the largest variety of nonhuman primate species in the world and, as it happens, to a collection of larynxes removed from primate cadavers. Using MRI and computed tomography, Nishimura and his colleagues inspected the larynx anatomy of 43 species from 25 genera. The painstaking work revealed that all of them had a structure that is known to be absent in humans: a thin membrane separated from the vocal fold by a groove known as a sulcus. The most parsimonious explanation is that this feature was present in the ancestors of today’s primates and that humans are the only extant species to have lost it.
One of the most important contributions of this study is its comprehensive survey across many species demonstrating that the presence of these vocal membranes “is the default state for primates,” says Princeton Neuroscience Institute biologist Asif Ghazanfar, who did not participate in this research but has previously collaborated with one of the authors. It’s “hard to compile” the data for such a large comparative analysis, he notes.
Nishimura and his colleagues also sought to understand the function of these membranes in primate sound production. Their approach was to test whether this anatomical structure vibrates during experimentally simulated vocalization in chimpanzees (Pan troglodytes), rhesus macaques (Macaca mulatta), and common squirrel monkeys (Saimiri sciureus)—a question they studied both in live, anesthetized animals and in larynxes removed from animals that had recently died naturally or by euthanasia unrelated to the experiments. In all instances, these thin membranes did vibrate, providing evidence that they are fundamental to sound production.
Together with previously published models and data from groups that included Ghazanfar and a coauthor of the current study, this team’s observations suggest that these membranes may facilitate the execution of louder, high-frequency calls used, for instance, when facing antagonistic groups, says Nishimura. But in addition, the new experimental evidence, accompanied by mathematical modeling, indicate that the vibration of these membranes makes vocalizations more prone to instability—similar to what happens when adult humans scream or babies cry. As a result, information is often not transmitted in a clear and structured way, Nishimura explains.
Overall, the authors propose that the loss of these membranes and the resulting vocal stability and harmony may have been a necessary step in the evolution of speech.
Kristina Simonyan, a neuroscientist and otolaryngologist at Harvard Medical School who was not involved in this paper but whose PhD advisor 15 years ago was one of the authors, says this is a “really interesting study,” yet she is not entirely convinced about two of the ideas it proposes. First, the anatomical modification the authors describe is not equivalent to a “loss of complexity in human vocal anatomy,” as the researchers put it in their paper’s title. “Complexity does not come with more components,” she notes, and just because the human larynx doesn’t have the groove that results in what the authors call the vocal membrane, it doesn’t mean that the structure is not extremely complex, she says.
Second, while she agrees with the authors “that probably losing the sulcus over the course of evolution made the vocalizations more regular,” she is not sure how this loss might have factored into speech evolution. “Speech is not vocalization,” she says, adding that, for example, people with a speech disorder where this groove is present as a pathology can still speak; their only problem is hoarseness. These findings are likely “pieces of a larger puzzle,” she concludes.
Nishimura says the study’s results suggest that the loss of the vocal membrane may have been “adaptive” for the evolution of speech, but acknowledges that it is unclear whether its contribution was direct or indirect.
It remains unknown whether “the selective force for losing the vocal membranes in humans is so that they can produce this ultrastable broadband signal,” Ghazanfar says, but adds that “I think it’s a . . . super cool hypothesis.” The study does not present irrefutable evidence for that hypothesis, but that’s the nature of evolutionary studies like this one, he notes. However, the authors provide “tantalizing evidence in favor of that idea.”