Language remains one of the great enigmas of evolution. While humans and other primates share many similarities, the ability to communicate through speech is not among them. However, a recent study conducted by researchers at Cornell University has shed light on the intricate language of orangutans. These great apes of Southeast Asia are known for their sophisticated vocal communication, but the subtleties of their repertoire have proven difficult for scientists to comprehend.
After three years of careful study, the researchers were able to decipher the complex patterns hidden within the roars, sighs, and other vocalizations of Bornean orangutans. This breakthrough allowed them to gain unique insights into the communication skills of these remarkable creatures.
The findings, published in the journal PeerJ Life & Environment, represent a significant advance in our understanding of orangutan communication. The research team bolstered their study by comparing AI detection methods with the work of biologists and bioacoustics scientists, who relied solely on their trained ears, intellects, and measurement tools.
To compile their dataset, the team analyzed 117 long calls from 13 male Bornean orangutans. They used 46 acoustic measurements on 1,033 different pulses detected in these calls. The researchers noted that these features greatly enhance the potential complexity of the orangutan’s signal, suggesting that we may soon decipher what these great apes are saying.
Dr. Wendy Erb, the lead author of the study, explained, “Our research aimed to unravel the complexities of orangutan long calls, which play a crucial role in their communication across vast distances in the dense rainforests of Indonesia.”
The research team employed a multifaceted approach, utilizing a cutting-edge unsupervised machine learning algorithm called Uniform Manifold Approximation and Projection (UMAP). This algorithm had previously been successful in decoding animal vocal repertoires for the University of California, San Diego in 2020. In addition to other supervised machine learning techniques, the team used statistical algorithms written in the programming language R.
Through a combination of supervised and unsupervised machine learning, the researchers classified three primary pulse types: “Roar” for high-frequency pulses, “Sigh” for low-frequency pulses, and “Intermediate” for those falling between the two categories.
While the focus of the research was not on the meaning behind the orangutans’ vocalizations, it did reveal that these animals employ a much wider range of sounds than previously thought.
This study has implications for our understanding of human evolution. Humans are the only primates capable of producing complex sounds, and there is a clear connection between how more primitive primates acquired these skills and how we did. To unravel this connection, scientists must first comprehend how the “graded” vocalizations used by animals like orangutans convey meaning so effectively.
By studying orangutan vocalizations, scientists may eventually uncover the secrets of human language acquisition. Each species develops its own vocal complexity as a result of evolutionary factors such as sexual selection, habitat details, social structures, and predation pressures.
