Researchers at MIT CSAIL and Project CETI have made a groundbreaking discovery in understanding the communication system of sperm whales, thanks to machine learning technologies. Their study, titled Contextual and Combinatorial Structure in Sperm Whale Vocalizations, sheds light on the intricate codas used by these marine mammals for communication.
The research focuses on the codas, which are series of clicks that serve distinct linguistic purposes. Through their study, the team revealed a previously unknown variation in the coda structure, indicating that coda types are not random but actually form a complex combinatory coding system.
Drawing inspiration from the work of the late marine biologist Roger Payne, who extensively studied humpback whale songs, the researchers utilized machine learning to analyze a dataset of 8,719 sperm whale codas collected off the coast of Dominica. By studying the sounds in context, as exchanges between whales, they identified a sperm whale phonetic alphabet that explains the variability in coda structure.
The team’s method marked a shift from traditional analysis that focused on individual codas, allowing for a deeper understanding of the communication system. The phonetic alphabet uncovered by the researchers demonstrates a sophisticated communication system within sperm whales, akin to the linguistic concept of duality of patterning found in human languages.
While this discovery is significant, the researchers acknowledge that there is much more to learn about sperm whale communication and potentially extend their findings to other species like humpback whales. As they continue their research, they aim to unravel the complexities of marine mammal communication systems and further explore the fascinating world of cetacean language.
