UC Berkeley and Project CETI find sperm whale calls have vowel-like patterns, hinting at language-like structure
Researchers from UC Berkeley and Project CETI report that sperm whale calls contain continuous, vowel-like and diphthong-like acoustic patterns, suggesting a layered vocal structure more complex than simple click codas and opening new paths for interspecies communication study.
Key takeaways
- Vowel-like structure: The UC Berkeley study finds acoustic bands and shifts resembling human vowels and diphthongs.
- Advanced methods: The team used signal processing and unsupervised machine learning combined with Project CETI recordings to reveal hidden structure.
- High field accuracy: Project CETI reports noise-detection, click categorization and dialect recognition accuracies up to 99.5%, 97.5% and 95.3% respectively (Audacious Project).
- Open questions: Findings indicate phonetic-like building blocks but do not prove grammar or decoded meanings.
Main content
What the study found
Researchers previously described sperm whale communication primarily as patterned clicks called codas, used for group identification or coordination. The new Project CETI–UC Berkeley analysis shows finer acoustic detail within those clicks: recurring spectral shapes and smooth frequency trajectories that behave like vowels and diphthongs.
“Instead of isolated beeps, we found gradient acoustic features that function like the sounds that make up syllables in human language,” said Gašper Beguš.
Key observation: rather than discrete on/off click events, some call elements display continuous resonance bands and shifting energy patterns analogous to steady-state vowels and moving diphthongs in human phonetics.
How the team analyzed sound
The study combined traditional acoustic analysis with unsupervised machine learning. Steps included:
- Advanced noise detection to clean ocean recordings.
- Unsupervised clustering to group similar call elements without pre-labeling.
- Spectral and timing analyses to identify continuous bands of energy and smooth frequency shifts.
Project CETI provided long-term recordings collected off Dominica and applied non-invasive robotics and AI pipelines to sort tens of thousands of calls, enabling detection of patterns, dialects and possible cultural transmission with high reported accuracy.
Why it matters for language science
Human language relies on basic acoustic units (vowels, consonants) assembled into larger structures. Finding vowel-like and diphthong-like elements in sperm whale calls suggests a possible phonetic-level building block in a nonhuman species. This is an important step toward assessing whether whales combine units flexibly—a hallmark of complex communication systems.
Important caveat: the presence of phonetic-like units does not equate to demonstrated grammar or syntax; behavioral and cross-disciplinary evidence is required to link sounds to specific meanings or rules.
Project CETI’s goals and future plans
Project CETI brings together linguists, marine biologists, AI researchers and engineers to catalog calls and build tools for interpretation. The project plans to expand datasets, refine models to test combinatorial use of units, and make data and tools publicly available to enable reproduction and extension of findings. Project CETI emphasizes public data sharing and partnerships to broaden engagement (Project CETI).
The project’s technology stack includes non-invasive robotics for higher-quality recordings and machine learning frameworks tuned for noisy marine environments, with field detection and classification metrics reported on the Audacious Project page.
What the evidence does and does not show
- Shows: robust acoustic structure and recurring patterns within click elements consistent with vowel-like and diphthong-like behavior.
- Does not show:
- Next steps:
Supporting context on whale social life
Sperm whales live in tight, matrilineal family groups and have very large brains. Long-term studies—such as those informing Project CETI’s Dominica recordings—have documented distinct dialects that can be transmitted across generations, providing a social substrate that can favor complex communication.
Implications for San Diego
Local and regional effects may include research collaboration, tourism impacts, policy discussions and public engagement. Specific areas of impact:
Economic impact
- Research partnerships: San Diego institutions like Scripps and local marine tech firms could collaborate on signal-processing and robotics work.
- Tourism: whale-watching operators may market educational trips that highlight complex whale communication, potentially boosting bookings.
Political consequences
- Ocean noise policy: renewed pressure to review shipping lanes and port activity to mitigate underwater noise.
- Funding priorities: potential prioritization of grants for marine bioacoustics and AI research in regional budgets.
Social effects
- Education: schools and museums can incorporate audio demos and public talks on the science.
- Community science: volunteer-based acoustic monitoring and citizen projects could expand local engagement.
Cultural relevance
- Coastal identity: discoveries may deepen local pride and fuel cultural work around marine life.
- Indigenous perspectives: outreach should respect and include tribal and fishing-community knowledge when integrating scientific narratives.
Practical applications for residents
- Support noise-reduction and safe boating practices to protect whale communication channels.
- Attend talks, exhibits and classroom activities to learn about acoustic science.
Sources and further reading
- UC Berkeley news release
- Project CETI website
- Project CETI on The Audacious Project
- PR summary on Morningstar
Note: Researchers emphasize this discovery is a significant step but not a final decoding of whale language; ongoing, multidisciplinary work will be required to link acoustic structure to behavior and meaning.
