Sperm whales communicate through codas, short rhythmic sequences of clicks. A study published on 15 April 2026 in Proceedings of the Royal Society B shows that these codas fall into two acoustic categories which, not content merely to resemble human vowels, behave like them along several dimensions. The result is fascinating. The vocabulary used calls for care.
What a vowel is, acoustically
This step is necessary, otherwise the rest makes no sense.
When a human speaks, the vocal folds produce a rich sound that the cavities of the mouth and throat then filter. Depending on the position of the tongue and the lips, certain frequency bands are reinforced: these are the formants. The difference between an “a” and an “i” lies not in the pitch of the voice but in the position of these formants. A vowel is therefore defined by a resonance structure, not by a note.
The sperm whale, for its part, has the largest sound-producing organ in the animal kingdom, a nasal complex of more than five tonnes in a large male, in which the sound bounces between air sacs before leaving the body. This path too creates resonances, and therefore formants.
What was already known about codas
This result belongs to a line of work that has to be recalled in order to gauge what it adds.
Codas have been described since the 1970s by their number of clicks and their rhythm, written down as sequences, for example three regular clicks or a run of five with a pause. These patterns are not distributed at random: groups of sperm whales that share the same repertoire form what researchers call vocal clans, which can occupy the same waters without mixing.
Some codas also appear to carry an identity function, at the scale of the clan as well as of the individual. It is one of the best documented cases of cultural transmission in a marine animal, set out in detail on our page on sperm whale culture.
The long-term monitoring carried out off Dominica, where the same families have been observed for more than fifteen years, has supplied most of this corpus, and it is that same dataset which feeds the recent analyses.
What the study establishes
On the basis of the number of formants, the authors distinguish two categories which they name a-codas and i-codas, by analogy with human vowels. What they add is not this acoustic resemblance, already described, but the demonstration that these categories are organised according to language-like regularities.
| Interaction with structure | the coda types defined by the number of clicks and their rhythm interact with vowel quality |
|---|---|
| Duration | a-codas are generally longer than i-codas |
| Distribution | the duration of i-codas follows a bimodal distribution, that is, two preferred values rather than a continuum |
This last point is the most significant. A two-peaked distribution indicates discrete categories rather than a gradient: the system works through distinct units, which is a property of phonological systems.
Measuring a resonance in a train of clicks
One technical difficulty is worth flagging, because it explains why this result is arriving now.
A sperm whale click is a very brief pulse, very broad in frequency, the opposite of a human voiced sound, which lasts and repeats. Looking for formants in a pulse means analysing the fine spectral structure of the pulse and of its internal replicas, produced by the sound travelling back and forth inside the animal’s nasal complex.
These replicas are not an artefact: they are the signature of the geometry of the sound-producing organ, and it is indeed from their spacing that the size of a sperm whale has long been estimated by ear. The step from that size information to information about a sound category is what the recent work explores.
The approach demands large quantities of clean recordings, attributed to known individuals, and analytical methods able to handle that volume. It is the conjunction of the two that makes the study possible today. Our sound library lets you hear these signals.
From phonology to meaning, the missing step
This has to be said plainly, because it is the point on which coverage of this work regularly goes off the rails. Showing that a system has phonological properties is not the same as showing that it carries meaning.
Phonology describes how a sound system organises itself into units and contrasts. It says nothing about what those units mean, nor even whether they mean anything. A system can be richly structured and carry very little information, as another piece of work by the same team on the slow clicks of males, described as a low-entropy signal, indicates.
The authors themselves use “vowel” as a technical analogy, framed by measurable criteria. The word is warranted within that frame; it becomes misleading as soon as it is taken out of the acoustic context. Our page on the brain and cognition of cetaceans holds to this distinction.
The limits of the study
Three reservations are worth setting out, and the authors conceal none of them.
1. The corpus. The analysis rests on recordings from a small number of populations, the eastern Caribbean supplying most of the material. Nothing guarantees that the same categories are found in clans of the Pacific or of the Indian Ocean, and the comparison between regions remains to be made.
2. The share of the individual. The spectral structure of a click also depends on the size of the animal producing it, and therefore on its age and its sex. Separating what belongs to a chosen sound category from what belongs to the anatomy of the speaker is a difficult problem, and it governs the interpretation.
3. The absence of a behavioural test. Showing that two categories exist in the signal does not show that sperm whales distinguish them. A demonstration would require a perception experiment, that is, observing a different reaction to two otherwise identical categories. Nothing of the kind has been carried out, and the practical difficulty is considerable in an animal that dives to a thousand metres.
These reservations do not weaken the measurement, which bears on a recorded signal. They delimit what can be concluded from it, which is the mark of a solid result on a subject where enthusiasm often outruns the data.
Why this result matters nonetheless
Because it shifts the question. For a long time sperm whale communication was described by the number of clicks and their rhythm, which yielded “coda types” used in particular to distinguish cultural clans. This study adds a second dimension, a spectral one, independent of the first.
A system with two combinable dimensions offers far more combinations than a system with only one. Whether or not sperm whales exploit that capacity remains open, but it exists, and it is a necessary condition for any complexity.
The link with the filmed observation of a birth off Dominica is direct: the team had noted changes of vocal style at the key moments, including vowel-like structures. Our page on the acoustic profile compares repertoires, and our sound library lets you listen to codas.
Source: Beguš G., Dąbkowski M., Sprouse R. L., Gruber D. F. and Gero S., “The phonology of sperm whale coda vowels”, Proceedings of the Royal Society B, 15 April 2026.
Did you find this article useful?
Add L'info des Cétacés to your preferred sources on Google. Our stories will be suggested to you more often, and the choice can be undone as easily as it is done.
Follow us on GoogleThis preference is saved in your Google account and can be changed at any time. It is neither a certification nor a recommendation from Google.
Leave a review