How a Small Cat Makes Such a Low Purr
The Purr Can Keep Its Own Time
A cat is a small animal with a surprisingly slow instrument. The repeating pulses of a purr can arrive around twenty-five times a second. At that rate, each cycle lasts forty milliseconds. That is not the pace of the cat's breathing: many of those little repetitions fit inside a single inward or outward breath.
The puzzle is how such a small voice box gets down there. One answer long taken seriously was that the muscles supplied the rhythm—contracting and relaxing over and over, under neural control, to interrupt the airflow. Instead of a small instrument naturally vibrating slowly, the cat would be actively making it do so.
A 2023 experiment showed that the instrument had more capacity of its own than that explanation required. But the experiment did not turn a living cat into an unnecessary accessory to its purr.
An organ without a metronome
Christian Herbst and colleagues investigated eight domestic-cat larynges in an excised-larynx setup. In their published abstract, they report that all eight produced self-sustained oscillations at typical purring rates. They could generate sounds in the 25–30 Hz range without neural input or muscular contraction. [1]
This is not a claim that tissue makes sound without an energy supply. Airflow supplies energy; deformable tissues interact with it and provide restoring forces. Under suitable conditions, the system keeps oscillating. It does not need an instruction to arrive separately for every repetition.
That is the core of myoelastic-aerodynamic phonation, a formidable name for a principle behind ordinary mammalian voices. The fast rhythm can emerge from the interaction of air and tissue rather than being imposed by a train of equally fast neural commands. A living animal still has to arrange the conditions for its voice and power its breathing. Keeping time and organizing the performance are different jobs.
The researchers also found connective-tissue masses embedded in the vocal folds. Their proposed explanation is that this specialization helps a small larynx reach unusually low frequencies. The University of Vienna's account describes the structures as pads and compares the resulting mechanism to human creaky voice, or vocal fry. [1, 2]
The important word is helps, not proves. Finding a structure alongside a capability is not the same as selectively removing that structure and showing that the capability disappears. The accessible abstract gives an anatomical explanation worth pursuing, not a pad-only causal demonstration.
My account of this experiment is limited to its complete abstract, reproduced by Mark Liberman, and the university's complete public release. I could not access the full paper's methods or figures. Those two accounts come from the same research, not two independent experiments. In particular, I cannot establish from them how completely the laboratory preparation reproduced natural purring through both directions of airflow.
Even within that limit, the result matters. A claim that purring-rate vibration is physically impossible without repeated active contractions is too strong. Eight isolated larynges provide a substantial reason to abandon that necessity. A claim that living cats never use rhythmic muscle activity is another proposition—and the authors explicitly leave open a contribution from active contractions. [1]
Vincent did it the other way
An earlier, independently authored study helps show why the intact animal still matters. In 2011, Susanne Schötz and Robert Eklund compared purring in four domestic cats: Donna, Rocky, Turbo and Vincent. They recorded the animals in a quiet home environment, using a microphone near the muzzle. To mark breathing direction, the researcher felt the expanding or contracting chest and said “in” or “out” during the recording. [3, Methods]
That simple detail gives the sound a bodily coordinate system. A purr is not merely a number on a frequency scale. Its repetitions continue through alternating inward and outward airflow.
Donna's mean repetition rate was 24.63 Hz while breathing in and 27.21 while breathing out. Turbo also ran faster outward. Vincent reversed the pattern: 23.45 inward, 20.94 outward. Rocky's modest difference was not statistically significant in the study's tests. That does not establish that his two rates were identical. [3, Table 1 and Results]
There is no universal rule about which breathing direction makes the faster purr hiding in those four cats. Nor does the fact that Vincent was sixteen while the others were six months old establish an effect of aging. One older individual cannot separate age from individuality.
What the study supplies is a more demanding target for an account of the whole behavior: a low repeating vibration within changing airflow directions, with differences between animals. Those differences need not mean different vibration mechanisms. Matching the approximate frequency is relevant. It is not the whole performance.
The recordings do not themselves settle the muscle question. The study measured acoustic patterns, not neural commands or the effects of selectively disabling a muscle. Its many analyzed breath phases still belong to only four cats. It complements the isolated-larynx experiment; it does not replicate its mechanism.
The rhythm and the animal
What changed my understanding was the removal of an assumed necessity: the idea that each slow pulse must be imposed by an active contraction. The tissues and airflow can generate a purring-rate rhythm without a muscle being commanded to twitch for every pulse.
That does not make the living control problem disappear. It locates it more precisely. How does the animal establish and maintain the right conditions? How do tissue motion, breathing direction and any rhythmic muscle activity cooperate in an ordinary purr?
The isolated organ tells us that the instrument can keep this kind of time. The living cat tells us there is still a performance to explain.
Sources and scope
[1] Christian T. Herbst et al., “Domestic cat larynges can produce purring frequencies without neural input,” Current Biology 33 (2023), 4727–4732.e4, DOI: 10.1016/j.cub.2023.09.014. Complete abstract read as reproduced in Mark Liberman's “How cats purr,” Language Log, October 9, 2023. Full 2023 paper, methods, figures and supplements not read. Liberman's accompanying quotation from Ingo Titze supplies context for self-oscillation, not another cat experiment.
[2] University of Vienna, “Cats purr differently than previously thought,” October 4, 2023, complete release carried by ScienceDaily. Same research lineage as [1]; anatomical interpretation, not independent replication.
[3] Susanne Schötz and Robert Eklund, “A comparative acoustic analysis of purring in four cats,” Proceedings of Fonetik 2011. Complete main text, methods, Table 1 and captions read. Small observational conference study; no new recordings, playback, figure-image analysis or raw-data reanalysis performed for this essay. Cycle duration is reciprocal-rate arithmetic, not a listening result.