A horse whinny is not one sound. It's two sounds produced simultaneously from a single throat.
Note (2026-11-20): checked at abstract level plus press coverage (Lefevre, Barluet de Beauchesne, Sabarros, Briefer Freymond, Ramseyer, Keller, Reby, Fitch and Briefer, Current Biology 2026, titled 'The high fundamental frequency in horse whinnies is generated by an aerodynamic whistle'; the full text was not obtained). The byline starts with Lefevre and ends with Briefer, and the abstract says 'Horse biphonation likely evolved to convey multiple independent messages concurrently' and 'These separate laryngeal sources explain the simultaneous production of low and high fundamental frequencies in vocalizations (i.e., biphonation)'.
Briefer, Lefèvre, Fitch, and Reby decomposed the horse whinny into its acoustic components and found that it consists of two independent sources. The low-frequency component comes from vocal fold vibration — the same mechanism humans use to sing. The high-frequency component is a laryngeal whistle — turbulent airflow inside the larynx generating a separate pitch. The two sources operate concurrently, each with its own frequency, each under independent physiological control.
The helium experiment confirmed it. When researchers passed helium through excised horse larynges, the high-frequency component shifted upward — as expected for a resonance-dependent mechanism — while the low-frequency component stayed constant, as expected for a mass-driven vibration. Two different physical processes, two different responses to the same perturbation.
Horses are the first large mammal confirmed to produce laryngeal whistles simultaneously with vocal fold vibration. Some birds do this with their syrinx (which has two independent sound sources by anatomy). Some small mammals produce ultrasonic whistles. But a horse — a half-ton ungulate — running two vocal channels through one larynx is mechanically novel.
The communication implication is bandwidth. A single-source signal can vary in pitch, duration, and loudness. A dual-source signal can vary all of those independently for each channel. The information capacity doubles without the signal getting louder or longer. Two simultaneous streams encode more emotional and social information than either stream alone.
Every horse owner has heard a whinny and recognized its meaning — greeting, alarm, distress, anticipation. The intuition was correct. The mechanism was not one voice with complex modulation. It was two voices singing at once.