Despite their differences, dozens of languages spoken around the world seem to beat according to the same cadence of about two seconds.
Researchers analyzed recordings in 48 languages from all inhabited continents. Their work, published in the Annals of the New York Academy of Sciences, covers more than 100,000 intervals. The result is surprisingly stable: their median duration is around two seconds, regardless of the language family or the person speaking.
Pixabay illustration
To understand what the scientists measured, one should not count words or syllables. They divided speech into sound sequences separated by silences. Each measurement corresponds to the time between the beginning of one sequence and the beginning of the next. This method makes it possible to study an overall cadence without depending on a language’s grammar.
This choice changes the comparison. At the scale of syllables or other small units, languages display many rhythmic differences. Over longer sequences, these variations become far less pronounced. The researchers thus obtain a kind of average beat close to 0.5 hertz, or about one cycle every two seconds.
The team then applied a comparison used in research on animal communication. Human cadence falls within a range already observed in certain species. The data included, in particular, the song of the skylark, the vocalizations of the Cape fur seal, and those of a fish called the common meagre.
The study focuses on the temporal organization of sounds. It indicates that human speech does not occupy a category of its own when its rhythm is measured in this way. Shared biological constraints could therefore help set this time scale.
Breathing is one possibility. A person cannot produce a vocal sequence indefinitely without taking a breath, which naturally imposes limits. The authors also mention working memory, the brain’s ability to briefly retain information while processing it. These explanations remain hypotheses, and the study cannot determine which of them is responsible.
The recordings come from the DoReCo corpus, which focuses in particular on poorly documented languages. This choice greatly expands the sample beyond the languages most often studied in laboratories. The observed regularity persists despite this diversity, making it possible to compare human speech more directly with the vocal rhythms of other species.