Your dog may prick up its ears at its name or run to the door when it hears “walk.” But can its brain separate unfamiliar words when it hears them without a pause?
In a new experiment, dogs listened to a continuous recording of invented words.
No pause indicated where one word ended and the next began.
Yet their brains responded to recurring consonants, a clue to which syllables belonged together.
The same clear word-level response did not appear when vowels defined the words.
The study published in Science does not show that dogs learned what the words meant.
It reveals a preference previously demonstrated only in humans: when dividing continuous speech into parts, the brain can prioritize consonant patterns over the louder vowels.
At first, that preference sounds odd. Vowels carry more acoustic energy and stand out more clearly than consonants. Why would a listener favor quieter sounds?
Across many human languages, consonants often provide more information about which word is being spoken.
The researchers wanted to know whether dogs, who spend their lives hearing us talk, might pick up on that usefulness too.
Boglárka Morvai and colleagues at Eötvös Loránd University in Hungary devised a way to ask without relying on dogs to follow commands.
They played recordings of meaningless, three-syllable words to 20 adult dogs while electrodes resting on their heads measured electrical brain activity.
Twenty adults heard the same recordings.
The speech flowed continuously. No silence announced the next word. In one recording, recurring consonant patterns grouped syllables into invented words. In another, vowel patterns did that job.
A third scrambled the syllables, leaving no recurring word pattern to discover.
The vowels were acoustically stronger. If dogs’ brains merely followed the most prominent sounds, vowels had the advantage. If the animals tracked the patterns useful for dividing up the stream, consonants might win despite being quieter.
The researchers measured whether brain signals lined up with the invented words’ rhythm despite the absence of audible boundaries. They also compared responses to the different streams.
The result was clear.
When consonants repeated, dogs’ brain signals synchronized with the rhythm of the three-syllable invented words, indicating that they followed the recurring units in the stream.
With vowel-defined words, the word-level response was not significantly stronger than in the scrambled recording. Humans also favored consonants, but their brains tracked the vowel-defined pattern more strongly than the scrambled stream too.
“What made the result especially interesting was that this bias had not previously been demonstrated in other species, not even in non-human primates, yet we found it in dogs,” Morvai and study co-author Dorottya Rácz told ScienceAlert in a joint response.
They considered two possibilities: Perhaps consonant preference requires the specialized language abilities of a human brain. Or perhaps repeated experience with speech can help an animal learn which sounds provide useful clues.
“This suggests that at least some aspects of speech processing do not require a brain that is specifically pre-wired for language; regular experience with speech may be enough for them to develop,” Morvai and Rácz said.
The experiment cannot establish what caused the dogs’ preference. Nor does finding the same bias mean dogs process speech exactly as humans do.
Both species’ brains followed the rhythm of syllables, but only humans showed clear tracking of individual speech sounds. “For us, each speech sound is a separate unit, whereas for dogs, syllables form the smallest units of processing,” Morvai and Rácz explained.
That does not mean dogs cannot distinguish individual sounds. Previous research from the team indicates that they can, although they may give small sound changes less attention when processing familiar words. The invented words in this experiment had no assigned meanings, so the results reveal pattern tracking, not comprehension.
The researchers also examined whether dogs needed to hear human speech from early puppyhood. Eleven dogs were thought to have entered a speech-rich home before 12 weeks of age, while nine were thought to have done so later. Both groups showed a consonant bias.
“Although the sample was small, we were still able to detect a consonant bias in dogs that had not lived in a speech-rich home during their first weeks of life,” Morvai and Rácz said.
Intensive early exposure may therefore not be essential, though the dogs’ histories before adoption were not always complete.
Dogs whose early lives are better documented could help answer when the preference develops. Comparing dogs with wolves could address a different mystery: whether the underlying capacity predates domestication or emerged during dogs’ long association with humans.
Your dog might know what follows the word “walk.” The surprising finding here is that even when the sounds carry no meaning, its brain can follow a pattern in the uninterrupted flow of speech.
The research has been published in Science.
This article was fact-checked by Fiona MacDonald and edited by Fiona MacDonald. While we pride ourselves on our process, we are only human. If you spot a mistake, please let us know.
