Dogs rely on consonants to identify words in continuous human speech

A shared auditory mechanism

Human speech flows continuously without distinct physical pauses between individual words, requiring listeners to parse acoustic streams into discrete units. While vowels carry higher acoustic energy and volume, human languages rely predominantly on consonants to establish word identity, as changing a consonant routinely alters core meaning. Research published on 24 September 2026 in the journal Science indicates that domestic dogs share this cognitive bias. Neuroscientists at the Department of Ethology at Eötvös Loránd University in Budapest discovered that pet dogs preferentially track consonant structures when segmenting continuous spoken sounds, a strategy previously documented only in humans.

Measuring brain activity through electroencephalography

To test speech segmentation strategies across species, researchers recruited 20 adult humans and 20 companion dogs between 1 and 14 years of age. The canine cohort comprised 11 crossbreeds alongside purebreds such as Border Collies and Siberian Huskies, with male and female subjects living in family homes. Scientists recorded real-time neural activity using non-invasive electroencephalography while subjects listened to artificial speech streams made of three-syllable nonsense words without pauses. The audio featured three distinct structural conditions: streams where consonants repeated in fixed patterns while vowels varied, streams where vowels followed fixed patterns while consonants varied, and control streams with entirely random sound distributions.

Co-author Dorottya Rácz described the neural alignment between both groups during the artificial speech tests.

We found surprising similarities in the way both species processed these speech sequences. Like humans, the brains of dogs segmented the continuous flow of speech into words better when the recurring pattern, that is, the backbone of the word, was formed by consonants.

Brain responses to consonant patterns

Neural recordings revealed that both dogs and humans generated stronger, synchronized brain responses when three-syllable words were organized by recurring consonant frameworks, such as combinations like BaNeK, BiNoK, and BuNaK. Although vowels delivered greater acoustic energy, subjects in both species prioritized the underlying consonant structure to detect word boundaries. Human infants initially rely on perceptually prominent vowels during early infancy before transitioning toward a consonant-centric bias around one year of age. The canine data demonstrates that this perceptual preference does not depend on complex linguistic capacity or the ability to produce speech.

Lead author Attila Andics outlined the structural distinction between vowels and consonants in verbal communication.

Although vowels are louder and more noticeable, consonants usually form the backbone of words.

Exposure to daily human speech

Because companion dogs share domestic environments with people and encounter spoken communication daily from early life, they provide a model for evaluating how auditory exposure shapes cognition. The findings suggest that recurring immersion in language allows the auditory cortex to extract statistical regularities without requiring formal grammar or vocabulary comprehension. Juan Manuel Toro, a professor at the Center for Brain and Cognition at Universitat Pompeu Fabra, addressed how continuous exposure influences canine speech processing.

And that is where this study is interesting, because it shows that this bias is not uniquely human. Domestic dogs are exposed to human language from birth and this study demonstrates that this experience is sufficient for them to also focus on consonants to identify words.