Why Your Dog Instinctively Matches Spoken Human Words With Geometric Shapes

A study published on October 14, 2026, in Proceedings of the Royal Society B: Biological Sciences revealed that domestic dogs spontaneously match spoken human syllables to visual geometric shapes.

When presented with abstract nonsense words, dogs look longer at rounded, undulating shapes when hearing the soft vocalization "Bouba". When hearing the sharp, clipped sound "Kiki," their gaze shifts to jagged, spiky polygons.

The discovery, led by animal behaviorist Eleonora Biffi and colleagues at the University of Parma in Italy, marks the first time that the "Bouba-Kiki effect"—a psychological phenomenon documented in humans for nearly a century—has been verified in domestic canines.

For generations, linguists regarded the matching of speech sounds to geometric contours as a uniquely human cognitive quirk, an evolutionary stepping stone that enabled our species to bootstrap arbitrary vocal utterances into meaningful words. The Parma findings overturn that anthropocentric assumption. By demonstrating that companion animals map phonetic acoustics to visual forms without prior training, the research reframes the perennial debate over how non-human animals process human vocal cues.

Tracing how this discovery unfolded reveals a decades-long scientific escalation: from early Gestalt experiments in the 1920s and disputes over animal intelligence to trials in Italian animal behavior laboratories.

1929: The Canary Islands Experiment and the Doctrine of Arbitrary Signs

The foundation of the Parma discovery began nearly a century ago on the island of Tenerife. In 1929, German psychologist Wolfgang Köhler set out to test how human minds extract meaning from novel visual and auditory patterns. Working at the Anthropoid Research Station on the Spanish island, Köhler drew two shapes on separate cards: one was an undulating, amoeba-like blob with smooth curves; the other was an explosive, sharp-edged polygon with needle-thin points.

Köhler presented these drawings to Spanish-speaking participants and introduced two made-up words: maluma and takete. He asked a simple question: which shape is maluma, and which shape is takete?

Between 95 and 98 percent of participants matched maluma with the rounded blob and takete with the spiky starburst.

At the time, the result troubled linguistic orthodoxy. The dominant theory of the era, established by Swiss linguist Ferdinand de Saussure, held that human language is fundamentally arbitrary. Saussure argued that the sound of a word bears no intrinsic link to the physical properties of the object it represents. The English word "dog," the French "chien," and the German "Hund" share no acoustic qualities, yet all designate the same quadruped.

Köhler’s maluma-takete test offered evidence of the opposite: sound symbolism, also called phonesthesia or iconicity. Some speech sounds carried an innate, intuitive physical texture.

Yet across the 1930s and 1940s, canine cognition research ran on an opposing track. Ivan Pavlov’s landmark work on classical conditioning had convinced the scientific establishment that animals possessed no abstract cognitive representations. In Pavlov’s laboratory in Saint Petersburg, dogs salivated to metronomes, buzzers, and bells not because the acoustic frequencies meant anything inherent to the canine nervous system, but because sheer mechanical repetition had wired a reflex.

For the next seventy years, animal behaviorists operated under a rigid consensus: human speech directed at domestic animals was merely acoustic wallpaper—an inventory of arbitrary conditioned triggers where pitch, volume, and repetition elicited trained motor actions, completely detached from internal cross-sensory concepts.

2001: The Neurological Resurgence of Bouba and Kiki

The phenomenon re-entered modern neuroscience in 2001, when neuroscientists Vilayanur S. Ramachandran and Edward Hubbard of the University of California, San Diego, revisited Köhler’s paradigm.

Ramachandran and Hubbard renamed the test words "Bouba" and "Kiki" to isolate distinct phonetic properties. "Bouba" paired voiced bilabial stops (/b/) with back rounded vowels (/u/), a combination that produces a smooth, continuous acoustic envelope and requires the speaker to round their lips into a circle. "Kiki" deployed voiceless velar stops (/k/) and high front unrounded vowels (/i/), creating sudden, abrasive acoustic transients and requiring a taut, angular mouth position.

Testing both American undergraduates and native Tamil speakers in India, Ramachandran and Hubbard found that 95 percent of subjects assigned "Bouba" to the curvy shape and "Kiki" to the jagged one.

Ramachandran and Hubbard put forward a neurological explanation: the human brain performs cross-modal abstraction, translating sensory data across distinct cortical regions. They hypothesized that this cross-wiring occurred in the angular gyrus, a parietal lobe structure positioned at the junction of the auditory, visual, and somatosensory cortices.

The angular gyrus, they argued, acted as a neural clearinghouse. It recognized a mathematical property common to both sensory streams: the sharp, abrupt acoustic transients of /k/ shared a rate of change with the sharp visual vectors of the spiky shape, while the smooth, low-frequency acoustic transitions of /b/ mirrored the gradual curvature of the blob.

Ramachandran and Hubbard pushed their thesis further: they proposed that this cross-modal sensory mapping was the evolutionary spark of human language. Before syntax, hominids used bodily and acoustic gestures whose physical shapes mirrored the world around them.

Because the angular gyrus and high-level language architectures were assumed to be uniquely developed in primates, particularly humans, the Bouba-Kiki effect was cataloged as a human evolutionary trait.

Non-human animals, researchers assumed, lacked the neural machinery to translate a phoneme into a geometric concept.

2004–2012: Canine Prodigies and the Mystery of Object Classification

At the turn of the century, researchers began re-evaluating canine intelligence. In 2004, Julia Fischer and her team at the Max Planck Institute for Evolutionary Anthropology published a paper in Science examining Rico, a Border Collie who had learned the names of more than 200 items. Rico could retrieve specific toys upon verbal command and demonstrated "fast-mapping"—the ability to deduce the referent of a new word through the process of elimination, a skill previously documented only in human toddlers.

In 2011, psychologist John Pilley of Wofford College expanded those boundaries with Chaser, another Border Collie. Over three years of intensive daily training, Chaser learned the names of 1,022 distinct objects, categorized them into functional groups like "frisbees" and "balls," and recognized syntax-like combinations of verbs and nouns.

These canine prodigies forced a fundamental question: when dogs hear human speech, what are they picturing?

The question of whether can dogs understand human language has long divided ethologists and linguists, swinging between claims of literal comprehension and dismissals of conditioned auditory reflexes.

To determine how dogs categorize words, University of Lincoln researchers Emile van der Zee, Helen Zulch, and Daniel Mills conducted a study in 2012 published in PLOS ONE. They tested Gable, a five-year-old Border Collie, to examine if dogs share the human child's "shape bias".

When human children between ages two and three learn a novel word for an object—such as "cup"—they generalize that word to other objects of the same shape, ignoring differences in color, size, or material texture. If you show a toddler a blue plastic cup and call it a "dax," and then offer them a choice between a red ceramic object of the identical cup shape and a blue plastic plate, the child will call the cup-shaped object "dax". Visual geometry serves as the cornerstone of human semantic vocabulary.

Van der Zee’s team presented Gable with novel objects differing systematically across size, texture, and shape.

The results exposed a split between human and canine cognition. After initial training, Gable generalized words based on an object's size, not its geometry. After longer exposure, he grouped objects based on surface texture—favoring how an object felt in his mouth or against his paws. Gable showed no instinctive shape bias.

The Lincoln researchers concluded that the canine cognitive blueprint prioritized tactile feedback and mass over visual contours. If dogs did not rely on geometric outlines to organize the vocabulary of their everyday world, the scientific community saw little reason to think a dog would ever instinctively match the sound of a human phoneme with an abstract visual silhouette.

2011–2022: Bioacoustics and the Non-Human Primate Dead End

While cognitive psychologists probed dog vocabulary, bioacousticians took a different path: exploring cross-modal perception through animal vocalizations.

At the University of Sussex, researchers David Reby, Karen McComb, and their colleagues began testing how mammals integrate acoustic signals with physical anatomy. In 2011, Reby’s lab published an experiment in PLOS ONE using resynthesized dog growls.

By altering the formant dispersion of canine growls—an acoustic metric that correlates directly with the length of the vocal tract and therefore the physical size of the caller—the researchers tested pet dogs using a cross-modal preferential looking paradigm. Dogs stood in front of two life-sized models: a large dog and a small dog.

When speakers played a growl calibrated to a large animal's vocal tract, the subjects looked significantly sooner and longer at the large model. When the acoustic profile shifted to formants characteristic of a miniature dog, their gaze swung to the small model.

The Sussex experiments demonstrated that dogs possess cross-modal mental maps. They did not perceive growls as isolated audio; they mapped acoustic frequency profiles to real-world visual dimensions.

Subsequent work led by Holly Root-Gutteridge and Anna Korzeniowska in 2022 confirmed that dogs display an abstract pitch-size correspondence. They matched arbitrary low-pitched mechanical tones to large physical blocks and high-pitched tones to small blocks, even when the objects had no biological relevance.

                     CROSS-MODAL SENSORY DISCOVERY TIMELINE
═══════════════════════════════════════════════════════════════════════════════════════
 1929  Köhler's Tenerife Trials: Humans spontaneously match "Maluma" to curves
       and "Takete" to sharp points. Structuralist linguistic dogma challenged.
───────────────────────────────────────────────────────────────────────────────
 2001  Ramachandran & Hubbard: "Bouba-Kiki" revival. Cross-modal cortical mapping
       identified; proposed as a human-exclusive language evolution catalyst.
───────────────────────────────────────────────────────────────────────────────
 2011  Sussex Bioacoustics: Dogs match resynthesized growl formant frequencies
       to visual physical body sizes of large versus small canine models.
───────────────────────────────────────────────────────────────────────────────
 2012  Lincoln Shape-Bias Failure: Border Collie "Gable" classifies named objects
       by size and texture, failing to show the human infant's visual shape bias.
───────────────────────────────────────────────────────────────────────────────
 2022  Primate Impasse: Kanzi the bonobo fails the Bouba-Kiki task. Primatology
       concludes sound symbolism requires uniquely human vocal-linguistic traits.
───────────────────────────────────────────────────────────────────────────────
 2023  The Avian Anomaly: 3-day-old domestic chicks exhibit spontaneous Bouba-Kiki
       associations, pointing to ancient vertebrate sensory cross-talk.
───────────────────────────────────────────────────────────────────────────────
 SEPT  ELTE Budapest (Science): EEG demonstrates dog brains track consonant rhythms
 2026  in continuous speech, revealing human-like phonetic stream parsing.
───────────────────────────────────────────────────────────────────────────────
 OCT   University of Parma (Royal Society B): 70 dogs match "Bouba" to round
 2026  and "Kiki" to spiky shapes; results hold under double-blind controls.
═══════════════════════════════════════════════════════════════════════════════════════

Despite these canine findings, primatologists remained convinced that the Bouba-Kiki effect itself belonged exclusively to humans.

In early 2022, a research team led by Vera Schietecat and colleagues published an investigation titled "Bo-NO-bouba-kiki: Picture-word mapping but no spontaneous sound symbolic speech-shape mapping in a language trained bonobo" in Proceedings of the Royal Society B. They tested Kanzi, the world’s most linguistically competent bonobo (Pan paniscus), who had spent decades communicating via a keyboard of lexigrams.

Kanzi completed thousands of trials matching auditory English words to visual lexigrams. Yet when presented with the Bouba-Kiki challenge, the bonobo performed at chance. He showed no spontaneous preference linking "Bouba" to rounded contours or "Kiki" to pointed edges.

The scientific consensus held firm: non-human primates lacked the spontaneous audio-visual sound symbolism that underpinned human speech. If humanity's closest living evolutionary relative could not detect geometry in phonemes, the ability was presumed impossible for a domesticated carnivore.

2023–2024: The Avian Curveball Upends the Evolutionary Timeline

The hypothesis that sound symbolism was a specialized human linguistic adaptation ran into a complication in 2023, arriving not from apes or dogs, but from an animal behavior laboratory at the University of Padua in Italy.

Researchers Maria Loconsole, Cinzia Chiandetti, and Giorgio Vallortigara tested three-day-old domestic chicks (Gallus gallus domesticus). Because the chicks were tested mere hours after hatching, they had no prior exposure to human language, no cultural training, and minimal visual experience.

The chicks were placed in an experimental arena where they could choose to approach either rounded or spiky cutouts while speakers played "Bouba" or "Kiki" sounds.

The chicks spontaneously associated the auditory stimuli with the shapes. They approached rounded shapes when exposed to "Bouba" and favored spiky shapes when hearing "Kiki".

The Padua chick study dismantled the theory that sound-shape symbolism depended on human language or an evolved human vocal tract. A bird possessing a tiny, lissencephalic brain lacking a mammalian neocortex could decode the geometry of human phonemes.

The biological mechanisms supporting cross-modal correspondences were not late-stage primate adaptations. They were ancient vertebrate sensory systems, likely dating back hundreds of millions of years.

The revelation cast the canine lineage in a new light. If birds possessed cross-modal sensory bridges, why had dogs failed early shape-bias tests? Had previous dog experiments tested the wrong variables, or had scientists miscalculated how canine sensory systems interact with human vocalizations?

September 2026: The Cortical Bridge in Budapest

In late September 2026, the canine cognition puzzle gained its missing neural link. A research group led by Attila Andics at the Department of Ethology at Eötvös Loránd University (ELTE) in Budapest published a study in Science.

Andics and his team had spent a decade using non-invasive functional magnetic resonance imaging (fMRI) to monitor the brains of awake, unrestrained family dogs. In earlier work, they had established that canine brains process human speech hierarchically: the left hemisphere processes lexical meaning, while the right hemisphere processes acoustic intonation, closely mirroring the lateralization seen in the human brain.

Their September 2026 study tackled a more granular puzzle: how dogs parse the stream of human speech into individual components.

Using non-invasive scalp electroencephalography (EEG), the Budapest researchers exposed 20 pet dogs and 20 humans to artificial speech streams that contained no physical pauses between words. Some speech streams contained regular, repeating consonant sequences; others contained repeating vowel patterns; and a control stream was randomized.

The EEG readouts revealed that dog brains synchronized their neural oscillations to consonant-based rhythms, matching the neural tracking displayed by the human subjects.

The Budapest findings showed that dogs are not tone-deaf to human phonetics. Rather than listening solely to the emotional prosody or pitch of a human voice, the canine auditory cortex tracks the acoustic boundaries created by consonants.

When asking whether can dogs understand human language, scientists must distinguish between lexical vocabulary and lower-level acoustic processing. Dogs might not master syntactic grammar, but the ELTE study proved their nervous systems track the phonetic components of human speech.

The stage was set. Canine auditory systems parsed phonemes; bioacoustics showed they processed cross-modal sound and size; and avian studies proved sound-shape links were biologically ancient.

All that remained was to test domestic dogs in a Bouba-Kiki trial.

October 2026: The Parma Discovery

The study published in Proceedings of the Royal Society B on October 14, 2026, provided that missing evidence.

Headed by doctoral researcher Eleonora Biffi, co-first author Giulia Pedretti, and senior ethology professor Paola Valsecchi at the University of Parma—working alongside Simona Normando, Maria Loconsole, and computational specialist Teddy Lazebnik—the team designed an experiment to test whether domestic canines map speech phonemes directly to visual geometry.

┌─────────────────────────────────────────────────────────────────────────────┐
│                 UNIVERSITY OF PARMA TESTING APPARATUS                       │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                             │
│      [ Left Screen ]                                   [ Right Screen ]     │
│    ┌─────────────────┐                               ┌─────────────────┐    │
│    │  *  /\   *  /\  │                               │    (       )    │    │
│    │ / \/  \ / \/  \ │                               │   (  BLOB   )   │    │
│    │ > SPIKY SHAPE < │                               │    (       )    │    │
│    │ \_/\  /_/\  /_/ │                               │     (_____)     │    │
│    └─────────────────┘                               └─────────────────┘    │
│      [ Audio Spkr L ]                                  [ Audio Spkr R ]     │
│             \                                                 /             │
│              \                                               /              │
│               \                                             /               │
│                \                 Center                    /                │
│                 \                  │                      /                 │
│                  ─────────► [ Dog + Owner ] ◄─────────────                  │
│                                    │                                        │
│                           [ 4K GoPro Camera ]                               │
│                                    │                                        │
│                        [ AI Gaze-Tracking Engine ]                          │
│                                                                             │
│  Phase 1 (0-20 sec): Leashed fixation; AI head-orientation vector tracking  │
│  Phase 2 (20-30 sec): Unleashed physical exploration of preferred display   │
└─────────────────────────────────────────────────────────────────────────────┘

The Parma researchers recruited 70 adult pet dogs of various breeds and mixed backgrounds. The study was split into two stages, each following a within-subject design:

Study 1 (51 Dog-Owner Pairs): Each dog entered a testing arena featuring two high-definition displays flanking a central observation point. One screen displayed a black, multi-pointed, jagged starburst; the other displayed an undulating, curved, amoebic blob. Speakers behind the displays played recordings of a female Polish phonetician pronouncing either "Bouba" or "Kiki". The vocalizations were repeated 15 times over a 30-second trial, alternating once every two seconds. The shapes swapped sides across four trials to prevent spatial bias.
Study 2 (19 Dog-Owner Pairs): The team added a control to eliminate the Clever Hans effect—the risk of owners unintentionally tipping off their pets through micro-movements, eye shifts, or muscle tension. In this cohort, owners wore sound-isolating headphones blasting rock music throughout the trial, blinding them to which audio stimulus their dog was hearing.

During the first 20 seconds of each trial, owners held their dogs facing the central midpoint between the displays. A calibrated 4K camera filmed the dogs at 30 frames per second.

To prevent human observer bias, the video footage was analyzed by an artificial intelligence model that calculated each dog’s binocular gaze vector and head-orientation angle in real time.

For the final 10 seconds of each trial, owners released their dogs to physically approach and inspect the screen of their choice.

The results showed a clear statistical pattern:

┌─────────────────────────────────────────────────────────────────────────────┐
│                       EXPERIMENTAL LOOKING PREFERENCE                       │
├─────────────────────────────────────────────────────────────────────────────┤
│  Auditory Cue: "BOUBA" (Voiced Bilabial Stop + Back Rounded Vowel)          │
│  ├── Gaze toward Rounded Blob:  56.0% looking duration (p < 0.01)           │
│  └── Gaze toward Spiky Polygon: 44.0% looking duration                      │
│                                                                             │
│  Auditory Cue: "KIKI" (Voiceless Velar Stop + High Front Vowel)             │
│  ├── Gaze toward Spiky Polygon: 63.0% looking duration (p < 0.001)          │
│  └── Gaze toward Rounded Blob:  37.0% looking duration                      │
│                                                                             │
│  Study 2 Double-Blind Control (Owners wearing audio-masking headphones):    │
│  └── Sound-shape concordance maintained; Clever Hans effect excluded.       │
└─────────────────────────────────────────────────────────────────────────────┘

When "Bouba" played through the speakers, the dogs looked significantly longer at the rounded, soft-edged shape.

When "Kiki" cut through the room, their attention swung toward the jagged, spiky polygon.

During the unleashed exploration window, dogs directed their physical approaches toward the matching visual shape.

In Study 1, dogs showed an overall baseline interest in the spiky shape, likely drawn by the higher visual contrast of sharp angles. But when Study 2 muted the owners' hearing with rock music, that general bias flattened out, leaving the sound-shape correspondence intact.

"Dogs looked significantly more at the rounded shape while 'Bouba' was played, and they looked at and explored the spiky shape when 'Kiki' was played," Biffi and her co-authors wrote in their report.

The team confirmed that domestic dogs spontaneously connect the acoustic properties of spoken human syllables to abstract geometric contours.

The Biophysical Code: Why Stop Consonants Are Sharp and Vowels Are Curved

How does a dog’s nervous system translate sound waves into geometric angles? The answer lies in the physics of speech acoustics and how mammalian sensory systems process incoming stimuli.

                 ACOUSTIC-TO-VISUAL MAPPING MECHANISMS
═══════════════════════════════════════════════════════════════════════════════
 Acoustic Stimulus: "KIKI"              Acoustic Stimulus: "BOUBA"
 Phonemes: /k/ + /i/                    Phonemes: /b/ + /u/
───────────────────────────────────────────────────────────────────────────────
 Physical Acoustic Characteristics:    Physical Acoustic Characteristics:
 • High-frequency transients           • Low-frequency spectral energy
 • Sudden, abrupt attack envelope       • Smooth, continuous acoustic envelope
 • Rapid spectral transitions          • Gradual rise and decay times
───────────────────────────────────────────────────────────────────────────────
 Neural Sensory Translation:            Neural Sensory Translation:
 Sudden energy spikes activate neurons  Sustained acoustic energy activates 
 sensitive to abrupt temporal change.  cortical maps sensitive to continuity.
───────────────────────────────────────────────────────────────────────────────
 Cross-Modal Visual Representation:     Cross-Modal Visual Representation:
 • Jagged, pointed contours             • Smooth, undulating contours
 • High-frequency spatial vectors       • Low-frequency spatial curvature
 • Abrupt directional changes           • Continuous, gradual edges
═══════════════════════════════════════════════════════════════════════════════

When human vocal tracts produce the word "Kiki," two acoustic mechanisms occur:

Voiceless Velar Stops (/k/): The tongue completely seals the airflow against the soft palate before releasing it suddenly. This creates an abrupt silence followed by a high-energy acoustic burst known as a transient. In signal-processing terms, this transient is a sharp spike in the acoustic waveform.
High Front Vowel (/i/): The resonant frequencies of the vocal tract produce high second and third formants (often exceeding 2,500 to 3,000 Hz). High frequencies are characterized by rapid, tightly compressed sound waves.

When a dog’s middle and inner ear processes this signal, the auditory nerve fires in sharp, high-frequency bursts. In the visual cortex, sharp-angled geometric figures possess high spatial frequencies—abrupt, high-contrast shifts in luminance and directional vectors. The canine brain connects these inputs because they share an underlying mathematical rate of change.

By contrast, the word "Bouba" produces an acoustic profile built on low-frequency formants and smooth energy transitions:

Voiced Bilabial Stops (/b/): The lips close softly, allowing the vocal cords to vibrate throughout the consonant's production. There is no sharp, empty silence followed by an explosive acoustic transient; the sound envelope rises and falls in a continuous, gradual swell.
Back Rounded Vowel (/u/): Lip rounding elongates the vocal tract, dropping the first and second formants into low-frequency territory (often below 800 Hz). The sound waves are long, wide, and smooth.

To the canine nervous system, "Bouba" provides an acoustic signal free of sudden temporal shifts. In visual processing, this continuous acoustic flow maps onto the gradual, uninterrupted contours of a curved shape, which lacks high-frequency spatial vector changes.

Rather than asking simply whether can dogs understand human language, researchers must now investigate how sensory cortices translate acoustic contours into non-auditory properties like mass, texture, and shape. The canine brain is not merely listening to arbitrary labels; it is running real-time cross-sensory translations across sight and sound.

Domestication Adaptation or Vertebrate Inheritance?

The Parma discovery leaves behavioral biologists facing a clear evolutionary question: why does this mapping exist in a dog?

Two competing schools of thought have emerged to explain the findings:

1. The Domestication Hypothesis

Under this model, the canine Bouba-Kiki effect is an evolutionary byproduct of 30,000 years of domestication.

Over thousands of generations living alongside human families, dogs were intensely selected for their ability to read human emotional states, facial expressions, pointing gestures, and vocal nuances. Wolves do not track human gazes or gaze back into human eyes to elicit oxytocin release; domestic dogs do.

Proponents of this view argue that dogs evolved a hyper-tuned auditory sensitivity to the acoustic subtleties of human speech. By living in human households, domestic canines may internalize the structural and sound-symbolic associations of human language through passive everyday immersion.

2. The Ancient Biophysical Architecture Hypothesis

The opposing view—supported by the 2023 chick discoveries—holds that domestication has little to do with it. Instead, cross-modal correspondence is an ancient survival adaptation common to all terrestrial vertebrates.

In the natural world, physical properties are bound by acoustics:

Small, sharp, brittle things—like snapping twigs, splintering bones, or needle-sharp claws scraping stone—produce high-frequency, transient acoustic sounds.
Large, heavy, rounded things—like rolling boulders, large predators, or heavy body masses shifting against earth—produce low-frequency, continuous, resonant acoustics.

An animal that maps high-frequency, abrupt acoustic transients to sharp, angular physical structures gains an immediate survival advantage. It does not need to learn that a hiss or a snap warns of a pointed threat; its sensory system binds those sensory dimensions together automatically.

"If animals without human-like language also spontaneously associate sounds with shapes, this suggests that at least some of the mechanisms underlying sound symbolism may not be specific to language," Biffi noted following the study’s release. "Instead, they may reflect more general principles of how the brain integrates information from different senses".

To resolve this debate, Biffi’s research team has expanded its experimental program. The Parma laboratory is currently gathering comparative data from three distinct cohorts:

Pre-Weaned Puppies: Testing domestic puppies before they have experienced substantial human vocal communication, isolating innate perceptual hardware from learned home experience.
Wild Canids: Testing hand-reared red foxes (Vulpes vulpes) and captive gray wolves (Canis lupus) to isolate the specific genetic impacts of domestication.
Domestic Cats (Felis catus): Testing domestic felines, which share the human home environment but possess an evolutionary history and domestication trajectory entirely different from canines.

"If cats also show the Bouba–Kiki effect, this would provide further evidence that sound symbolism may reflect general perceptual mechanisms, rather than adaptations specifically related to dogs' long history of cooperation and communication with humans," Biffi explained.

Implications for Dog Training and Working Breeds

While evolutionary biologists debate ancestral sensory maps, the realization that dogs possess an intuitive sense of sound symbolism carries immediate consequences for professional trainers, working handlers, and companion animal owners.

For over a century, traditional dog training dogma maintained that verbal commands are entirely arbitrary acoustic tags. Trainers were taught that a dog could just as easily learn to lie down on the word "banana" as the word "down," provided the reward schedule was consistent.

The Parma discovery reveals that training commands do not land on a blank acoustic slate. A dog’s nervous system exhibits innate biases toward the geometry and physical texture of vocal sounds.

Commands with sharp, voiceless consonants and abrupt, high-frequency transients—such as "Sit!", "Tuk!", "Hup!", or "Quick!"—share acoustic profiles with the word "Kiki." These sounds naturally evoke sharp, sudden physical adjustments, making them well-suited for fast directional changes, emergency stops, or precise postural shifts.

Conversely, cues built around long, low-frequency, voiced vowels and continuous envelopes—such as "Whoa," "Slow," "Settle," or "Gooood"—share acoustic characteristics with "Bouba." These sounds map to gradual curves and fluid postures, making them effective for encouraging steady movement, soothing physiological arousal, or prompting a dog to relax.

Working handlers are already taking note. In search-and-rescue, military working dog handling, and agility sports, matching the phonetic envelope of a command to the physical geometry of the desired behavior could reduce cognitive latency in high-stress, noisy environments. A dog asked to perform an abrupt emergency stop may respond fractions of a second faster to an acoustic command that matches the sharp physical action required.

The Unresolved Horizons of Canine Cognitive Science

The Parma study has opened a new chapter in comparative cognitive science, leaving several technical questions unresolved for the next wave of animal cognition research:

Breed Variations and Sensory Biases: The 70 dogs in the Parma experiment were analyzed as a collective group. Do sight hounds (such as Greyhounds and Whippets), bred to emphasize visual processing, display a stronger sound-shape correspondence than scent hounds (like Bloodhounds and Beagles), whose brains dedicate massive neural real estate to the olfactory bulb?

The Limits of Interspecies Semantics: While asking if can dogs understand human language in its full grammatical complexity remains contested, Biffi’s research establishes that dogs already inhabit the perceptual foundations from which human vocabulary arose. A dog hearing human speech is not processing sterile acoustic noise; it is navigating a multisensory landscape where sounds possess tangible weight, texture, and geometry.

Cross-Modal Integration Across Modalities: Future investigations will look beyond sight and sound. Does the canine olfactory bulb participate in this cross-modal architecture? Do certain molecular scents carry "spiky" or "rounded" acoustic profiles in the canine mind?

The discovery that dogs see shapes in the syllables we speak demonstrates that the bridge between humans and canines was not built on behavioral conditioning alone. Long before humans invented words to name the physical world, mammalian nervous systems were already translating the rhythms of sound into the architecture of physical form.