Baby chicks associate ‘bouba’ and ‘kiki’ with round and spiky shapes, like humans

Researchers at the University of Padova report in Science (covered by NPR on Feb. 19, 2026) that newly hatched domestic chicks link the nonsense sounds “bouba” and “kiki” with rounded and spiky shapes, respectively. In controlled trials, three-day-old chicks trained to seek a combined blob-and-spike panel then preferentially approached the rounded figure when they heard “bouba” and the jagged figure when they heard “kiki.” One-day-old chicks shown moving images on screens produced the same pattern: “kiki” drew them to pointy motion, “bouba” to rounded motion. The finding mirrors the long-studied bouba-kiki effect in human infants and suggests conserved sensory associations across vertebrates.

Key Takeaways

• Study subjects: newly hatched domestic chicks tested at 1 day and 3 days old by researchers Maria Loconsole, Silvia Benavides-Varela and Lucia Regolin at the University of Padova (Italy).
• Primary result: chicks showed a statistically significant preference to approach rounded shapes when hearing “bouba” and spiky shapes when hearing “kiki.”
• Training protocol: three-day-old chicks were trained with a combined blob/spike panel paired with food; they then chose between separate rounded and spiky panels during sound playback.
• Younger test: one-day-old chicks responded similarly in a screen-based moving-object test that measures approach tendency to animate stimuli.
• Comparative context: prior experiments found the bouba-kiki effect in human infants (from about four months) but not in great apes such as chimpanzees and gorillas.
• Interpretation: authors and outside linguists propose the result reflects primitive cross-modal sensory mappings that predate human language.
• Publication and coverage: the experiments are reported in the journal Science and summarized in an NPR report on Feb. 19, 2026.

Background

The bouba-kiki effect has been documented for decades: across many cultures, people tend to label a rounded, blob-like figure with the soft-sounding nonsense word “bouba” and a jagged, angular figure with the sharp-sounding “kiki.” Psychologists have observed the pattern in adults and in human infants as young as four months, which has prompted debate about whether the mapping arises from shared perception or learned experience. Some researchers have argued these cross-modal correspondences might have offered a perceptual scaffold for early vocal symbol formation.

Attempts to detect bouba-kiki-like mappings in animals that are close relatives of humans — notably studies with chimpanzees and gorillas — have not produced clear evidence of the effect. That mixed comparative record motivated the Padova team to test a very different branch of vertebrates and to examine animals at the earliest post-hatching ages. By using chicks at one and three days old, the researchers aimed to minimize the influence of individual learning and social experience, strengthening the case that any observed mapping could be innate or developmentally early-emerging.

Main Event

In the three-day-old cohort, experimenters first trained chicks to go behind a panel decorated with a composite figure that included both rounded and spiky elements; successful location behind that panel produced a food reward. After training, each chick faced a two-panel choice: one panel showing a rounded, blob-like motif and the other showing a spiky, angular motif. During the decision window, the team played repeated recordings of either the syllable “bouba” or the syllable “kiki.” The chicks’ choices shifted in line with the sound: “bouba” trials produced more approaches to the rounded panel, while “kiki” trials produced more approaches to the spiky panel.

To test even earlier predispositions, the researchers ran one-day-old chicks in a screen-based task. Two video displays presented moving shapes — one rounded, one spiky — that normally elicit approach behavior in newly hatched chicks. When the recorded sounds were played simultaneously with the displays, chicks approached the screen showing the shape congruent with the sound: pointy motion during “kiki” playback, rounded motion during “bouba.” The pattern was consistent across multiple subjects and replications in the report.

The authors caution that the experiments do not imply chicks possess language-like semantics. Maria Loconsole emphasized that the results show basic cross-modal linkages between auditory and visual features that may have later been co-opted in humans for symbolic mapping. Outside linguist Marcus Perlman of the University of Birmingham noted surprise at the finding but argued it supports the idea that vertebrate sensory systems are biased to expect certain regularities — a low-level constraint that language may have exploited.

Analysis & Implications

The Padova experiments strengthen an argument that some sound-shape correspondences are not exclusively human cultural artifacts but reflect more general perceptual tendencies in vertebrates. If chicks display these mappings within days of hatching, the associations are unlikely to be the result of individual learning from language exposure. That developmental timing supports hypotheses that cross-modal correspondences can be rooted in conserved neural or sensory mechanisms, such as statistical correlations between acoustic properties (e.g., abrupt high frequencies) and visual texture or contour.

Evolutionary interpretation, however, requires caution. Demonstrating a predisposition to match sounds and shapes does not equate to a capacity for symbolic reference or combinatorial syntax; those latter faculties remain distinctive to humans. The chicks’ responses indicate a tight sensory coupling that could facilitate initial labeling, but language-building requires additional cognitive capacities — social learning, abstraction, and intentional signaling — that the current experiments do not address.

From a comparative perspective, the result reframes earlier negative findings in great apes: absence of evidence in apes does not mean vertebrate-level absence of cross-modal mapping. Instead, species differences in sensory ecology, developmental windows, and experimental method may explain divergent outcomes. The chick result prompts fresh cross-species testing with standardized protocols to map where in the vertebrate tree such mappings appear and how robust they are across modalities and contexts.

Comparison & Data

The table summarizes published and reported outcomes: human infants and neonatal chicks show congruent sound-shape mappings, while studies with great apes have not found consistent evidence. Differences in methodology — e.g., forced-choice labeling versus spontaneous approach — may shape detectability. Future work should standardize stimuli, control for motivational factors (food reward, social cues), and examine neural correlates to test if homologous mechanisms underlie the behaviors.

Reactions & Quotes

Several researchers and commentators highlighted the study’s implications while urging restraint about claims on language origins. Below are brief excerpts placed in context.

Loconsole framed the study as a window into early life predispositions and explicitly warned against overinterpreting the results as evidence of avian language:

“I’m not saying that chicks have human language.”

Maria Loconsole, University of Padova (researcher)

This remark followed the team’s description of the experiments and was intended to emphasize that the observed mappings are basic perceptual correlations rather than symbolic communication equivalent to human language. The authors argue these linkages could represent sensory regularities that human communication later exploited.

“I was surprised by it,”

Marcus Perlman, University of Birmingham (linguist)

Perlman, who was not part of the study, noted his surprise given prior negative results in great apes. He interpreted the finding as evidence that vertebrate sensory systems are predisposed to expect certain regularities in the environment, a bias that might have provided raw material for the cultural evolution of speech.

“So that’s interesting. It shows that the vertebrate’s sensory systems are primed to expect certain regularities in the world.”

Marcus Perlman, University of Birmingham (linguist)

Observers stressed follow-up work: replication across breeds, cross-modal neural measurements, and tests with other taxa to map how widespread the mapping is and whether it arises from shared ancestry or convergent sensory constraints.

Unconfirmed

• The claim that the bouba-kiki mapping is universally innate across all vertebrates remains unconfirmed; wider taxonomic sampling is needed.
• Whether the observed chick mappings reflect homologous neural mechanisms shared with mammals or convergent perceptual processing is not established.
• Any direct causal link between these mappings and the historical emergence of human spoken symbols is speculative and unproven.

Bottom Line

The Padova chick experiments add a compelling data point to the literature on sound-shape correspondences: newborn chicks, like human infants, prefer to associate “bouba” with rounded forms and “kiki” with spiky forms. Because these responses appear within days of hatching, they strengthen the argument that at least some cross-modal mappings arise very early in development and may reflect conserved perceptual biases rather than learned, language-specific conventions.

However, the presence of sensory linkages is not the same as possession of language. The study reframes questions about the perceptual prerequisites for symbolic communication and invites broader comparative and neural work. Replication across species, standardized methods, and mechanistic studies will be necessary to determine how these predispositions contributed — if at all — to the cultural evolution of human language.

Sources

• NPR — news coverage summarizing the Padova team’s findings (news media).
• Science — peer-reviewed journal reporting the original experiments (scientific journal).
• University of Padova — research institution where the experiments were conducted (academic institution).