Image source, BBC News Mundo/Daniel Arce
Look at the two figures above; the white one on the left, with its pointed shapes and numerous vertices and edges. The fuchsia on the right, rounded and soft.
Now, imagine that one of them is "kiki" and the other is "bouba." They are made-up words, which mean nothing in any language, but which is which?
If you thought that "bouba" is the round one and "kiki" is the pointed one, you are part of an immense majority. And the strange thing is that there is no logical reason for it.
Probably when you named them you didn't think that a "b" is soft or that a "k" is sharp. However, the answer came effortlessly, as happens to millions of people around the world.
This little mental exercise is known today as the bouba/kiki effect, and it has been intriguing psychologists, linguists and neuroscientists for a century, as it could hide a clue to one of the great mysteries of our species: how we start to talk.
Of that, there are only hypotheses. But as for how we started talking about this particular phenomenon, it all started in 1924, thanks to an article by Georgian psychologist Dimitri Uznadze about how we name things.
In his laboratory in Tbilisi, he gathered ten participants, gave them a list of made-up words, showed them six abstract drawings for five seconds each, and asked them to choose the most appropriate name for each figure.
If the choice had been pure chance, the answers would have been distributed without rhyme or reason. It was not like that: the participants agreed more than chance would allow.
End of Most read
Uznadze concluded that when we give a name to something we follow certain regularities of the mind (or of the "human soul," in the psychological vocabulary of the time).
That assertion was a small heresy, since, at that time, linguistics maintained - without a doubt - that the relationship between a word and its meaning was purely arbitrary.
However, the German psychologist Wolfgang Köhler, a key figure in Gestalt psychology (how the mind organizes what it perceives into patterns or structures), echoed Uznadze's finding in his book Gestalt Phsychology (1929).
image copyrightGetty Images
Initially, it was he who simplified the experiment, with two shapes, one rounded and the other angular, and two invented words: takete and baluba, which became maluma in the second edition of the book (1949).
He assured that people assigned the names without hesitation, although he did not give figures; He limited himself to taking the result as obvious.
But although some paid some attention to the experiment, its mention was reduced to a few lines and two illustrations in a foundational work in which other topics called for more reflection.
Observations on takete and maluma were considered a curiosity, and were not properly explored... until half a century later.
The rise to fame
All the news you need is on BBC World Service
Download it here
End of Ad app
In 2001, neuroscientists V. S. Ramachandran and Edward Hubbard of the University of California, San Diego, published their research on synesthesia, a phenomenon in which a sensory stimulus triggers an additional experience, such as seeing colors while listening to music.
Ramachandran suspected that connections between the senses were not exclusive to synesthetic people, but that we all share a certain capacity to establish intersensory associations.
To explore this idea, he and Hubbard revisited the experiment and studied whether people without synesthesia could also spontaneously relate sounds and shapes.
But they used two new words: bouba and kiki.
These are the ones who have named the effect since then, since it was their article Synaesthesia: a window into perception, thought and language, published in the Journal of Consciousness Studies, that popularized it.
They explained that since kiki is pronounced with consonants that cut the air suddenly and with an intense high-pitched vowel, it visually resonates with a sharp edge, while bouba forces you to round your lips and lets the sound flow, like a curve.
For them, the human brain easily connects what it sees, what it hears and how it moves its mouth, a kind of limited synesthesia that we all share.
Their conclusions derived from their discovery: they reported that a surprising 95% of the participants had labeled the figures with the made-up words in the same way.
They noted that to test whether the associations depended on language or culture, they had done the experiment with English-speaking students in the US and Tamil speakers in India. However, they did not detail how many people participated in the experiment or how the responses were collected.
So, rather than offering an exhaustive experimental demonstration, they presented the phenomenon eloquently and convincingly, and managed to make it visible.
The science that came later was in charge of verifying whether this striking association held.
Many times yes
The scientists were testing the effect in other groups.
Daphne Maurer and her collaborators found it in children about two and a half years old, who barely speak. In 2013, Ozge Ozturk and his team detected it even in four-month-old babies, who looked longer at one of the options when the sound and shape "fit."
That same year, Andrew Bremner and his colleagues traveled to northern Namibia to test the Himba, a semi-nomadic people with no written language and very little contact with the West. They showed the same pattern, which ruled out the suspicion that it was all because the "o" in "bouba" is round on the paper.
The most ambitious test came between 2021 and 2022, with a study led by linguist Aleksandra Ćwiek in the journal Philosophical Transactions of the Royal Society B.
The team tested the effect online with 917 people who spoke 25 languages from nine different language families and used ten writing systems, including Arabic, Japanese, Georgian and Cyrillic.
The conclusion was that the effect is robust in all these cultures and does not depend on the alphabet, although the same authors indicated that in some languages, such as Mandarin, Turkish or Romanian, the effect was weak or did not appear clearly.
Other works have documented failures.
When psycholinguist Suzy Styles and linguist Lauren Gawne tested the bouba-kiki effect with speakers of Syuba, a language from the Nepalese Himalayas, they seemed confused by the made-up words, possibly because they didn't sound like any real words in their language.
It was like being presented with the made-up word "ngf" and asking if it's round or pointy. You would probably find it difficult to answer.
Without humans
If the effect were a cultural invention, only humans should notice it. So scientists did something curious: test it on animals.
The first attempts with chimpanzees, gorillas and a bonobo were unsuccessful. But in February 2026, a team from the University of Padua led by Maria Loconsole published something surprising in Science: newborn chicks, one and three days old, approached the pointed figure when "kiki" was played and the rounded one when "bouba" was played.
The researchers highlighted that it is not that chicks speak, but rather that the connections between senses may be much older than language.
In October 2026, a team from the University of Parma, led by Eleonora Biffi, published something similar with dogs in Proceedings of the Royal Society B: 51 pets looked longer at the round figure when hearing "bouba" and the pointed one when hearing "kiki."
There was a second test with 19 other dogs, whose owners wore headphones with music so they couldn't accidentally give them cues, and the pattern held.
The authors urged caution: the sample is small and more research is still needed.
A mere curiosity?
Up to this point, everything sounds like after-dinner fun. However, the bouba/kiki effect could play a role in solving a puzzle we may never fully clarify: how language evolved.
The question is very old. In the Cratylus, one of Plato's dialogues, two friends of Socrates argue about her, and do not reach an agreement. Hermogenes maintains that names are pure convention, an agreement between people. Cratylus defends the opposite: each thing has a correct name by nature.
Over the centuries, others continued to try to understand the origin of language, including the 18th-century philosophers Jean-Jacques Rousseau and Johann Gottfried Herder. In the 19th century, the idea that the first words imitated the sounds of nature became popular.
The philologist Max Müller, in a mocking spirit, baptized it as the "bow-wow theory", and ridiculing it worked better than any refutation.
At that time, linguistics was beginning to consolidate itself as a scientific discipline and, in its desire to establish rigorous limits, it closed the door to this type of discussion. In 1866, the Paris Linguistics Society even included in its statutes that it would not accept any communication on the origin of language.
Half a century later, in the Course in General Linguistics (1916), published after his death, Ferdinand de Saussure elevated the arbitrariness of the sign to a principle: there is no natural reason why a dog is "perro", "dog" or "chien". For much of the 20th century, that was the only respectable possibility.
But little by little, the topic returned to the table, hand in hand with cognitive science, evolutionary anthropology and genetics. An article by Steven Pinker and Paul Bloom, in 1990, proposed that language could be explained as a result of natural selection and reopened the debate.
If all words had been arbitrary from the beginning, how could our ancestors agree to fix the first ones? And how did those who invented them get others to understand their meaning so that the language spread?
The bouba/kiki effect suggests a possibility: perhaps the first words did not start from scratch, but relied on natural connections between sound and shape, size or movement that we all shared. On that scaffolding, language could grow until it became largely arbitrary, like today.
It is a hypothesis, not a fact.
From size to flavor
Once you know how to search, you start to see this phenomenon elsewhere. Already in 1929, the same year as Köhler's book, linguist Edward Sapir did an experiment with two words: "thousand" and "bad."
He asked people to say which one named a big table and which one a small one. Most associated "thousand" with small and "bad" with big. The "i" is pronounced with the mouth almost closed, and the "a" opens it completely.
The Spanish seem to support it. To make something small we use the "i": little house, little bird, little one. To enlarge, we use more open and serious vowels: perrote, hombre, casana. It is not an iron law, but the tendency is there, and it is strong enough to be noticed in grammar.
Nor are the correspondence of sounds limited to shapes.
In 2010, psychologist Julia Simner and her colleagues Christine Cuskley and Simon Kirby at the University of Edinburgh published a study in the journal Perception titled "What does that sound like?"
With different basic flavors, they found that people tend to associate sweet with soft, continuous sounds and sour with staccato, high-pitched sounds.
It seemed to indicate that our brain has a small map that links sound with taste...
...And taste with shape: in 2013 Cadbury rounded the ounces of its Dairy Milk tablet and many consumers protested because it tasted sweeter. The company insisted it had not touched the recipe.
Oxford psychologist Charles Spence published an opinion article in Flavor magazine where he posed a hypothesis: perhaps the rounded shape, without changing a single gram of sugar, was enough to evoke sweetness. Their reasoning was supported by previous studies according to which people associate milk chocolate with round shapes and dark chocolate, which is more bitter, with angular shapes.
But let's go back to the two figures from the beginning.
If the correspondence seemed obvious to you, it's really amazing. There is nothing to explain why "bouba" seems rounded to you.
Perhaps it is the echo of a very ancient connection between what we see, what we hear and what we do with our mouths; perhaps it is the link that allowed the first sounds of our ancestors to begin to mean something.
It is true that most of our words are arbitrary, as Saussure stated. But bouba and kiki seem to indicate that the pact between sound and sense is less capricious than previously believed.
Discover more
Video
End of Video
