A game about language · six experiments
The language in your head
You learned a language before you could tie your shoes, and you have used it every waking hour since. Most of what it does happens below the line of attention. These six experiments, each taken from a published study, bring a little of it to the surface. Play first: the explanation waits for you.
Nothing you choose or type here leaves your device, unless you share your portrait.
Bouba and kiki
One of these shapes is called kiki, the other bouba. Which one is kiki?
What happened
In his book Gestalt Psychology (1929), the German psychologist Wolfgang Köhler drew two shapes and asked readers to match them with two invented words, takete and baluma; later printings changed the second to maluma. In 2001 the neuroscientists Vilayanur Ramachandran and Edward Hubbard renamed it bouba and kiki and reported that about 95% of the people they asked gave the spiky shape the name kiki.
That figure came from students in the United States and India. The widest test so far, by Aleksandra Ćwiek and colleagues in 2022, asked 917 speakers of 25 languages, written in ten scripts. 72% made the usual match; the effect was clear in 17 of the 25 languages. Strong, then, but not universal.
Why
Say both words slowly. For bouba your lips round and the sound flows. For kiki your tongue snaps off the roof of your mouth and the vowel is high and narrow. The mouth seems to draw the shape. Language is meant to be arbitrary: nothing about the word tree looks like a tree. At its edges, though, sound and meaning leak into each other. Linguists call this sound symbolism.
Mil and mal
In an invented language, mil and mal both mean table. One of them is the big table. Which table is the mal?
What happened
In 1929 Edward Sapir, one of the founders of American linguistics, asked people exactly this kind of question. He gave 500 children and adults pairs of invented words that differed only in a vowel. Depending on the pair, about 75% to 96% of them gave the larger meaning to the word with a; mil and mal are his own example. The vowel a needs a wide-open mouth and a large resonating space; i is made with the mouth almost closed.
Is it only a lab trick?
In 2016 Damián Blasi and colleagues compared basic words in 6,452 word lists, covering 62% of the world's languages and 85% of its lineages. Words for small contained the sound i more often than chance predicts. Words for nose leaned on n, for tongue on l, for round on r. Look at small things in a few languages:
- teenyEnglish
- petitFrench
- piccoloItalian
- micRomanian
- kicsiHungarian
- chiisaiJapanese
And the counterexample sits in English itself: big has the i, small does not. A tendency across thousands of languages is not a rule inside any one of them.
Where green ends
Drag across the band and stop where green turns into blue.
Use a mouse, a finger, or the arrow keys.
Now look at these two pairs. Which pair looks more different?
What happened
Both pairs are exactly the same distance apart in a colour space built to match human vision (OKLCH). One pair straddles the line you drew; the other sits on one side of it. When the pair across a name boundary looks more different, researchers call it categorical perception: the name stretches the colours on either side. In the lab the effect is small and is measured in milliseconds, so a single choice proves nothing either way.
Not every language draws your line
In the World Atlas of Language Structures, 68 of 120 sampled languages use one word for green and blue together; linguists call it grue. Only 30 have separate words like English. Russian goes further the other way: light blue, goluboy, and dark blue, siniy, are separate basic colours, as distinct as pink and red. In 2007 Jonathan Winawer and colleagues found that Russian speakers told two blues apart faster when they fell on opposite sides of that line. English speakers showed no such advantage, and for Russians it disappeared when they had to keep a long number in mind by rehearsing it, which keeps the language system busy.
A sentence that trips you
Read two sentences one word at a time, at your normal pace. Tap the box, or press the space bar, for each next word.
Tap here or press space to start
Did the second sentence make sense?
What happened
It is a correct English sentence. Read it as The horse that was raced past the barn fell and broke its leg, just as you would read The car driven past the house broke down. Your reading system does not wait for the full stop. It bets on the likeliest structure word by word, and raced is far more often the main verb than a hidden passive. When fell arrives there is no place left for it, and you have to go back and rebuild the sentence.
Psycholinguists call these garden-path sentences, after the phrase to lead someone up the garden path. The psychologist Thomas Bever used the horse in 1970; the version here adds a few words so that the trap does not sit at the end of the sentence, where everyone slows down anyway.
Why it matters
Prediction is what makes reading fast. The same habit that trips you here saves you time in nearly every other sentence you will read today.
The letters you like
Quickly, without thinking about it: choose the six letters you like best.
Now type your first name. It is used for one calculation on this page and then cleared; it is never stored or sent.
What happened
In 1985 the Belgian psychologist Jozef Nuttin found that people prefer the letters of their own name to other letters, without realising why. Two years later he repeated the test at 13 universities in 12 European languages: a letter that belonged to the chooser's name was picked about 30% of the time, any other letter about 20%. It is now called the name-letter effect, and it is strongest for initials.
Your name is the word you have heard addressed to you more than any other. Fondness for the whole seems to spill over to its parts. One person's six letters prove nothing; the effect shows up only across many people.
The law in your words
Write for a minute about anything: what you did yesterday, a place you love, a person you miss. Or paste something you have written. At least 80 words.
What happened
In the 1930s and 1940s the American linguist George Kingsley Zipf noticed that in any long text the second most common word turns up about half as often as the first, the third about a third as often, and so on down. In the Brown Corpus, a million words of American English printed in 1961, the appears 69,971 times, nearly 7% of all words, and of 36,411 times, slightly over 3.5%. The dashed line on your chart is that ideal. In large collections of text, between 40% and 60% of the distinct words appear only once.
The pattern turns up in every language where it has been looked for, and in texts no one planned. You just wrote one. A short text is noisy, and your slope will wander; the longer you write, the closer it usually gets to −1. Why language settles into this shape is still argued over.
Your portrait in language
Each experiment adds something to this picture: the shape is how you heard bouba and kiki, its size your mil or mal, the colour your green line, the ring your reading of the trap sentence, the letters your favourites, and the dots your own words, ranked.
Sources
Every figure on this page comes from the studies below.
- Köhler, W. (1929). Gestalt Psychology. New York: Liveright.
- Ramachandran, V. S. & Hubbard, E. M. (2001). Synaesthesia: a window into perception, thought and language. Journal of Consciousness Studies 8(12), 3–34.
- Ćwiek, A. et al. (2022). The bouba/kiki effect is robust across cultures and writing systems. Philosophical Transactions of the Royal Society B 377: 20200390. doi:10.1098/rstb.2020.0390
- Sapir, E. (1929). A study in phonetic symbolism. Journal of Experimental Psychology 12(3), 225–239.
- Blasi, D. E., Wichmann, S., Hammarström, H., Stadler, P. F. & Christiansen, M. H. (2016). Sound–meaning association biases evidenced across thousands of languages. PNAS 113(39), 10818–10823. doi:10.1073/pnas.1605782113
- Kay, P. & Maffi, L. (2013). Green and Blue. In Dryer & Haspelmath (eds.), The World Atlas of Language Structures Online, chapter 134. wals.info/chapter/134
- Winawer, J. et al. (2007). Russian blues reveal effects of language on color discrimination. PNAS 104(19), 7780–7785. doi:10.1073/pnas.0701644104
- Bever, T. G. (1970). The cognitive basis for linguistic structures. In J. R. Hayes (ed.), Cognition and the Development of Language. New York: Wiley.
- Nuttin, J. M. (1985). Narcissism beyond Gestalt and awareness: the name letter effect. European Journal of Social Psychology 15(3), 353–361; and (1987) Affective consequences of mere ownership: the name letter effect in twelve European languages. EJSP 17(4), 381–402. doi:10.1002/ejsp.2420170402
- Zipf, G. K. (1949). Human Behavior and the Principle of Least Effort. Cambridge, MA: Addison-Wesley. Brown Corpus counts as given by Fagan & Gençay (2010); hapax share from Kornai, Mathematical Linguistics (2008).
Another experiment on how reading takes over attention, the Stroop effect, has its own page: The colour or the word?