Phonosemantics
A made-up word can feel round before it means anything. In 1929, Wolfgang Köhler showed Spanish speakers two shapes and asked which one was maluma and which one was takete; the soft word went to the curved shape, the hard word to the jagged one. Ramachandran and Hubbard repeated the trick in 2001 with bouba and kiki, and reported about 95% agreement among English and Tamil speakers. That is the crack in the old claim that language is arbitrary all the way down.
How it works
Phonosemantics is the study of whether speech sounds carry non-arbitrary associations with meaning. The modest claim is easy to defend: humans often map sound features onto shape, size, speed, texture, or emotion. The larger claim is where the fight begins: some researchers think these mappings are tiny residues inside a mostly arbitrary system, while others think they helped language get started.
The mechanism is probably not one thing. Rounded vowels like /u/ pull the mouth into a rounder shape. Plosives like /k/ and /t/ arrive as sharp acoustic events. Front vowels like /i/ often feel small or bright; back vowels like /u/ often feel large or dark. None of this turns sound into destiny. It gives a speaker a weak prior before dictionary meaning takes over.
A compact way to say it:
meaning_guess = convention + sound_bias + context
Convention still does most of the work. The sound bias is the interesting residue.
Where it shows up
The cleanest evidence comes from nonsense words, because nobody can hide behind learned vocabulary. The 2021 cross-language study by Ćwiek and colleagues tested speakers of 25 languages across 9 language families and 10 writing systems. The effect survived, which weakens the claim that Roman letters alone explain it.
| Example | Sound pattern | Usual association | Source |
|---|---|---|---|
| bouba / kiki | /b u/ vs /k i/ | round vs spiky | Ramachandran & Hubbard, 2001 |
| maluma / takete | sonorant vowels vs plosive stops | curved vs jagged | Köhler, 1929 |
| mil / mal | high vowel vs low vowel | smaller vs larger | Sapir, 1929 |
| Himba participants | no written language in test group | similar shape-sound match | Bremner et al., 2013 |
This matters because writing is a late technology. If pre-literate children and communities with little exposure to alphabetic forms still show the mapping, the signal sits closer to perception than typography.
What's contested
The effect is real enough to study, but its size depends on the task. Forced-choice experiments can exaggerate certainty because the participant must choose between two options. A person matching bouba to a blob in a lab is not proof that Sanskrit, Japanese, or English vocabulary is secretly encoded by sound.
The deeper dispute is about origins. Ferdinand de Saussure made arbitrariness central to modern linguistics in Course in General Linguistics in 1916. Phonosemantics does not defeat Saussure; it narrows him. The sign is mostly arbitrary, but not perfectly arbitrary.
Why this has to do with other realms
The link to concept information theory is sharper than it looks. Shannon stripped meaning out of communication in 1948 so he could measure signal transmission cleanly. Phonosemantics puts a little meaning back into the signal before the dictionary arrives.
There is also a strange bridge to mission voyager 1. The Golden Record carries greetings in 55 languages, but an alien listener would not inherit Earth’s phonetic priors. Bouba may feel round to many humans; it may mean nothing to a receiver built under another atmosphere.
An open question
If sound symbolism gives children a small head start in word learning, how much of early vocabulary is shaped by perception before culture hardens it into grammar?
Key sources
- Wolfgang Köhler, Gestalt Psychology (1929) - the maluma/takete starting point.
- Edward Sapir, “A Study in Phonetic Symbolism” (1929) - early evidence that vowels bias size judgments.
- V. S. Ramachandran and E. M. Hubbard, “Synaesthesia: A Window Into Perception, Thought and Language” (2001) - the bouba/kiki version that became canonical.
- Andrew J. Bremner et al., “Bouba and Kiki in Namibia?” Cognition (2013) - Himba data that tests the culture and writing objection.
- Aleksandra Ćwiek et al., “The bouba/kiki effect is robust across cultures and writing systems,” Philosophical Transactions of the Royal Society B (2021) - 25-language replication.
Further reading
- Ferdinand de Saussure, Course in General Linguistics (1916) - the cleanest statement of arbitrariness as the default rule.
- Mutsumi Imai and Sotaro Kita, “The sound symbolism bootstrapping hypothesis for language acquisition and language evolution” (2014) - why this may matter for child language.
- John Ohala, “The frequency code underlies the sound-symbolic use of voice pitch” (1994) - a path from pitch to social meaning.
- concept fermi paradox - a useful antidote to assuming human signals travel with human meanings.
Abhishek's take
What grabs me here is the weakness of the effect, not its strength. A 95% lab choice sounds like a law until you remember that real words survive by convention, memory, trade, prestige, and accident. Phonosemantics is useful because it shows meaning entering through the side door, not because it replaces the dictionary.
Tags: #language #cognition #sound-symbolism #semiotics #perception
See Also
- concept information theory
- concept fermi paradox
- mission voyager 1
- dest proxima centauri
- mission breakthrough starshot