For best experience please turn on javascript and use a modern browser!
You are using a browser that is no longer supported by Microsoft. Please upgrade your browser. The site may not present itself correctly if you continue browsing.
Bekijk de site in het Nederlands

Languages develop structure as they are passed down from generation to generation. The same is true for the sound structure of languages, according to UvA researcher Tessa Verhoef. Her work marks the first study into the emergence of structure in speech sounds using continuous signalling with participants in a lab environment.


Language adapts

Language is something that is passed down from one generation to the next, each unconsciously introducing its own subtle changes. ‘Language adapts to what our brains find easy to learn. Things that people can’t remember, won’t be used and reproduced. And because people make unconscious mistakes and leave out structures or combinations of sounds that are hard to remember, language eventually gains regularity and becomes more learnable’, explains Verhoef.

‘Chinese whispers’

Natural languages around the world are already highly ordered. Thus, to enable her to study the emergence of structure in language, Verhoef decided to use an entirely new language, one that is whistled instead of spoken. Basically, the format of the experiment can be compared with the game of ‘Chinese whispers’. In this game, a message is whispered from one person to another until it has been transmitted several times. The last person then speaks the whispered words aloud. Often, the original message has changed substantially. This process demonstrates on a small scale what happens when languages are repeatedly learned and passed down through generations.

Exhibiting structure

In Verhoef’s experiments, simple, artificial languages were transmitted from person to person. Participants had to produce the words of these languages with a slide whistle. Participants were given one hour in which to learn the artificial whistled language, which they then had to reproduce. Those reproductions were then used to train the next participant and so on. ‘In the beginning, no structure could be detected in the artificial languages at all’, says Verhoef. ‘But after several rounds of passing it on to new participants, the languages began to change and they became more regular. The chaos of sounds characterising the original language was reduced to a set of basic elements that were systematically reused and combined. In other words, the artificial whistled languages began to exhibit structure, while the participants did not consciously invent this.’

The way in which structure emerges in the artificial languages of Verhoef’s experiments dovetails with phonological theories on the principles of efficient coding.


T. Verhoef, Efficient Coding in Speech Sounds: Cultural Evolution and the Emergence of Structure in Artificial Languages. PhD thesis supervisors: Prof. P.P.G. Boersma and Prof. S.M. Kirby. Co-supervisor: Prof. B.G. de Boer.

Time and location

The graduation ceremony will take place on Friday 27 September at 12:00.
Location: Agnietenkapel, Oudezijds Voorburgwal 231, Amsterdam