PhD defence: How our brain keeps vowels apart
Cup or cap, the difference between these two words is just a vowel. But in order to understand what someone is saying, it is important that we are able to recognise such differences properly. Nadine de Rue studied how Dutch vowels are stored in our brains. She will defend her thesis on 27 January.
Although the difference between two words is sometimes only one sound, this has major consequences for the meanings we give to them, as in the case of the Dutch words 'buur' (neighbour) and 'boer' (farmer). Our brain distinguishes the different characteristics of sounds without us even being aware of it, and it does so incredibly quickly. Linguist Nadine de Rue investigated how this is possible.
Contrasts
‘In the case of vowels, the brain needs to know which differences are important enough to be noticed and which are not. We call those important differences phonological contrasts', says De Rue. The difference between the sounds /u/ as in boer (/bur/; ‘farmer’) and /y/ as in ‘buur’ (/byr/; ‘neighbour’), for example, is caused by a contrast in tongue position. The /u/ is a back vowel, where the back of the tongue plays a role. The /y/ is a front vowel, where the front of the tongue comes up. Another important contrast in Dutch is lip rounding: the lips can be rounded (as with the /o/ in boot (/bot/; ‘boat’)) or not rounded (as with the /e/ in in been (/ben/; ‘leg’)). The latter contrast is found in very few languages. Together, the lip roundness contrast and the place contrast ensure that in Dutch you have a contrast between three types of vowels: non-round front vowels (like /e/ in been (/ben/; ‘leg’)), round front vowels (like /ø/ in neus: /nøs/; ‘nose’)) and round back vowels (like /o/ boot: /bot/; ‘boat’)).
Mental representations
But how can a Dutch listener recognise these differences? ‘You have to compare what you hear with what is stored in your brain. We call these stored forms mental representations. An important question within phonology is which information is stored in these representations', says De Rue.
Based on previous literature, De Rue assumed that not all phonological features are stored in our mental representations. For example, the 'front' feature of a front vowel is not stored, while the 'back' feature of a back vowel is. ‘A change from a back vowel (boot: /bot/; ‘boat’) to a front vowel (neus: /nøs/; ‘nose’) would therefore be more noticeable to the listener than the other way round', says De Rue.
'Objectively speaking, the difference between two sounds is of course the same, regardless of the direction of the change'
This is how it works: if a front vowel changes into a back vowel, there is no mismatch between the stored characteristics of the front vowel and the 'back' characteristic that the listener hears in the back vowel. But if a back vowel changes into a front vowel, there is a mismatch, because the heard characteristic 'front' of the front vowel clashes with the stored characteristic 'back' of the back vowel. De Rue: 'This has already been demonstrated in earlier research. In German, there is a greater perceptual difference when the back vowel /o/ changes to the front vowel [ø] (V/o/gel – V[ø]gel (‘bird – birds’)) than the other way round, while objectively speaking, the difference between two sounds is of course the same, regardless of the direction of the change.'
Experiments
In order to test how the mental representations of the tongue placement contrast (front versus back) and the lip roundness contrast (round versus not round) are stored in Dutch vowels, De Rue conducted several perception experiments, including EEG. EEG is used to register the electrical brain activity, in this case the reaction of the brain to vowel changes.
During the experiment, the participants heard a stream of individual sounds. For example, they heard an /o/ over and over again, but this stream was sometimes interrupted by another sound, such as an /e/. The listener then expects an /o/, but hears an /e/. The characteristics of the deviating vowel are compared with the characteristics of the expected vowel. De Rue was thus able to see whether the response differs in both directions and is thus asymmetrical or not. 'This gives information about whether a feature is present in the mental representation,' she explains.
Surprising asymmetry
The results of the experiments indeed showed asymmetries in vowel perception in Dutch. De Rue found an asymmetry in both the place contrast (front/back) and the lip rounding contrast. The asymmetry for lip rounding was surprising, because there is no such asymmetry in German. German and Dutch listeners therefore perceive the differences between the same vowels differently. Apparently, it not only matters which vowels a language has, but also how exactly they are used in a language.
Comparing languages
This is the first time that this asymmetry has been found in the contrast between round and non-round vowels. This asymmetry cannot be explained in exactly the same way as the asymmetry in place contrast that was already known. De Rue therefore proposes in her dissertation why there is an asymmetry for lip rounding in Dutch.
Follow-up research should show whether similar asymmetries can also be found in languages other than Dutch. Detailed perception experiments that compare different languages could provide more insight into the similarities and differences between different languages and language systems. In this way, more insight can be gained into whether certain phonological phenomena are universal or language-specific. De Rue: 'An eye for detail is thus important for insight into the big picture.’