Speech (and nonspeech) learning in adults

Perceptual adaptation to an atypical talker/accent in Greek listeners

In this experiment, we investigated whether upon hearing atypical realizations of some of words, adults attribute the atypicality to the speaker at hand and expect all his speech to be accented. 

Greek listeners were first exposed to atypical productions of their native words such as καφές or μπύρα (which are typically produced as [kafes] and [bira]) ‘coffee’ and ‘bier’, sounding more like [kafis] and [bera]. The listeners were then tested on how they categorize isolated instances of Greek vowels, produced by the same speaker. Our results showed that listeners adapted their comprehension of this atypical speaker, and did so for all speech sounds. This means that they created language generalizations over the input they heard. Interestingly, they did so even when the speaker produced nonsensical words. The adults could thus learn even without comprehending the lexical content (pretty much as very young babies do before understanding the meaning of words). This work was was done with Václav Jonáš Podlipský and Anastasia Chionidou (University of Thessaloniki), and has been published in the Journal of Experimental Psychology.


Adults’ ability to shift existing speech sound categories and to learn new ones

The goal of this experiment was to find out whether it is easier to learn second-language speech sounds that are similar to native ones, or whether it is easier to learn second-language sound that are entirely new.

One group of Spanish listeners was trained to shift a speech sound contrast that they have in their native language and another group of Spanish listeners to create an entirely new contrast. The first learning situation is comparable to a case where, for instance, a native speaker of Spanish (or Greek) starts to learn Portuguese (or Italian) and has to realize that the /i/ and /e/ she knows from the native language are in Portuguese realized slightly differently. The second learning situation is comparable to a case where the native speaker of Spanish starts to learn Czech and has to realize that a short /e/ is contrasted with a long /e:/ and that changing a short to a long vowel can change the meaning of words in the new language.

We recorded EEG from forty Spanish-speaking adults and assessed how they processed the novel speech sounds before and after 9 minutes of auditory training. We found that they could learn in the first situation (e.g. the Spanish learning Portuguese /i/ and /e/) but less so, or not at all, in the second one (e.g. the Spanish learning Czech /e/ and /e:/).

Data collection was done during Katerina’s PhD at the University of Amsterdam. The work is now under review.


How native language affects the learning of second-language speech

More info to come soon.


Projects