By Maria Dokovova
1. Why did I start this?
It is widely perceived that second language listeners are better at understanding second-language accents rather than first-language accents. For example, as a Bulgarian whose second language is English, I am expected to be better at understanding Bulgarian-accented or foreign-accented English, rather than native English accents. Other people have put a name to this belief, calling it the Interspeech Intelligibility Benefit Hypothesis.
Different studies in the past have come up with contradicting results for this hypothesis. The majority reject the hypothesis; they find that hearing the accent of another language does not make it easier for the listener to understand what is being said, even if that other language is their native one. So, Dutch-English bilinguals do not benefit from hearing Dutch-accented English. There are some exceptions however. If the listeners have low proficiency in their second language, then they might find it easier to hear their first language accent while listening to the second language. So, as a learner of English in high school, it might have been easier for me to understand the accent of my fellow Bulgarian classmates rather than the American teacher.
Due to the fact that there was insufficient research in this area, as well as the fact that these papers were in constant disagreement with each other, I was inspired to study this specific area for my PhD at the Speech and Hearing department of Queen Margaret University.
2. What did my participants do?
In one of my experiments I collected data from 94 Bulgarian-English bilinguals living in the UK. I made an online experiment that I shared through social media among different Bulgarian groups across the UK.
I measured the participants’ English proficiency using a tool called Lextale; by testing to see if the participants can recognise and distinguish the real words from false words on the screen, such as real words “carrot” and nonsensical words “raccot”, Lextale calculates and comes up with a percentage score.
The participants’ understanding of accents were measured using a similar tool. They listened to real English words or nonsense words pronounced by either two native English speakers or two Bulgarians who spoke English as a second language. After hearing each word, they pressed a button on their keyboard to respond whether it was a real or nonsense word. I measured their accuracy and the speed at which they answered. It was assumed that a lower speed and lower accuracy meant that they struggle more at understanding the particular accent.
3. What did my bilingual participants actually do?
Overall, most participants responded more slowly and less accurately to a Bulgarian accent than a native English accent. Meanwhile, the listeners with the lowest English proficiency had no difference between the two accents, while the bilinguals with the highest English proficiency had slower and less accurate responses with Bulgarian-accented words. The latter even demonstrated slower and more inaccurate responses than the lowest proficiency participants. This means that the listeners who had the lowest proficiency in English were better at understanding Bulgarian-accent English than the listeners who had highest proficiency in English. However, because this study focused on Bulgarian bilinguals, this may only be the case for Bulgarian-accented English.
4. So what?
The main finding of this experiment is that no one was better (faster and more correct) at understanding Bulgarian-accented English than native English speech. In other words, this study did not support for the Interspeech Intelligibility Benefit Hypothesis.
It seems that Bulgarian-English bilinguals with higher proficiencies in English are specialised to native English speech at the expense of Bulgarian-accented speech.
As usual a study raises more questions than it answers! I wonder, would people who are just starting to learn English benefit more than more advanced learners from being taught by someone with a Bulgarian accent?
All of my participants lived in the UK – so would high English proficiency matter less if they lived in Bulgaria and were surrounded by speakers of Bulgarian-accented English?
I would love to see other people do similar research with different language pairs so that we know how much we can generalise the results. Another future step (and a challenge!) would be to study these relationships in a more natural setting, focusing beyond single-word recognition.
More details on this research available here.
