Ready for everything – Accommodating variability in accented speech

In an increasingly globalized world, one of the most important things we have to do is understand people from different parts of the world and the things they say. Not only are there an estimated 5,000 different languages on Earth today, but also the number of non-native speakers of English vastly outnumbers the number of native speakers.

Of course, this problem generalizes to all sorts of language communities where immigration or international collaboration is happening. It is for this reason that it is important for speakers of a language to be able to successfully adapt to non-native accents and grammatical structures.

We constantly have to adapt to everyone’s voice and style of speaking, even people who share the same dialects! Everyone’s voice is unique; so one person’s “ah” might sound like another person’s “eh”. If we did not figure this out quickly, we might not understand what others were saying until minutes later into the conversation. So how do we do this? We have to be very good at predicting – for example, we can make good guesses about what my vowels are likely to sound like due to my gender.

Prediction is much harder for non-native speakers, who rarely ever achieve native pronunciation and intonation in their second languages if the languages are acquired later in life. Even though non-native speakers can very easily achieve native-like grammar, with rich vocabularies and idiomatic expressions, pronunciation difficulties may pose the greatest barrier to understanding people in their non-native language.

One consequence of sounding like a non-native speaker is that we are trusted less, and what we say is not taken at face value. One particularly striking example is that foreign accents make us less likely to believe statements made by non-native speakers, as measured by a truth rating task. For example, if you hear a statement that “Ants don’t sleep”, you will likely rate the truth value of that statement lower if it is spoken with a heavy foreign accent than without.

This is true even when we are merely summarizing facts said by native speakers. It is clearly maladaptive to disregard something someone says just because they have an accent. We ought to put in a reasonable amount of effort to listen to non-native speakers.

But just how difficult is this?

Recently, Witteman, Weber, and McQueen published a study in Psychonomic Bulletin and Review looking at one factor which plays a role in our ability to understand non-native speech: the consistency or predictability of the accent itself. In contrast to native speakers of a language, non-native speakers exhibit considerable variability in the way they articulate sounds. Speakers of German, for example, alternate in their use of [g] and [k] in English in a way that is often consistent with German pronunciation, but which seems to have no discernable pattern to an English speaker. So, a word like “hog” would sound like “hock”, even though a German speaker can say words like “green” with an English [g].

Thankfully, we usually handle this inconsistency very well, and rely more on other cues in the language such as the topic of conversation or the grammar to identify the word that was said. When we know someone is not a native speaker, we are better at accepting different pronunciations of things than if we believed that person was a native speaker.

Perhaps we are willing to accept just anything from a non-native speaker, or maybe we expect foreign accents to behave like dialects. Witteman and colleagues used a cross-modal lexical decision task which allowed them to look at whether a German speaker using a consistent accent in Dutch made words easier to identify. If a speaker uses an inconsistent accent, it should take more time to determine whether a word was simply mispronounced or not a word at all.

To test this, Witteman and colleagues played words and nonwords over headphones. Those same words and nonwords were then presented on a screen, where participants discriminated the written words from nonwords in a lexical decision task. The words could be said in either a consistent accent, or an inconsistent accent. Consistent accents had a predictably German vowel for a small percentage of the stimuli (e.g. hend for hand), while inconsistent accents used both the Dutch and German vowels (hend and hand). The authors predicted that the speaker with the consistent accent would make the discrimination of words and nonwords easier in the lexical decision task.

Indeed, the results showed that when the German speaker used an inconsistent accent in Dutch, the words were less identifiable than when words were produced with a consistent accent, which listeners grew to anticipate. Although in both cases people were able to adapt very quickly to the novel accent, it took people more time to adapt to the speaker who used both hend and hand instead of just hend.

The results of this study have the potential to have an impact on language instruction. First of all, it suggests that simply “having an accent” is not a real barrier to being understood. Initially, what matters is how predictable the accent is. This is good news for language learners! We are very likely to be understood in our foreign languages with some initial effort from our conversation partners. Witteman and colleagues note that “inconsistencies […] seem to create no major problems for native listeners: Although their adaptation is slowed down, they need only a few minutes to catch up.”

At the same time, there are still gaps in language instruction. Accent minimization clearly has a number of tangible and intangible benefits, such as being understood earlier or being believed about what we say. Because instruction and education at the university level is becoming increasingly international, it is important to see what other factors contribute to comprehensibility and what both listeners and speakers can do to close the gap.

You may also like