If you’ve ever caught yourself talking to someone and thought, “Gee, I sound just like them”, it might be a sign that you’re engaged in the conversation or the task at hand.
The same goes, as a new study shows, for solving puzzles in an immersive virtual game environment.
Over the years, researchers have found that when two people discuss an alluring topic, the sounds of their speech are more likely to converge.
Without prompting, speakers may unintentionally start to say certain words like the other person. They may also change the syntax of their sentences to better align with their converser or alter their pronunciation to match one another.
Numerous experiments have shown the mimicking of sounds, or phonetic convergence, to be a prevalent feature of human speech.
But these changes seem to be subtly different depending on the content and context of a conversation, including the sex, race, and conversational role of the speakers, as well as the goal of the overall chat.
In 2018, experimenters found that when speakers are highly engaged in a task – in this specific case, dictating the colors of particular objects in the video game Minecraft to a partner – they emphasize words differently compared to simply reading colors off a boring old computer screen.
This leaves the possibility that the goal of a task and the level of participant engagement can influence the sounds of a person’s speech.
The new study from US-based researchers adds to the literature by extending the 2018 video game experiment.
The authors of the new research wanted to examine how speech production changes when a task is highly engaging and involves a conversational partner, as opposed to working on a rather dull task together.
Researchers split 52 native English speakers into pairs. These pairs then had to complete a series of tasks together.
Talker A had to help their partner, Talker B, who was seated in an adjacent room, navigate around a colorful Minecraft scene using target words laid out in the virtual maze that both participants could see on their respective computer screens.
In the less engaging task, Talker A merely had to read aloud words on a screen to Talker B, who then selected the right words on their screen.
The more engaging the task, the more convergence the researchers ultimately heard in the speech of Talker A and Talker B.
“By way of example, imagine Talker A is identified as having durationally longer productions compared to their partner, Talker B, at the start of an interaction,” the authors explain.
“If Talker A shortened their productions over the course of the experiment, then this would be taken as evidence that Talker A converged towards Talker B. Likewise, if Talker B lengthened their productions over the course of the experiment, then this would be taken as evidence that Talker B converged towards Talker A.”
The study suggests that a highly engaging conversation leads to more mimicking. Speakers in the experiment were more likely to adapt their speech to the sounds of their partner when their attention was more invested in a game.
This suggests that phonetic convergence might be a way for humans to create synergy with one another and reduce the chances of being misunderstood.
Like many before, the study is limited in that it only focuses on native English speakers and certain acoustic features of speech. Its sample size was also too small to tease apart the subtler acoustic changes that might occur during a conversation.
“It is one thing to focus on the sound level, but other things happen at the language level, like using words someone may not normally use,” admits communication scientist Navin Viswanathan from Pennsylvania State University.
It also must be noted that in the current study, the high engagement task resulted in significantly more conversation between the partners overall. This means it’s possible that speakers in the video game scenario simply had more opportunity to learn and mimic the nuances of their partner’s speech than in the more straightforward task.
Further studies are needed to explore how speech length may ultimately impact speech convergence, but these initial results suggest mimicry is not always mockery.
For better or worse, the voices of others can easily rub off on us.
The study was published in Speech Communications.