By Alex Fradera
We have a mostly impressive ability to identify people we know based on the sound of their voice, but prior research has uncovered an intriguing exception – we’re not very good at discerning identity from laughter. Now Nadine Lavan and her colleagues have published research in Evolution and Human Behavior that looks into why this might be and what it says about our evolutionary past.
There are two main reasons why laughter may be hard to read. It might be because when you laugh for real, your vocal apparatus and lungs act in an involuntary way, unlike in many other vocalisations, and this difference in production might block the identity cues normally found in the voice. Alternatively, listeners of genuine laughter might get distracted by it’s contagious emotional content, meaning they don’t pick up on the identity cues.
To test these possibilities, Lavan’s team recruited helpers to produce fake laughs (while trying to sound natural) and genuine laughs, provoked by playing them amusing sound and video clips. A team of judges then listened back and rated all the laughs as highly authentic sounding genuine laughs, unauthentic sounding fake ones, and also genuine and fake laughs that sounded somewhat authentic.
The researchers categorised the laughs this way is because genuine laughs can sometimes sound flat, whereas a forced laugh can sound remarkably authentic (as you’d know if you’ve ever been around laughter yoga – or participated, as I did this month, in a live-action game session where you could use gestures to command people to laugh at anything you said).
Next, 37 participants, around 24 years old, completed hundreds of trials, in which they listened to two of the recorded laughs, each a few seconds long, and had to decide whether they came from the same individual or not. As in prior research, the participants’ accuracy was poorest when judging the identity of two spontaneous laughs, rather than deliberate laughs.
But what about authentic sounding fake laughs? If fact, participants were just as good at discerning identity from authentic sounding fake laughs as from non-authentic sounding fake laughs (and the degree to which spontaneous laughs appeared authentic also made no difference to performance). It was only the fact of being spontaneously produced that made laughs harder to identify, which suggests the difficulty reading identity isn’t due to listeners getting distracted, but by the fact the sounds are squeezed out in an involuntary way that is so different from human intentional speech.
The new findings fit with an evolutionary account that is speculative but intriguing. It goes like this: animal vocalisations evolved primarily to signal information about threats or food, not identity, and even today our evolutionarily old, spontaneous vocalisations, including laughter, remain as personally undifferentiated as those of animals (which would explain why can’t distinguish slowed-down spontaneous laughter from slowed-down animal calls). However, as humans became increasingly comfortable with making voluntary sounds through developing speech, we began to lean towards peppering those sounds with elements that are distinctive. In ever-larger social groups, it became important for us to signal identity, a prerequisite for developing trust and reciprocity over time. According to this story, the distinctive voice is a hallmark of our species that we acquired on the road to becoming savvy social beings.