If you happen to be walking through Australian bush, you might hear a loud, deep rumbling sound that conjures images of a ferocious wild boar – and no one could blame you for panicking and running for dear life.

You might feel a little sheepish, though, after learning the sound came from a cute little koala (Phascolarctos cinereus).

These wily marsupials are not the only animals to make sounds that defy their body size, called “dishonest signalling” by biologists.

“Nature is full of animals like squeaky Rottweilers and tenor-Chihuahuas,” says Andrea Ravignani from the Max Planck Institute in Germany, co-author of a position paper published in the journal Biology Letters.

As a rule, body size is related to the frequency of sounds animals make (known as “acoustic allometry” in the world of biologists). But some find ways to fake their size and sound smaller or bigger than would be expected.

One impetus to produce different sounds is to attract mates who are drawn to larger body size or superior singing skills that reflect better reproductive success.

Could sexual selection for “fake signalling” have led to the ability to learn new sounds?

Although it’s rare in mammals, humans, seals, bats, elephants, whales and dolphins seem to be particularly clever at copying and learning novel sounds outside their vocal repertoire.

To explore whether animals that fake their body size are the same that can learn new sounds, Ravignani teamed up with Maxime Garcia, from the University of Zurich in Switzerland, to marry their respective work in these related but disparate fields.

The pair analysed the sounds and body size of 164 different mammals ranging from mice and monkeys to aquatic mammals such as the subantarctic fur seal (Arctocephalus tropicalis) and Amazonian manatee (Trichechus inunguis).

They combined methods from acoustics, anatomy, and evolutionary biology to compare the different sorts of animals in the dataset.

Results showed those deviant animals that acoustically fake body size were indeed better at learning and copying new sounds, paving the way for new evolutionary insights, according to Garcia.

“Importantly,” he says, “this shows that, in some mammal species, the ability of vocal learning may derive from sexual selection leading to an increased volitional flexibility of sound production.”

He says there may be a link to human speech evolution. “We believe that a ‘dishonest signalling’ strategy may be a first evolutionary step towards learning how to make new sounds of any sort.”