Don’t Bother Trying To Direct An Elephant To Lunch By Pointing


Asian elephants fall into this odd space between wild and domesticated. They’ve been used as beasts of burden in places like India and Thailand for hundreds of years, but they have never been domesticated. Instead, they are trained, beginning at the age of just two or three, and then put to work for the rest of their lives, sometimes until after they reach age 60.

That odd space got some scientists, led by the University of Cambridge in England, wondering how Asian elephants might respond in something called the “object-choice experiment.” In that test, two containers are placed in front of an animal subject. One of the containers is filled with food, and the other is empty, but the animal can’t tell just from looking at them. Then a human experimenter points or gazes at the container with the food.

Some animals figure out where the food is, but many don’t. Domesticated species — including dogs, cats, horses, and goats — can perform this trick, and so can some wild species. There’s a theory, though, that domestication gives animals an evolutionary heads up in the task, priming them to understand the gaze or actions of humans.

But maybe Asian elephants could also do this, the researchers posited, since they spent so much of their lives around people. And so the scientists tested seven elephants at the Golden Triangle Asian Elephant Foundation in Thailand, setting them up with their personal mahouts (caretakers) as the experimenting human, as seen here in drawing A:


Drawing B is what’s supposed to happen. Unfortunately, the scientists were wrong, and the elephants proved unable to follow the pointing finger or directed gaze of their mahouts. The pachyderms were successful, though, when they were given verbal instructions meaning things like “left” or “right.”

So what does this mean for the domestication theory? Well, it might be another notch showing that the theory holds true, but perhaps not. As the researchers note in their paper in PLOS ONE, the mahouts use more than verbal cues when directing their elephants. They sometimes point, directing an animal to pick up a tourist’s dropped flip-flop, for example. And so the mahouts were actually upset when their elephants didn’t seem to pick up their non-verbal instructions in the experiment; after all, their animals did something similar every day. Why not then?

Maybe it was just that the experimental setup was an unfamiliar situation, the researchers suggest, and when put in such circumstances, the elephants react better to verbal rather than visual cues.

I’d be curious to see what would happen if the elephants knew the test better, or if the experiment was adapted to be more like what the animals encounter every day. Perhaps then they’d show, in the experimental setting, that they can pick up on the directions from their mahouts as well as they do in their day-to-day lives.

Top elephant image courtesy of flickr user Werner Witte; Diagram of experiment setup from Plotnik JM, Pokorny JJ, Keratimanochaya T, Webb C, Beronja HF, et al. (2013) Visual Cues Given by Humans Are Not Sufficient for Asian Elephants (Elephas maximus) to Find Hidden Food. PLOS ONE 8(4): e61174. doi:10.1371/journal.pone.0061174