According to an article published last week in Current Biology, African elephants in captivity “can use human pointing clues to find hidden food.” Elephants aren’t great at this. But they are as good as human two-year olds. And that’s pretty good.
The bottom line: you can show an elephant where you hid the food by pointing.
Scientists have wondered whether our ancestors domesticated dogs (and other animals) because of their natural ability to read our gestures, eye movements and body language. Or whether they have developed sensitivity to us as a result of the long history of their domestication, whether we have bred this into them.
The authors of the elephant study, Anna F. Smet and Richard W. Byrne, were interested in this question when they turned to study elephants. On the basis of their finding, they conclude, “the elephant’s native ability in interpreting social cues may have contributed to its long history of effective use by man.”
Pointing is certainly a remarkable communicative act and understanding pointing is no mean feat.
Humans point. We point with our hands, and eyes, and even our chins. Pointing is for us an effective way to direct the attention of another to what interests us.
When you think about it, it is a puzzle that pointing works at all.
Point to something in your vicinity such as a chair. Now point to its color. Now to its location. Now to the materials out of which it is made. Now to its shape.
Did you do something different each time? Very likely the hand gesture didn’t change.
It would seem, then, that in order to know what someone is pointing at, you need to know, already, what they are pointing at.
And yet we find it natural and straight forward to point and we can usually tell what a person is pointing at.
How does this work?
The solution to this puzzle, as the philosopher Ludwig Wittgenstein noticed, is to appreciate that pointing requires lots of stage-setting
We point in context, in the course of communicating. We say things like “look at the crazy color of this chair” and we accompany these words with a pointing gesture. And we add: “No, not this one, that one,” as we make another directing movement of the hand.
There is no such thing as bare, context-free pointing. And so we don’t confront the task of deciding, absent all context, what a person is pointing at.
The point is not that we always use words to disambiguate our acts of indication. The point is that we always indicate things by pointing within a context of shared interests, goals, concerns and saliencies.
Which brings us back to elephants. As I understand Smet’s and Byrne’s findings, the conclusion should not be that elephants can treat an act of a pointing as a clue to the location of hidden food, as their article’s title suggests. That wouldn’t be understanding pointing at all. After all, when I point to the chair I am not giving you a clue to what I’m interested in. I’m telling you. No, the real story, it seems, is that the elephant understands.
This understanding, just as in the human case, requires shared context. And that’s just what Smet and Byrne provide. The food in their experiments was placed into buckets in full view of the elephants, but in such a way that the elephants couldn’t tell into which bucket the food was put. Now the buckets are placed before the elephant and it is shown, with a pointing gesture, where the food is. Elephants like food. And so the elephant cares. It pays attention. It understands.
The fact that elephants can understand what we’re are pointing at suggests that we and elephants have more in common than we have previously admitted.