JB Say What?

A few days ago, I started a post about a recent article in PLOS about tool use in rodents, but I got distracted. So today, as promised, I’ll talk about this study, and what it means about animal intelligence.

It seems to me that we often think about intelligence in other species in exclusively anthropocentric ways. This was brought home by the experience of one of the members of my lab who recently returned from a trip to Thailand and Cambodia. At one of her stops, she went to a demonstration of “Elephant Art.” The handler sets up an easel with a canvas, gives an elephant a specially modified paint brush dipped in paint, and the elephant goes and paints her art. Here are a couple of pictures, unretouched, showing part of the process:

It is important to note that no human hand was involved in the elephants painting the picture. In fact, you can see on the whole process on this YouTube video. It is somewhat of a long video, so if you’re bored, just skip to the end and take a look at the picture. What you’ll notice is that it is the exact same picture created by the exact same elephant but at a different time. What has happened, obviously, it that that particular elephant was taught to paint those particular lines, which end up looking like a “self portrait.” It turns out that each “Elephant artist” has his or her own specialty, which they recreate on a daily basis, as can be seen in this photo.

Now don’t get me wrong. I’m plenty impressed with the skills of the trainer and the steady trunk of the elephant. As anyone who has seen my own attempts at representational art will attest, I really couldn’t do much better. And I don’t doubt for a minute that elephants are intelligent animals. But this parlor trick is no more a sign of intelligence or artistic ability than any shaped animal behavior, such as trained dolphins or seals at an aquarium.

So what does this have to do with the article in about rodent tool use. Basically, these researchers trained degus, rodents that are indigenous to Chile, to use a small rake to grab food pellets that were just out of their grasp. Although the article is quite technical, it is worth navigating to the web page to see the movies of these critters at work. As someone who has seen his fair share of rodents, the behavior of these degus is pretty impressive. The authors are quite careful demonstrating that these animals acquired the “idea” of the tool, and were not just trained to use this particular tool in this particular way. But in reading through the article, you find that it took these animals 2700 trials over 56 days in order to reach a point where they were correctly using the tool 75% of the time.

So let’s think for a minute about what this means in terms of intelligence. To repeat: This animal was presented with a rake every day for nearly two months and after 2700 trials, he accurately employed this tool 3/4 of the time. Can you think of any task that this animal would have to perform in the wild where this kind of training regimen and success rate would be acceptable? Avoiding predators? Finding food? Attracting mates? Any animal that took 2 months to figure out how to do any of these tasks wouldn’t survive very well. But wait a minute, I hear you say, these are instinctive behaviors: They aren’t intelligence.

So if I hear you right (and I think I do), you are making the tacit assumption that instincts aren’t intelligence. That, my friend, is where you’re wrong. Instincts are very much intelligence. In humans, we have, to coin a phrase, a language instinct: most scientists agree that our brains are uniquely wired to acquire language. It is not taught. With regard to non-human animals, the behaviors and skills that they have that enable them to survive in their niches are very much intelligence as well.

I’ll close with brief story. Researchers have brought dogs into labs and tested their memory. While a restrained dog is watching, they place a doggie treat in one of two identical dishes and then put up a barrier between the dog and dishes. If they remove the barrier and let the dog loose within 5 seconds, he will immediately go the correct dish. If they delay the dog by 30 seconds, however, the dog will perform at chance level—he will be equally likely to chose either dish. Now imagine this same dog given a bone that he then buries in the backyard. That dog will find that bone months later without any hints of any kind. So in the lab, the dog’s “memory” is less than 30 seconds, but in the wild, his memory is measured in months. The key here is that there is no ecological significance of the experimental paradigm in the laboratory for the animal. What seems so obvious to us—”Hey, dummy, the treat is in this bowl over here”—is not at all relevant to the dog. When put in its own niche, the dog performs perfectly intelligently.

This post is way too long already, so I’ll stop for now. More in future days.

Posted by Glenn on Thursday, April 3rd, 2008 at 8:28 pm.