There are numerous reports of ravens dropping sticks, stones, and other objects. It is usually assumed that the birds are acting deliberately for some purpose. However, most reports of object-dropping behavior that I’m aware of are too incomplete for conclusions. Bob Sam told me of a raven in Sitka, Alaska, who took an unshelled walnut, flew up with it several times, and dropped it on the concrete. The shell didn’t break. Finally, after the raven dropped it on the street, he perched atop the Sitka Hotel. A car came along and ran over the nut. Then the bird flew down and ate the nut’s contents. Hilmar Hansen, a railway worker from Montana, told me of ravens placing deer leg bones onto rails he was inspecting, and coming back to feed on the marrow. These acts may or may not have been deliberate. More anecdotes, and especially more details to each anecdote are needed to come to conclusions as to what drives the raven’s behavior. Have there been a thousand unreported instances of ravens flying with nuts and eventually dropping them at random? Would a raven go to pick up the pieces of any bone or nut that was run over by a vehicle? Had the bird previously dropped a tough walnut in frustration at random, to be rewarded with food by a passing car? Did they learn by trial and error, and then know by insight what they had done? Intelligent behavior may result from a combination of curiosity, exploration, persistence, patience, keen observation, learning, and opportunism; but insight is difficult to prove because it is not necessarily a prerequisite for clever behavior.
Roger Smith of the Teton Science School wrote me: “I was banding young ravens on the east side of Teton Park in 1992. The nest was in a Douglas fir, about 35 feet up. While banding the third of five young, one of the adult ravens landed on a branch approximately 1.5 meters from me. This bird began vigorously tapping and rubbing its beak on the branch, vocalizing, and moving back and forth on the nest branch. I then noticed it had pulled a cone off the branch and began making low vocalizations with the cone in bill. By this time, I was banding the fifth nestling when, to my surprise I was hit in the face with a cone. I stopped and watched as the bird walked down the branch about two feet and pulled off another cone. At his point, I tried not to make eye contact with the bird but watched the behavior more closely. The bird walked back along the branch to the same distance from me, began the same beak tapping and vocalizations, then flicked the cone toward me, only this time the cone landed in the nest. The ‘throwing’ movement was incredibly quick. The bird flew off the branch only moments after this second encounter.”
It is not possible to say with certainty if that raven knew what it was doing, although it seems unlikely the bird had learned the act of throwing, that cones can be pulled off for throwing, or that throwing them at nest predators might deter them. Insight could account for all three.
Mental representation of successive images that are projected in the mind like a movie is consciousness and the modus operandi of intelligence. We use mental projection so routinely in almost all activity that we take it for granted. When we throw a ball, we see a trajectory to the intended target. We make choices to achieve very specific anticipated results. Without anticipating or projecting the results of the various alternatives to generating a sequence of actions, no intelligent strategy is possible. In effect, the fundamental capacity to develop strategy requires a capacity to visualize that which is out of sight, and that which has not yet happened but can happen. To ravens, that which is out of sight is, as with us, also not necessarily out of mind, as can be seen even in trivial examples.
One time, I had a white opaque plastic bag with a fresh DOR (dead on road) woodchuck. Goliath had seen me cut off a leg for him, and then put the chuck back into the bag. I walked away down a woods path. I hadn’t gone far before he was behind me, begging. Odd, I thought. How could he be hungry after I had just given him a whole woodchuck leg? I left him another leg. Instead of eating from it, he hurriedly hid it as I was walking off, and then flew right behind me, again begging some more. “Okay, Goliath, you know there is more in that bag. You want the whole thing. So here it is.” He does not follow me if he sees me put a piece of food into a bag, and I then retrieve it and give him the whole thing. He keeps track of objects even after he no longer sees them. Is that capacity advantageous in the wild?
Kristi Dahl published an anecdote in 1996 in Wyoming Wildlife that suggested the immediate usefulness of keeping mental track of objects out of sight. From her porch in Grand Teton National Park, Kristi observed Uinta ground squirrels scurrying around in the muddy sage grassland when the shadow of a raven overhead made them dash for their burrows. “We watched carefully as the adult raven landed nearby and approached a squirrel burrow…. The bird began to peck at the dirt, scooping loose soil with its beak. Suddenly, the raven stopped digging and gave a series of high-pitched, throaty yells. It was immediately joined by a young raven, evidently its offspring. The young bird began to beg and call as the adult continued digging…. Finally, about eight inches below the surface, the raven found its lunch. Pulling up a full-grown ground squirrel from the hole, it stabbed the animal several times…. The squirrel was torn apart and fed to the young raven before the pair flew off.”
It is only a small step from seeing something in the mind when it is out of sight, such as prey or an enemy, to remembering past moves and anticipating future moves and reacting appropriately. There are many suggestive examples that are consistent with the idea that ravens are capable of all of the above, although most of the examples are not tightly enough constrained to allow unequivocal interpretations. Here are a few of them:
As reported in the Manchester (England) Guardian of June 25, 1995, a trapper in Prince Albert National Park in northwest Saskatchewan observed a raven feeding on an animal killed by a wolf. The raven occasionally interrupted its meal to lie still on its back. Eventually, the trapper noticed that this happened every time ravens flew by overhead. He conjectured that the raven played dead so that to the birds overhead the scene would look like yet another raven had died from poisoned meat bait and would give the place a wide berth. Again, we need more details to draw such conclusions. Ravens roll on their backs in play. What, precisely, were the timings of the back-rolls versus the overhead flights?
Some of the varied uses of the bill.
When the woodfrogs were chorusing near my aviary, I got four from a local pool. I put the first frog out of sight into a square, brown, foot-long PVC drainage tube. The birds had played with the tube before and presumably knew it was hollow, but they were shy at their first sight of a frog. They perched and watched from up in their loft, about thirty feet away. As I walked away from the tube, Houdi flew down and first looked in one end, then ran around and checked the other end. The frog could not be reached from either end, but she picked one end of the tube up and the frog slid out. She grabbed it, flew back up onto her perch, ate the frog, then came down to walk on the aviary floor. When she walked past the tube, she did not peek in, apparently remembering that I had put just one frog in and she had taken that one frog out.
Several hours later, I put the second frog into the tube. Fuzz and Houdi were on full alert, intently peering down from their perches at the tube as I walked out of the way. Fuzz was the first one down this time. He bent his head down to ground level and peered into one end of the tube. He must have been able to see the frog, but he could not reach it. Was it therefore unavailable? No problem. Fuzz did not hesitate even a fraction of a second. He quickly walked around to the other end, reached in, and pulled the frog out. While he was eating it, I put the third frog into the tube. Houdi peered at the tube from her perch for a half minute before she flew down and looked in. Since the frog apparently had not moved, she simply reached in and pulled it out. After eating it, she retrieved the tennis ball she had played with previously and stuffed it back into the tube.
While the ball was solidly stuck into one end of the tube, I put the last frog in the other end, presuming it would escape deep into the tube to lodge against the ball. Houdi left her perch and looked in. The frog could not be reach
ed. It must have already moved down the tube to the ball. So she walked to the other end of the tube and tried to remove the ball. Fuzz joined her. He first peered into the open end of the tube, then walked to the other end and tried to remove the ball. It must have been stuck solidly, because it took him a full half minute to remove it. When he finally did, he grabbed the frog. Had he “filled in the blanks”—imagined where in the tube the frog lay? Do ravens have X-ray vision or can they can reconstruct spatial relationships in their mind? I speculated that the latter was more likely than the former.
Two days later, I introduced two round, white PCV drainage tubes into the aviary, four feet long and four inches in diameter. The tube was a different kind than ones I had used in previous experiments. As they always do with new objects, both birds inspected this new tubing from up on their perches, craning their heads down, twisting and turning their heads rapidly. Fuzz made deep rasping caws, a normal response to feared predators or strange things. Then he came down to make an inspection. He stood sleek and tall, and jumped nervously all around the tube as if performing a dance. He leaned down with his head close to the ground and peered into one end. He walked the four feet to the other end and peered in from that end also. Houdi meanwhile watched him from her perch. When he had finished and had gone back up to the perch, she joined him without examining the tube herself. I left the tube, being satisfied that they knew or would soon learn that it was hollow, not a white log.
In the meantime, I had introduced them to snakes. They had learned that snakes are good to eat and that they slither. Weeks later, I brought a live, foot-long green snake. Holding it by the tip of the tail, I dangled it in front of the white PVC. Then I dropped it. The snake slithered into the tube, and I stepped back. Both ravens had been watching me from their perches thirty feet away. Fuzz came down, walked to the mouth of the tube, looked in, then hopped rapidly the whole four feet around to the other side. He reached in, pulled the snake out, crushed its head, fed from it, and cached the rest, covering it with leaves. Houdi watched. A few minutes later, she went to Fuzz’s cache, retrieved the remains of his snake, and ate her fill also. Fuzz tolerated the theft because they were a pair by then, and he wasn’t very hungry. Later, with another group of birds, I dropped food into the same tube, now held up vertically. In that case, when I dropped food in front of them, instead of looking first from the perch at the end where they saw me put the food, they hopped down to the ground and went on their first try to the end where the food had ended up (snug on the ground and where it could not be seen except after digging). That is, they had anticipated its movement through the tube.
I wrapped a lump of butter in paper as the ravens watched me, and put it into the tube, stuffing wads of green foliage behind it. Fuzz did not hesitate to pull out all the foliage to get the butter. I then put an egg in so far that it could not be reached from either end. Fuzz checked into the tube from both ends, walking back and forth, for a total of eight inspections, first at one end and then at the other, as if not believing that what could be seen at one end could not be reached from the other. After the eighth look, he finally picked up one end of the tube. The egg rolled out, and he ate the yolk. One explanation for him picking up the tube might have been that in frustration he would try almost anything. Another is that he knew the egg would roll out. Not all apparently deliberate acts are as ambiguous.
A pair of wild ravens on the frozen carcass in front of my spruce blind were chipping little pieces of meat off one at a time by partially opening their bills and using the pointed tips of their lower mandibles as chisels backed by the force of the momentum of their swinging heads. In contrast, they pulled off softer meat in small chunks by grasping and pulling, using the small hook at the tip of the upper bill. Chunks of meat were stacked piece by piece into a pile. Finally, the birds grabbed their whole pile and flew off with it.
Ravens in crowds always act differently than those alone or in pairs. They never stack meat, perhaps knowing that any loose piece would instantly be taken by another bird. They instead either fly off with only one large piece at a time as soon as they have detached it, or fill their throat pouch with small pieces before flying off.
A large chunk of beef suet in the woods in the back of my house could not be pulled apart into pieces like meat. The frost-hardened suet could, however, be handled by hacking into it, chipping off small, loose pieces similar to how woodpeckers, crows, blue jays, chickadees, and nuthatches invariably do it. Ravens normally feed on suet in this way as well. But one day, one raven did something imaginatively different.
The raven pair that fed there never allowed me to get close. They flew off even when they saw me near a window. I had gone to the food at the edge of the woods, as usual, to provide new food, and as I started to walk there, the raven that had been feeding flew up. I had not observed its feeding behavior directly, but it had left its tracks in the snow. Most interestingly, it had also left an intriguing record of what it had done. This bird, rather than randomly picking off small chips for immediate eating, had carved a groove around one corner of the fat. This groove was carved using precisely aimed blows. The object had obviously been to cut off a manageable chunk from a larger, immovable one. Many of the small pieces of fat that had been chipped off during the cut were not even ingested.
I want to indicate briefly why I was so excited about the seemingly trivial fact of a raven carving a groove in fat, when birds are capable of infinitely more complex innate behavior, such as weaving ornate nests or navigating by sun-compass, using an internal clock. That the raven had carved the groove was a fact, and given that no other birds and perhaps only a very rare raven would do such a thing, it was a highly singular fact. One robin does not make a spring, but this one fact was visible proof that the bird had forgone immediate gratification for a reward later on. That’s planning. There were no secret raven trainers out there in the woods, it was not a learned plan. It was a plan derived from mental visualization. It was an invention. It made the raven’s life easier but was hardly necessary. Given the constraints of the alternative interpretations, what the raven had done was news. I felt it should be splashed across the front cover of Nature and Science. Of course I knew it would not be. As expected, it was rejected several times for publication because it was “just an anecdote.” Science, to be publishable, is almost defined as that which is strictly replicable, and there was no way that anyone could drop off a hunk of suet and expect a raven to slice a piece off by hacking a groove through it. Not even my aviary birds would do it. This was a raven Einstein.
The record of peck marks in the suet seemed to suggest that the raven not only had thought ahead, but also had acted on that thought and shown intelligence. Intelligence is not merely consciousness or awareness alone. It is not just complex behavior. Intelligence is not just super-detailed memory, rapid learning, complex vocal communication, play behavior, or tool use. Intelligence may or may not be related to all of these things, and some kinds of intelligence require them, but they are not what intelligence is. Intelligence is doing the right thing under a novel situation, precisely as this bird had done. Intelligence is understanding the world, and reacting appropriately to it, not just perceiving it. Intelligence is about awareness, and about testing responses in the head rather than the “real” world, where such activity may be time-consuming, harmful, or fatal.
Solving the meat-on-the-string puzzle.
TWENTY-SIX
Testing Raven Intelligence
RAVENS HAVE RELATIVELY LARGE brains, and they do many things that look intelligent, but there was still no experimental proof that they are intelligent. I needed new ideas to get out of a rut, but really new, novel ideas, seldom come to us by forethought or design. They come by mucking around even as we try not to. I got my idea for a new approach on examining raven intelligence by leafing through a copy of Ranger Rick magazine, a present to my then young son, Stuart.
The magazine contained a short article on the “clever” things you can w
atch birds do, like pulling up food suspended by a string. I could not believe that chickadees could actually do that. If they did, it seemed to me, they would have to be trained. And if you trained them first, then insight or mental visualization needed to perform the task could follow the task, but could not precede it. I dismissed the whole “intelligence” aspect of this, as I had with so many other stories I had heard.
My next thought was that if ravens were as intelligent as they sometimes seemed to be, it was possible that a rare individual among them might figure out how to pull up food on a string without having to go through lengthy trial-and-error learning. That is, insight might precede or accompany learning to produce the same behavior. The bird might evaluate the situation, play out a mental scenario in its mind, and from that insight perform the task quickly.
Testing experimentally for the existence of insight—the evaluation of different choices to make an intelligent decision without overtly trying them—may seem impossible to do because everything that an animal does includes prewired responses and learning. Neither of these can be simply excised from its brain.
Any one behavior is a combination of innate programming of blind or unconscious responses, learning, and insight. For example, our sexual preferences are largely innate, but giving flowers during courtship—preferably daisies rather than skunk cabbages—takes insight as modified by learning. There is no apparent limit to the complexity of animal behavior that can be innate or learned, provided it relates to the same conditions the animal has encountered predictably over millions of years. Insight requires consciousness, but it is more. It is the mental visualization of alternative choices that then guide judgment of new situations at the moment. It could be common in all sorts of behavior, but we can only assign it to play a likely role in behavior when it fulfills three criteria: First, to eliminate the innate component, it has to be extremely rare and exclusive of what the animal normally encounters and does. Second, it has to solve a problem. Third, it can’t be a learned response.
Mind of the Raven: Investigations and Adventures With Wolf-Birds Page 35
