乌鸦考古原文，求确认~ - ChaseDream

Some of you know that Bernd Heinrichhas spent many winters studying ravens and their behavior. This month, Heinrichand his colleague, Thomas Bugnyar, published an article in Scientific American that explores the intelligence ofravens. In this article, they investigate the question; do the birdsconsciously contemplate alternative behaviors and choose the most appropriateones, or are they merely relying on instinct or learning to perform specificactions by rote?

They begin by noting that ravens arenot the only birds that are reputed to behave intelligently. They state thatother relatives of the ravens -- the corgis, such as crows, jays, magpies andnutcrackers -- appear to possess surprising and sophisticated mental abilities.They even mention that these birds' capacities appear to be equivalent to or toeven surpass those of the great apes. For example, nutcrackers have thecapacity to recall thousands of locations where they have cached food items --a capacity that exceeds that of humans.

Which leads us to ask; do corvidsrely on logic to solve problems or are they relying on instinct? Do corvidsdistinguish between each other and alter their behaviors accordingly? To moreprecisely determine the mental capacities of ravens, the largest of thecorvids, Heinrich and Bugnyar designed several tests. The first experimentconsisted of food hanging from a string below the bottom of the wire cage(pictured right, bigger).To get this treat, the bird had to reach down from a perch and grasp the stringin its beak, pull up on the string, place the loop of string on the perch, stepon this looped segment of string to prevent it from slipping down, then let goof the string and reach down again and repeat its actions until the morsel offood was within reach.

They found that some adult birdswould examine the situation for several minutes and then perform this multistepprocedure in as little as 30 seconds without any trial and error -- as if theyknew exactly what they were doing. Because there was no opportunity for thebirds to be confronted with a similar problem in the wild, the simplestexplanation is that they were able to imagine the possibilities and to performthe appropriate behaviors. The authors also found that successfully performingthis behavior required maturity: immature birds were unable to do it whileyear-old birds performed a variety of trials before they were able to succeed.

But was it logic that the birdsrelied on to solve this problem? The authors assert that, basically, knowinghow to do something requires few or no trials, whereas trial-and-error learningrequires no logic. In fact, it was possible that the birds were rewarded byhaving the meat become closer with each looping behavior. So as a result, theauthors designed another experiment to find out how the birds were solving theproblem by presenting them with a situation that was not immediately rewardingbecause it was counter-intuitive: a string that must be pulled down to causethe food to move upwards towards the bird (pictured left, bigger).

In this situation, the ravens werestill interested in the food but none of them managed to solve the problem ofobtaining it even though they would have had to use the same sequence ofactions. The authors concluded that the pull-up method of obtaining the meatwas mastered quickly because it was logical -- a capacity that is lacking orpresent only to a limited extent in most animals.

Thinking and logic can be quiteunreliable and can cause their own set of problems. For example, paper waspsrely on precise hard-wired behaviors to manufacture paper into a nest with avery precise architecture. No learning is required to create the nest, althoughthe environment can modify some genetically programmed behaviors. So why arecorvids different? What is special about their social environment that favoredthe evolved of intelligence as the source for complex behaviors?

Much of the natural history ofravens suggests that they evolved under circumstances that required them tocope with rapidly changing short-term situations. These birds are opportunistswho do some hunting on their own, but are mainly dependent upon food that otheranimals have killed. The predators that inadvertently provide the birds withfood are unpredictable and thus, can also kill ravens. Under thesecircumstances, trial-and-error is evolutionarily untenable because the firstmistake in dealing with an unpredictable predator could cost the birds theirlives.

Food bonanzas provided by mammaliancarnivores are often quickly consumed by them. As a result, it pays ravens toget an early start in feeding -- often, side-by-side with these carnivores. Todo this, the birds must be able to predict the carnivores' behavior, such aswhen they might attack, how far they can jump and how to distract them, andsome of that knowledge needs to be in place before the bird itself isdistracted by feeding.

Juvenile birds learn these thingsearly in life by interacting with the predators through testing theirreactions. Juvenile ravens often will land nearby and nip them from behind.This so-called risky "play activity" is dangerous but ultimately aidsin the birds' survival by providing information about the capabilities ofvarious predators. By deliberately provoking them, ravens learn which animalsthey can trust and how far away they must stay to remain safe.

Ravens also cache food -- busilyhauling it away, burying it in secret locations and eating it later. Becauseravens have a nearly nonexistent sense of smell, they must memorize the preciselocation of this stored food, as is the case for other birds that also engagein caching. However, unlike most other caching birds, ravens observe caching oftheir competitors and thereby memorize the precise locations of not only theirown caches, but also those of their competitors. Because of this, ravens preferto cache their food in private.

As newly-fledged birds that arestill being fed by their parents, young ravens practice caching by hidinginedible items. Not only are the young birds learning which items are edible,but equally important, they were also learning to predict their siblings'behavior -- namely, cache pilfering. To better understand practice caching andpilfering behavior, the authors acted as surrogate parents to several youngbirds. One person was designated the "thief" and always stole a youngbird's cache, whereas the other person consistently examined the young birds'caches but never pilfered them.

When the thief was nearby, theauthors found that the young birds significantly delayed the time they waiteduntil they cached their food, and they relocated those caches they previouslymade. In contrast, the presence of the nonthief did not elicit these behaviors.These experiments reveal that the young birds improved their food-cachingskills after others raided them, but they also learned to distinguishindividuals, in this case, human thieves from human nonthieves.

Because wild ravens typically feedin groups, it is nearly impossible for them to cache food without another birdcatching them in the act, thereby learning the cache's precise location. Thus,it is important that ravens identify individual birds, just as they are able todistinguish between humans. So the authors designed another experiment wherethey tested this ability in ravens.

A large aviary was designed forcaching. Inside a smaller cage within this aviary were two ravens; one was a"knower" bird that was able to observe the test bird's cachelocations, and the other was a "nonknower" bird that had not observedthe test subject's caching behavior. The cacher was allowed to make threecaches and then that bird was removed from the aviary. Within five minutesafter the termination of the caching behavior and the removal of the cacher,either the knower and nonknower birds were allowed to go into the caching arenato search for food.

Knowing that caching birds oftenretrieve their food when robbery seems imminent, the authors tested the cachingbirds' behavior when they were in the arena privately, when the nonknower birdwas present, or when the knower bird was present in the aviary. They found thatthe caching bird retrieved its food stores more often when the knower bird camewithin two meters of the cache. Thus, the authors speculated, the caching birdwas able to identify which bird had observed it making its caches and was able todiscriminate between it and the "nonknower" bird.

The authors also found that theknower birds were careful about their intentions; they did not go to the cacheswhen the caching birds were nearby, but instead, they waited until they were atsome distance. This suggests that both birds had the ability to attributeknowledge to specific individuals and to anticipate a particular response.

The authors used another version ofthis same experiment to determine if the caching birds made subtle cues thatthe knower birds might be able to decipher. To do this, a human make thecaches, and then stood by passively while either two knower birds, or a knowerand a nonknower bird were placed in the aviary together. As predicted, theknower birds were quick to pilfer human-made caches when paired with anotherknower. However, when paired with a nonknower, the knower bird's reactionsdepended upon social rank. When a knower was paired with a socially dominantnonknower bird, the knower would delay approaching the cache -- instead,waiting until the dominant bird was some distance away before pilfering thestash. Thus, the authors concluded that it was unlikely that the knowers wereproviding some behavioral cue that cache raiders might use.

In conclusion, the results of thestring-pulling experiment indicate that ravens rely on logic to guide theiractions. The results of the pilfer anti-pilfer experiments show that ravensreact to their competitors based on what they remember them being able toobserve, that they can accurately attribute the capacity of knowledge to theircompetitors, and that they integrate this knowledge together with social statusto make strategic decisions when retrieving food caches.