In the span of a few thousand years human beings have achieved some remarkable feats, innovating and crafting a rich web of traditions and beliefs that we pass from one generation to the next. The young learn from the old, and not only master traditional techniques but reshape them, creating a dynamic culture in which creativity and inventiveness yield increasingly complex solutions. But beyond intelligence and opposable thumbs, how did humans develop culture in such myriad, complex ways, when other brainy, dexterous species did not? Tests comparing young children, chimpanzees and monkeys suggest that collaboration, rather than competition, may be the key to human culture.

Some behavioral scientists suggest that although cultural traditions exist in many species, only humans possess a culture that accumulates. Put simply, if culture is evidenced when an innovator's idea, behavior or tool spreads across a population and persists over generations, then many animals possess culture. Dolphins, for example, appear to pass from mother to offspring clever foraging uses for conches and sponges. But in cumulative culture, individuals transform and improve on a cultural idea, increasing its complexity and sophistication. Psychologist Robert Kurzban of the University of Pennsylvania has compared an example of consistently simple tool use in chimps, such as stick tools to fish for termites, with the evolving advancements in human technology, such as the transition from an abacus to a calculator.

"In cumulative culture, you end up with something that's way more complex than anything anybody could have devised for themselves," says Kevin Laland, a biologist at the University of Saint Andrews in Scotland.

Last week Laland and his colleagues at Saint Andrews, The University of Texas M. D. Anderson Cancer Center, the University of Strasbourg in France and Durham University in England published a study that they believe reveals why humans alone possess a capacity for cumulative culture. Using a learning test, they examined the performance of children three to four years old and compared it with that of chimpanzees and capuchin monkeys. Given that all three species have evidenced some degree of culture, the researchers hoped to glimpse the evolutionary variation in different parts of the primate family tree.

Each species was studied in small, mixed age and gender groups and presented with a special puzzle box. The box had three increasingly difficult levels. When done in order, each level could be solved in part by using a trick learned at the previous level and adding on a solution. For example, the first level involved sliding a panel, whereas the second involved sliding a panel and pushing a button. The model therefore tested for cumulative learning, a foundation for cumulative culture.

Opening each level of the box led to a reward: fruits and vegetables for the nonhuman primates and stickers for the children. The rewards scaled up in quality and quantity with each level, meaning bigger, more sparkly stickers for the kids and more coveted snacks, like grapes, for the chimps and monkeys.

While the subjects worked on the puzzle box, the scientists observed each group to evaluate various behaviors—such as vocalizing, watching one another, imitation and sharing or stealing rewards. They also attended to the social rank, sex and age of subjects to assess their influence on how subjects learned. Overall, the children outperformed other primates at more difficult puzzle levels. Although all groups performed well at the first level, most chimps and capuchins were stymied by the second and third. In short, only the children demonstrated cumulative learning.

So what did the kids do differently? During the task they actively talked to, instructed and imitated one another. They were cooperative and even shared their rewards. These behaviors, which were not observed in either of the other two species, were therefore strongly correlated to success in solving the puzzle box. The researchers concluded that these so-called pro-social behaviors indicate the importance of social cognition in developing cumulative learning and culture. "There was so much more going on in the children," Laland says.

In contrast, the chimpanzees and capuchins treated the box competitively. Rather than instructing their offspring, mother chimpanzees snatched treats from them, and neither group seemed interested in learning from another's actions.

The conclusion: humans possess cumulative learning because of the way we help one another learn. "What sets us apart is the manner in which we humans work together to solve problems," Laland says. "We help each other, imitate each other, we give each other verbal instruction. These abilities seem to be critical to cultural learning."

Thus, we may owe our technical prowess and social sophistication to our outgoing natures and teaching skills. Kurzban, who co-authored a perspective piece in the same issue of Science, praises this experiment for its attention to some of the social and cognitive mechanisms that may influence human culture. "It foregrounds some of the most important psychological systems that influence human learning," he says.

Still, he cautions, the study's findings are principally correlative, and more experiments will be needed to examine the causes of cumulative learning. For example, preventing the children from speaking to one another could be a useful way to evaluate whether verbal communication gave the children an advantage over their fellow primates. Additionally, bridging the study of cumulative learning with the evolution of cumulative culture will require more complex investigation, and we are a long way from understanding how we evolved these capacities.

Kurzban also points out that other abilities may influence cumulative learning. For example, studies find that not only can children learn by imitation, as other animals do, but that they detect intention and act to achieve another's intended goal without relying on mimicry. "It's actually amazing because a person's intention is sort of invisible, and yet from a relatively young age kids can make inferences about invisible desire," Kurzban says.

The ability to recognize that other individuals have separate and discrete minds—or theory of mind—has also been attributed to nonhuman animals, and like culture the claim is contentious. Just this week a study published by theoretical biologists at the University of Groningen in the Netherlands, challenged a claim that the western scrub jay has this capacity. The jay—which caches food to save for later—will move and re-cache the same food item repeatedly when another bird is present. Some see this as evidence that the jay recognizes the other bird's intention to steal cached food and that the jay in turn understands that moving his stash will protect the treat. Using computer modeling, however, the Groningen biologists have demonstrated that the same behavioral outcome occurs if the jay simply caches food as part of its stress response. Like Occam's razor, when a simpler explanation for an animal's behavior exists, claims of culture or theory of mind are difficult to defend.

Yet other explanations persist: Primatologist Frans de Waal of Emory University has written extensively on the existence of culture and empathy in chimpanzees. He believes that the study comparing children, chimps and capuchins suggests humans are better equipped for certain forms of learning but does not rule out the possibility that the chimps and capuchins could be capable of cumulative learning.

"The absence of a capacity is far more difficult to demonstrate than its presence," de Waal says. He observes that although the puzzle-box experiment has successfully avoided some of the pitfalls of comparative studies, there are confounding factors, such as the children's ability to speak as well as the unequal value of rewards. The children cooperate and share stickers, which though prized have a largely symbolic value. In contrast, even a well-fed animal recognizes the value of food as something crucial to survival.

Furthermore, de Waal believes that there is already evidence for cumulative culture in chimpanzees, such as populations that create and use tools for cracking nuts and extracting termites. "It is hard to imagine that chimpanzees socially learn these more complex behaviors without first learning more basic forms," de Waal says.

Laland, for his part, is concerned that the evidence for culture in nonhuman species is still circumstantial. The innovations are basic enough that it does not rule out the possibility these tools developed separately rather than building on one another. One species, Laland concedes, has unusually persuasive evidence: the New Caledonian crow, an animal far away from humans in evolutionary terms. The species, which like the western scrub jay is a member of the Corvidae family, makes tools by shaping pandanus tree leaves to forage for insects in wood. A survey of these tools has revealed three distinct designs, and the geographic overlap in their use suggests they did not evolve independently but may have been created by modifying one form of tool into another—evidence for cumulative culture.

The crow example may seem primitive when compared with the technical achievements of humans. Yet it raises questions about the capabilities involved, how they evolved and how many animals might possess these capacities. If nothing else, it reminds us that we are looking at a spectrum of abilities—and the differences, as Darwin suggested, are more likely to be of degree than kind.