BioEd Online

The capacity to learn from others is one of the traits that have made humans such a global success story. Relying on it too much, however, could have contributed to the demise of past populations, such as the Maya of southern Mexico in the eighth and ninth centuries and Norse settlers in Greenland 1,000 years ago.

Over-hunting, deforestation and over-population are well-worn routes to societal collapse. Now, Hal Whitehead of Dalhousie University, Halifax, Nova Scotia, Canada, and Pete Richerson of the University of California, Davis, have modelled how different learning strategies fare in different environments. They found that conformist social learning — imitating and emulating what the majority are doing — may also cause the demise of societies. When environments remain stable for long periods, behaviour can become disconnected from environmental demands, so that when change does come, the effects are catastrophic¹.

Environments often change in unpredictable ways and over timescales from the seasonal to millennial. Rainfall and temperature change both seasonally and annually; populations of predators, prey and pests rise and fall; soil conditions change.

Behaviour that is genetically determined can adapt to environmental change by the slow process of natural selection, but only when that change is also slow. Rapid change puts a premium on the capacity of individuals to learn through exploration and experience, and to adapt their behaviour accordingly.

Figuring things out for yourself can be time-consuming, however, and a waste of time when others around you have already acquired the relevant knowledge. Once individual learning is common, the less costly strategy of conformist social learning can get a foothold.

The road to oblivion

Previous models have suggested that evolution will favour a mixture of the two strategies, in which individual learning is used in some circumstances and social or cultural learning in others. This means that populations retain the ability to learn new solutions to emerging problems but also benefit from the low cost of spreading these solutions through social learning.

Unlike previous models, Whitehead and Richerson's have explored how the different kinds of learning fare in 'red-noise environments' — when environmental variation is concentrated in large and infrequent changes, a pattern they say is characteristic of many historical records.

"During long periods with only modest amounts of change, conformist social learning is a more successful strategy than costly individual learning," says Richerson. In red-noise environments, this advantage remains true in the short term but can lead to long-term problems. "The mix of individual and social learning that evolves during quiet periods in red-noise environments tends to have too little individual learning to cope with rarer big changes," he says.

These results might explain some well-known historical crashes, say the authors. Mayan civilization famously went into free fall some 1,200 years ago — the cessation of temple building shows this — and was plausibly driven by a combination of ecological change and cultural inertia.

Similarly, Norse settlers in Greenland around AD 1000 were culturally conservative, carrying on much as they had in Scandinavia. They failed to adapt to an increasingly harsh environment or adopt the more effective behaviour of the Inuit, and eventually died out, possibly as a result of malnutrition.

Luke Rendell, a biologist at the University of St Andrews, UK, says the role of conformism in population collapse is plausible, even if surprising. "People might find it difficult to believe that humans, in all their complexity, would do something so stupid as to copy themselves to extinction, but in my view that may rely on an overly rosy view of human omnipotence," says Rendell. "What matters to most people is how they are doing as individuals right now, and longer-term considerations are very easily pushed down the priority listing."

Prizing innovation

Whitehead and Richerson's models highlight the perils of cultural conformism in red-noise environments, particularly when populations are small, but also show how other styles of learning can mitigate the problems. For instance, 'prestige bias' means that people only copy successful role models, rather than simply imitating what everyone else is doing.

"Societies should promote individual learning and innovation over cultural conformity, and the models for social learning should be individuals who have demonstrated that they understand how to live with the current environmental trends," says Whitehead.

The validity of these prescriptions may, however, turn on further work. "We lack empirical data on human behaviour with which to test these models," says Rendell. "Once we know more about how humans really employ learning strategies, we will be in a better position to evaluate the relevance of models like these to our own history and to our present."