In certain societies, people鈥檚 views naturally migrate over time towards extreme ends of a political spectrum, where they will become entrenched. This, at least, is what a mathematical analysis of human behaviour by Andr茅 Martins of the University of S茫o Paulo in Brazil has shown.
Martins鈥檚 computer model is designed to simulate the way opinions can spread through a society. It is based on a mathematical model of the way atoms align their magnetic fields. The underlying idea is that individuals鈥� opinions can be influenced by the views of their neighbours, just as the orientation of an atom鈥檚 magnetic field tends to line up with that of its neighbours.
While such an approach undoubtedly simplifies human behaviour, it has already shown some success by predicting people鈥檚 voting patterns in the build-up to an election (New Scientist, 24 August 2002, p 42). In that study, people鈥檚 views were modelled as two simple states of mind. Now Martins has extended this to allow each human 鈥渁gent鈥� to hold a range of opinions across a defined spectrum.
Advertisement
Martins assigned to each agent a view close to the middle of the spectrum, and then left them to interact with their neighbours. As the simulation progressed, each individual鈥檚 conviction shifted, based on what they learned about their neighbours.
The game always ended up with most of the agents鈥� views entrenched at one or other end of the given spectrum (). Martins says he was surprised by this, as he had expected the agents鈥� opinions to drift back and forth indefinitely.
Martins accepts his model is 鈥渃learly a simplification of the real world鈥�. In most real societies, he points out, those with extreme political opinions are in the minority. 鈥淯nder some circumstances, though, it might be inevitable,鈥� he says. 鈥淎n interesting next step is to try to understand when it will happen and why.鈥�
鈥淚n some circumstances, extreme opinions may be inevitable. An interesting next step is to try to understand when and why鈥�
Scott Atran at Michigan State University, who specialises in the psychology and anthropology of extremism, speculates that the model does seem to reflect some societies in the Middle East. 鈥淚n Palestine and Lebanon extreme attitudes towards Jews have permeated all of society,鈥� he says.
Dietrich Stauffer of the University of Stuttgart in Germany, who ran the election simulations, thinks Martins鈥檚 approach based on a range of opinions should be used with other models aiming to analyse the shift of societal views. 鈥淚f they all give the same kind of result 鈥� that extremism emerges automatically 鈥� then this is something to take seriously.鈥�
Not all researchers are convinced. Sociologist Duncan Watts at Columbia University in New York dismisses this approach to modelling social networks as 鈥渆xtremely unrealistic鈥�.
