It seems obvious that both the zenophobic and the altruistic aspects of human nature must have originated as part of the development of the human social calculus in early group environments, so it is reassuring that Santa Fe Institute researcher Samuel Bowles and colleague Jung-Kyoo Choi of Kyungpook National University in South Korea have shown how game theory can support co-evolution of such traits. But how realistic is it for them to go on to suggest that the bad bits can be eradicated from people just by education? That sounds suspiciously like 'tabula rasa' thinking.

All the evidence points in a different direction. In reality, the tiny proportion of humans who have managed to restructure themselves away from their genetically murderous propensities get swept aside once the mob takes over. You can take the human out of the group, but you can't take the group out of the human.

What is to be done, then? Will it one day be possible to chop neural activity so finely that aggressive instincts against 'others' can be damped down without dulling humans' entrepreneurial and innovative skills? Will a New Model Human be constructed in robotic form which will simply take over from the existing model? Will people learn to control themselves using cognitive implants which override 'negative' impulses in certain types of interactive social situation?

It is easy to say goodbye to war. But it will be a brave committee that sits down to define the limits of behaviour allowed to fans at football matches.