Making sense of irrationality


I came across an article on the other day about a fairly new branch of economics called neuroeconomics. This branch of economics takes a cue from behavioural economics by incorporating another of the sciences, neuroscience, into its analysis of why and how people make decisions.
One of the breakthrough reports cited in the article used active MRIs to determine that failure to choose an optimal outcome in the ultimatum game correlated with an area in the brain that is said to be involved with reward and punishment decisions. For those of you not familiar with the ultimatum game:

"Take two people and tell them they have the opportunity to split $10. Furthermore, tell one person that, as first mover, they get to make a one time offer, and tell the other person that, as second mover, they get the opportunity to either accept or reject this offer. If the offer is rejected they both go home with zero."

Economists have found that many people would rather have nothing than accept the ungenerous offer of a dollar. Meaning that maybe for some people the incentive of a monetary reward to choose the classically defined optimal outcome is tempered by the emotional/mental reward of punishing the other person for their low-ball offer.
Some economists, and in particular two from Princeton, criticise neuroeconomics and behavioural economics because, in their opinion, it doesn’t matter what the mental process is; just the final outcome is relevant to the discipline.
On the other hand, I think there is a lot of appeal in studying why and how people make decisions. Firstly because it's interesting, and secondly because people make irrational decisions daily – even in the realm of economics.
Of course, people usually make decisions that provide the most benefit to themselves and to society – that's why the total sum of individual decisions will guide society towards optimal welfare like an invisible hand.  Yet it is equally clear that sometimes, some people don’t act rationally. It's worth considering why.