There's an argument that in the strict sense, everyone is rational because they are all maximising some utility - it just isn't clear what that utility is, eg. money, sex, power, the satisfaction of having a well ironed shirt, it could be anything really.

Which is why this definition of rationality isn't very practical. In the traditional economic sense, rationality means maximimising money. With this narrow definition (and big assumption), the economic equations become tractable and can actually be analysed. Of course this doesn't represent reality. Going back to Kahneman and Tversky with their study on decision-making under risk (probably even further back than them since this is pretty obvious), psychologists have shown that humans aren't very rational.

But economists still tend to rely on this assumption so that the game theory equations work. Having said that though, I'm not sure whether or not this assumption holds (at least reasonably well) in the macro sense.