A theory should primarily match the data and thus we should first seeif it in fact does so. Proving a characterization theorem that shows theequivalence of one formulation of a theory to another formulation doesnot, by definition, prove anything about the data. The two formulationswill be just as close to, or far away from, the data. So you have to askyourself: ‘Why am I doing this and what purpose does it serve?’But from my point of view, it is more important to axiomatizeparadigms than theories. With respect to the axiomatization of aparadigm, I can tell a coherent story of scientific development whereaxiomatization would play a major role: scientists are dealing withvarious problems and we could help them. For example, manyeconomists are developing models. And the questions that we canaddress are: ‘What models and theories should they be using?’ and‘In which language should they be formulating their models, whenthey develop them?’ Here, an axiomatization can help. And in such a
way it could also help in behavioural economics and neuroeconomics.But I do not think that these fields have yet developed such a paradigm.To the extent the behavioural economics has a paradigm, it seems to bethe same rational-choice paradigm of economic theory.
Com vocês, o grande Itzhak Gilboa (que tem um livrinho bacana sobre escolha racional).