Understanding ignorance

[Aad Ruiter]

Dear all,

During the conference we heard different interpretations of ZI. On the one hand, there is ZI as the emergence of (rational, intelligent) order from randomness; on the other, there is a  more instrumentalist interpretation of ZI as a reflection of ignorance.

In this latter approach, constraints capture what we believe we know, e.g. institutional constraints (which are easier to verify than, say, preferences). Interestingly, institutional constraints go a long way in explaining aggregate outcomes, or to be more precise in reducing entropy.

I believe that the instrumentalist approach can be fruitful in behavioral economics too: for instance, a constraint may reflect that people tend to discard dominated strategies.

The goal would be to find a set of constraints (institutional and / or behavioral) that is as tight as possible, without violating observed data. For a very impressive application of these ideas to Wolfe's dice, see "Understanding Ignorance" of Rodriguez (in Maximum-Entropy and Bayesian Methods in Science and Engineering (Dordrecht: Springer, 1988), pp. 189-204).