I then introduced the concept of paradigm and specifically the concept of exemplars or shared examples. The argument here was that exemplars are necessary in the teaching of a discipline and without them there will be controversy. I claimed that thermodynamics does not have exemplars.

The crucial question then becomes «Why not?». I tried to answer this by looking at the role played by statistical mechanics in thermodynamics and kinetic theory since we might expect to find exemplars in statistical mechanics. But statistical mechanics is only a method of calculation. Without a consensus theory of probability embodied in models, there are no exemplars with which to structure the teaching of thermodynamics.