The exposition of quantum phenomena concentrates on superposition and interference phenomena (the fact that quanta exhibit in certain classic experiments both wave and particle aspects) and the informal exposition of the theory highlights the dual aspect of quantum evolution, that is the deterministic evolution governed by the Schrodinger equation of an isolated system and the non-deterministic, discontinuous evolution at measurement described by the projection postulate
I will argue that the lengths to which one must go to avoid the projection postulate in the modal interpretation are in general too severe, especially because, as I suggest in Section 5, the usual worries about the projection postulate do not apply to it as it is (or should be) used in modal interpretations.
The statement of the projection postulate is apparently clear:
The change of state implied by the projection postulate is called 'collapse'.
One might try to soften the sting of these questions by motivating the projection postulate using notions borrowed from classical probability theory.
Can a similar argument be used to justify the projection postulate in quantum mechanics?
First, the projection postulate as stated above entails a collapse not when somebody discovers the state of a quantum system, but when a 'measurement' occurs.
This reply fails, however, because it destroys the very classical basis on which we hoped to justify the projection postulate.
Modal interpretations can get away clean with the projection postulate.
3 Why modal interpretations should adopt the projection postulate
Consistent with their desire to uphold the in principle predictive completeness of quantum theory, KD walk this tightrope without introducing any new elements into its formalism - not even the projection postulate.
But it is at least implicit in yon Neumann's  treatment of measurement, supplying him with the motivation for introducing his projection postulate.