'If we look at it simply as a theoretical model, the Keynesian system is sound enough. It is consistent in the sense that, if we grant the premises, the conclusions will follow: the "level of incomes and employment" will be determined by the well-known determinants. The real issue is precisely whether the premises can be granted: to what extent they reflect reality. In Schumpeter's words, The realism of Keynes's "vision," not the logical consistency of his system is at issue.' -- Capital, Expectations, and the Market Process, p. 135
This, of course, is completely contra the people who claim that Keynes's system is illogical, self-contradictory, etc. The very first time I had to teach Keynes, and actually had to study him, I realized that not only was he consistent, there is also a beauty and elegance to his model. Of course, that says nothing about how closely it corresponds to reality.
The people who claim that Keynes's system is illogical actually mean one of two things, I suspect:
1) It went over my head, so the system must not make sense.
2) I don't like the policy conclusions he reaches, so the system must not make sense.