I must say I would like to support the early observations of Brad 
Bateman that this discussion is flawed in parts, as well as Chuck 
McCann's observation that the work of Tony Lawson in particular 
(though amongst others) could usefully be consulted.
Lawson's JPKE paper "Probability and Uncertainty in Economic 
Analysis" (JPKE, Fall 1988), see 
<http://www.jstor.org/pss/4538115>http://www.jstor.org/pss/4538115 
is especially important.  Here Lawson explicitly contrasts the 
various conceptions in the literature (incidentally contrasting the 
objectivist 'logical' account of probability of the early Keynes with 
the subjectivist account of say Savage, and well as with the accounts 
of Knight and others).  Lawson also shows on many occasions that it 
was an opposition to the frequentist interpretation that led Keynes 
to oppose GE Moore's views on correct action, and to work on his A 
Treatise on Probability in the first place.  Moreover the question of 
right conduct taxed Keynes greatly as is clear in his (still largely 
unpublished) philosophical writings. The latter are well covered in 
Lawson's "Keynes and Conventions" (Review of Social Economy, 1993) 
and elsewhere.

Jane Sperry