It is interesting that there are mid-19th century antecedents, 
whether in von Thunen or Mill or others.  Of course, Keynes's 
Treatise on Probability came out the same year as Knight's book, 
1921, and many observers would say that his analysis of the matter 
was far more sophisticated, allowing for gradations between pure 
objective probabilities such as flipping a fair coin to pure 
uncertainty, varying with weights of subjective belief by 
observers.  Also, I have seen a claim that Schumpeter made the 
distinction while discussing entrepreneurship in the German edition 
of his Theory of Economic Development, published a good decade before 
the dual appearance of the famous books by Knight and Keynes.

Given how well known Keynes's work is, quite aside from all this 
earlier work that is less well known, I find it bizarre that someone 
would assert that "we all know" that the distinction was due to Knight.

Barkley Rosser