I read Klaus Havelund's thesis on the Fork Calculus:
He develops the Fork calculus for reasoning about concurrent functional programs, the motivation being Concurrent ML.
In chapter 5, he writes:
First, we add a recursion operator, allowing us to define processes with infinite behaviour. Recursion is not difficult to deal with (thanks to the use of operational semantics), except that the axiomatisation cannot be shown complete in presence of recursion.
There is no literature reference here, so I ask you: where can I read about why recursion makes a programming language's axiomatisation cannot be proved complete, and in precisely what meaning?