Isn't a method worrying about whether it needs a lock or not and being coded appropriately way better (since the writer of that method/function thinks through it when writing it) than relying on someone who is maintaining code to think through the use-cases and call the appropriate non-recursive version of a child method/function depending on whether what they are writing has already taken the lock or not? In an ideal world, yeah, the maintainer would carefully worry about every line of code; but in the real world where the maintainer is enhancing functionality or fixing issues, it's too much to expect diligence on the part of the maintainer and to err on the side of safety at the cost of recursively taking locks. Recursively taking locks doesn't cause real problems except when there are other locks taken in between which might cause dead-locks with other parts of the code that take the same "other" locks without the recursive lock being taken.
