- You have to define an end condition. (Really this is no different than the 2nd clause of a C-style for() loop, or the expression evaluated by a `while()` loop.)

- You have to explicitly specify the inputs for each iteration of the loop (e.g: explicitly declare what you need in the function signature, instead of just using whatever variables happen to be in the parent scope as working memory.)

- It makes optimization concerns very quickly apparent, as you'll pretty quickly blow the stack or start waiting for the heat-death of the universe. The best heuristics I have for a quick optimization pass are: is it tail-recursive, and does it needlessly recompute values?

Once I realized recursion is just another way to tell the compiler "repeat this thing until [x] is true", it became a lot less mystical and much easier to reason about.

I'd agree with OP to an extent that syntax did make a huge difference in how I viewed recursion. I understood recursion on a theoretical level for a while, but avoided it whenever possible because I too thought it was difficult to reason about.

It wasn't until I learned Elixir that it "clicked" for me. Between pattern matching and multiple function heads: Elixir just made recursion really easy to think about. (Not having any traditional loop construct probably helped a bit, too.)