You can interpret the question of whether an abstraction “map naturally to most people's brains” in two different ways:
(0) Is it easy to understand how the abstraction works in principle?
(1) Is it easy to actually use the abstraction to build large robust systems?
Clearly, (0) is a necessary precondition for (1). But it's not sufficient. Pointer arithmetic is an example of an abstraction that's easy to understand “in principle”, yet hard to use right (for “most people”, anyway). I'm afraid the same is true of explicit continuation (“callback”) passing. Programming with tools that only work “in principle” is hell.
(0) Is it easy to understand how the abstraction works in principle?
(1) Is it easy to actually use the abstraction to build large robust systems?
Clearly, (0) is a necessary precondition for (1). But it's not sufficient. Pointer arithmetic is an example of an abstraction that's easy to understand “in principle”, yet hard to use right (for “most people”, anyway). I'm afraid the same is true of explicit continuation (“callback”) passing. Programming with tools that only work “in principle” is hell.