Willingham explains that knowledge often begins as inflexible and appears rote but, given more time and practice (and crucially, more related knowledge) inflexible knowledge becomes more flexible and transferrable.Â
Here is Daniel Willingham's article on rote knowledge.
I think the idea of inflexible knowledge explains why students who have been present a concept in only one context cannot generalize that concept well and have a hard time cuing up the necessary knowledge to problem solve when the problem isn't exactly what they learned in class.
Take for example this lesson on light being refracted. If all the examples done are of this form. The more dense fluid is always on top and the boundary between them is always horizontal...
Then it should be a surprise that students' inflexible knowledge produces these incorrect results on a test...