This is currently a raging problem in of all places, the quite foundational discipline of physics. In his recent book, The Trouble with Physics, the distinguished physicist Lee Smolin argues that Newtonian physics is good enough for macro, quantum physics is good enough for micro, but both rest on a common foundation that is another matter entirely. Thus, good enough is not good enough is you want an integrated theory of everything. The problem is assuming that something is foundational when it is not. Thus a final theory ain't so final.
For the psychology of motivation, the foundation is simple, namely a biologically informed and testable theory of how experience changes behavior. This is called a theory of learning. Unfortunately, a biological learning theory constrains how you can interpret observations about behavior. So when you ignore learning theory, you can't build psychological castles in the sky because there is no 'under' under. Thus, social, evolutionary, humanistic psychologies do without a neural foundation (in spite of their lip service to its importance) and equate approximate solutions with exact ones, which end up in a metaphorical sense, leaning. But that's OK if you tilt your head, and your reasoning within it, just so. But as this writer confesses, it is a common skill that I have never learned.