Several years ago a book I tried to write turned out a mess. Readers got angry with it, and with me for creating it, even though anger was the last emotion I had hoped to evoke. That spurred me to research, and become a big advocate of, narrative nonfiction techniques. The problem with that manuscript, I realized, was that I had set up a story and not provided an ending. So after investing themselves in the story, people were upset that I didn't deliver what I had implicitly promised.
Tilly has a similar though slightly broader view. He suggests our "reasons why" break into four categories: conventions, simple stories, codes, and technical stories. (Gladwell explains each more detail.) They are not necessarily hierarchical -- some reason-categories are better for some situations -- but we have evolved rules for what we expect. Certain situations require certain types of reasons, and if the reason doesn't fit the situation, we are upset or otherwise dissatisfied.
It's a new way to look back on my messy manuscript. I started with a technical story (as, I would argue, most books do -- and our affinity for "technical stories" may explain why many people become writers), but degenerated into conventions and simple stories. It's not that I didn't have an ending, but that my ending didn't fit the reasoning it should have.
Which provides me with an ending -- or reason -- to my own engagement with that manuscript.
I'm always interested in feedback, via info at johnclaytonbooks.dot.com