This seems consistent with what we've learned about Hacker News. What defies expectation tends to be interesting. The predictable is uninteresting.
Summary: What makes a new theory interesting? If it contradicts your assumptions, it's interesting. If it reaffirms your assumptions, it's not. Here is a list of sociology papers which the author considers interesting, with some historical context for each.
A funny thought: If you expected a theory to be interesting but it reaffirmed your assumptions would it not be interesting? :p