His research shows that we are more likely to absorb, retain, and share an idea if that idea is slightly anomalous. If an idea violates one or two basic, intuitive assumptions about a thing, it has a far stronger chance of being recalled and transmitted.

