There is a principle that I want to refer to in an upcoming talk, but I can't find a name for it. Someone must have named it after themselves by now. It is related to Ockham's razor, and to parsimony, but is distinct.
The principle is that an explanation for a pattern should preferably be applicable as broadly as the pattern is observed, but not more broadly. For example, an evolutionary explanation for group living that can be applied to all the group-living insects is preferable to an explanation that works only for one species of ant. The explanation should preferably not explain the pattern more broadly than it occurs. For example, an explanation for the fusion reaction of the sun merely in terms of the presence of hydrogen would also tend to predict fusion in many other contexts where it does not generally occur, making this explanation less desirable than a more complex one which also specifies the need for the physical conditions which encourage the hydrogen to fuse.
This is the sort of logical statement that is so obvious as to rarely need to be said, yet I need to say it for this talk. If you know what I should call it, please let me know.