> Fish are a paraphyletic group: that is, any clade containing all fish also contains the tetrapods, which are not fish. For this reason, groups such as the "Class Pisces" seen in older reference works are no longer used in formal classifications.
I think we mostly agree, because you wrote
> There's no question that in general, categories of living creatures don't have to be monophyletic.
But I think that the main effort is to make all the categories monophyletic, and drop/replace the other one when possible. Some non monophyletic categories are very useful in practice and they are keep to make communication easier.
The problem is that if you try to make "fish" a monophyletic category you must put "dolphins" inside them, because the dolphins are distant descendants of some fish (as you and me and all mammals).
But Pterosaur and Dinosaur are sibling categories, neither of them descend of the other. Using much smaller categories, it is like comparing penguins and orcas (both are black and white, both swim, both have to breath air, both eat fish), but neither of them descend of the other.