True. Linneaus erred on various groupings, and included a number of mythological entities in his scheme. However, that Dolphins were not fish is a fact originally recognised from their physical morphology not from any application of evolutionary theory.
Similarities relating to physical morphology can be misleading.
For example who would've guessed that a whale is more closely related to a cow than a walrus based on physical morphology alone?
While relatedness provides a useful proxy for similarity (because of evolution) it's really simularity, not relatedness, that matters.
But to get to the underlying similarities, those similarities that are more than superficial, do we not need to understand relatedness?
AbE - I have just read this topic through in more detail and it would seem that my above comments and this line of debate are essentially off-topic. So I won't carry on this conversation here. Apologies.
Edited by Straggler, : No reason given.