Could humans have evolved a parietal eye (or perhaps a "pineal eye") which then migrated to the back of the head?

It is worth saying, though tautologically true, that evolution can't do what evolution can't do, and that the genetic code for new structures and functions has to be combinatorially accessible to the evolving genome. But the modularity of genetic regulatory networks (GRNs) can produce big rearrangements at a single step: redirection of a single "pointer variable" could see hardwired data structures and algorithms evolved within one sensory/cognitive module suddenly duplicated within a new brain region. (See the idea of William Calvin and Derek Bickerton that nested syntax resulted from transposing tree structures used in arm-hand-finger coordination to vocal communication.) Alternatively, polyploid mutations can duplicate at the genetic level a whole GRN, which permits divergent evolution of recycled algorithms - the tree structures for syntax could then be modified without impairing manual dexterity.

So your point is valid, but don't underestimate the power of evolution!