Z can come up with some real world examples of this phenomenon.
Of words that have changed lexical category more than once through zero morphology? Can't off the top of my head, and I agree with you that something would've changed along the trail, e.g., meaning, form.
I think it's not an issue in NLP (natural language processing). I suppose it depends if one thinks that part of speech-ness is some property inherent in a word or whether, as many would think, that it is how a word is used syntactically that determines what it is. The former handles words like love (noun and verb) which not even peevologists find anomalous owing to its antiquity, but the latter would seem a more robust way to develop a word tagger. (It should also be easy enough to rewrite any tail-recursive function as an iterative one, and in writing any function one should take into account infinite recursion or endless loops.)
