Neither of these arguments sits well with me, either, although I understand your point in both cases. To me, ``unique'' is the only word that means ``one-of-a-kind'', without having to resort to such linguistic hacks as ``one-of-a-kind''. There are a multitude of words that could replace ``unique'' in this context, and offer more meaning. ``Extraordinary'', ``odd'', ``remarkable'', come to mind, but, without knowing what he meant, I cannot place any one of those adjectives in its place.

And that is the crux of my point. When words lose their meanings, or those meanings get diluted, how are we supposed to express ourselves precisely? A computer science rule states that one should be strict in output and liberal in input. To that extent, I knew what he meant. But if I wanted to express that something was one-of-a-kind, I'd much rather use a succinct term like ``unique'' (``singular'' has already gone by the wayside), but I cannot, because I'll have to qualify that, by ``unique'', I actually mean unique.

It is not a problem of the language changing, it's a problem that we're losing the ability to express ourselves in an accurate manner, and that loss of expressing, IMHO, is part of a domino effect. When we lose the ability to express our meanings, then those meanings themselves start to erode from our consciousnesses. Of course, you could dismiss that as metaphysical claptrap, and I would have no recourse, as that's a fundamental belief, with only slight evidence to back it up.

But, to speak explicitly to your points:

you could make it more humdrum and it would still end up unique

I do not believe that ``humdrum'' is on the same axis as unique. That, to me, is like saying that ``you could make it more purple and it would still end up soft'', which is obviously true, but not exactly relevant. If this, in fact, what he meant, then I would argue that he should have used a word that was the opposite of ``humdrum''.

both those words were spelt "unique" and the second has already died out

But that's obviously not the case, or this argument wouldn't have started.

almost all uses of the word "literal" are metaphorical ... I don't think you have to be aware [of denotations] in order to employ their common metaphorical uses correctly

No, you don't. However, when you're not aware of the metaphor, and assume that the metaphor is the definition, what happens when someone uses the word denotatively? And, conversely, what happens when the listener isn't sure what the speaker means because he isn't sure whether the speaker is even aware of the dichotomy at all?

My point, again, is that these sort of ignorance-based modifications make it harder for us to communicate, at least in the realm where we don't know the abilites and knowledge levels of the people with whom we are interacting. I feel confident in speaking with you, and having you understand what I mean based on the fact that you've proven yourself to be knowledgeable and outspoken, which at least implies confidence in your verbal abilities, but I have to dumb down my language so much to speak to the unknown individual that it becomes hard for me to express myself.