The other night someone I was with mentioned a _nauseous smell_. I thought: huh, interesting! I was of the impression that _nauseous_ was an experiencer-taking predicate (I feel nauseous, nauseous individuals), and that this (and no doubt many other) individuals had done the experiencer/stimulus dance to let that which causes nausea be called _nauseous_.
I later became rather embarrassed that I hadn’t remembered the old fake usage guideline that, in fact, _nauseous_ is only to be used for the stimulus, and _nauseated_ only for the experiencer. So in effect I not only didn’t have that guideline in my grammar, I felt sure (momentarily) that the standard was the exact opposite!
Now, first off, if you look in any dictionary or usage guide you’ll see that experiencer-_nauseous_ is widely accepted and basically unexceptional. At the same time, _nauseated_ is said to be rather rare (whether the frequencies take into account the sense of _nauseous_ is unclear; the lexeme is overall more frequent on Google, though interestingly not in the [BYU TIME corpus](http://corpus.byu.edu/time/)).
But further, I asked myself if I even _make_ the distinction between _nauseous_ and _nauseated_, ever. Certainly I don’t think I use stimulus-_nauseous_. Do I use _nauseated_? I have no idea. I don’t think so, but I couldn’t guarantee it. There must be some reason I thought the _nauseous smell_ use was non-standard, and I don’t think it’s because I had done some sort of strange prescriptive rule-reversal.
(Then there’s the unambiguously stimulus-selecting _nauseating_, and I’m pretty sure I use that.)
So, in conclusion…valence alternations and semantic change: it’s weird! (or am I weird from it?)
(for fun, search Google or whatever for “nauseated smell”)