Sorry, I was a little inexact. The way an Oreo tastes to someone whose tasting-system has been interfered with is a property of both the Oreo and the interference, and in some cases (e.g., someone hypnotized to think they're eating a raspberry) it may be a property of only the interference; in any case, the differences between different such tasters' experiences is largely a matter of the interference rather than the Oreo.

Something a bit like this is true even without interference, of course. Different people have different experiences on tasting the same foods. Very different, sometimes.

But, again, none of this is a good analogy for the p-zombie or inverted-spectrum experiments. The way the analogy is meant to work is:

• Person : inverted spectrum :: Oreo : tastes of raspberries.
• Inversion is a difference in the person :: raspberry taste is a difference in the Oreo.
• In both cases the change is purely internal.
  • "Inverted spectra" pose no sort of difficulty for physicalism if what's actually happening is that person 1 sees red and person 2 sees green because person 1 is looking at a tomato and person 2 is looking at a cabbage.
• In both cases the only change is supposed to be non-physical.
  • Otherwise there's no argument against physicalism here.

And timujin is suggesting that the Oreo version of this is obviously silly, and that we should apply the same intuitions to the other side of the analogy.

Your introduction of hypnosis breaks the analogy, because now (1) the change is no longer "internal": the raspberry taste is only there if a particular person is eating the Oreo, and that person has changed; and (2) the change is no longer non-physical: hypnosis involves physical processes and so far as we know is a physical process.