Pondering at the arguments at wiki:
In other words, if a brain in a vat stated "I am a brain in a vat", it would always be stating a falsehood. If the brain making this statement lives in the "real" world, then it is not a brain in a vat. On the other hand, if the brain making this statement is really just a brain in the vat then by stating "I am a brain in a vat" what the brain is really stating is "I am what nerve stimuli have convinced me is a 'brain,' and I reside in an image that I have been convinced is called a 'vat'." That is, a brain in a vat would never be thinking about real brains or real vats, but rather about images sent into it that resemble real brains or real vats. This of course makes our definition of "real" even more muddled.
We sometimes are aware of the dream and say I'm dreaming. When we wake up from a dream, we realize that we were dreaming, that is, suddenly we are connected to a second reality, and also retain memory of previous, disconnected reality. Also, somehow, we are able to differentiate between the two.
Now consider a brain in the vat. Let's say, the real camera feed from lab (where the whole vat thing is setup), is rendered when the vat brain dreams, and is woken up to the virtual world after sleep. So even in this upside down rendered reality, there must be some thing to convince the vat brain, to make it capable of differentiate between the two. (may be one of the rendered reality deliberately be made fuzzy etc)
So, doesn't this leads to the fact that the brain in the vat can
- acknowledge two realities?
- be convinced to some extent (even wrongly) to believe one of them?
- finally, have doubts about reality?