The Wikipedia definition certainly suggests that the demand falls for the "inferior" good because it rises for the "superior" alternative. It is reasonable to observe whether demand correlates with income, and it wouldn't be too shocking to see that demand for Busch correlates negatively with income, and demand for Czechvar correlates positively.
You're right, though, that there's no such thing as undemonstrated preference. We can say that the one drinking the Czechvar preferred it to Busch, and vice versa, but that's all we can say. We can't say what anyone's second choice would be, for example. So it's a methodological error to say that, because fewer rich people prefer Busch, therefore more rich people unprefer Busch. Nor would it be right to imagine one person moving up the income scale and abandoning Busch for Czechvar, since preferences are always time dependent, individual and subjective.
Psychologically, though, we could discover that a certain rich man used to drink Busch, and stopped when he became rich. So it's understandable why people fall for the temptation of interpreting "inferior goods" in psychological terms. That's invalid from an Austrian viewpoint, but I think mainstream economists do that sort of thing all the time.
--Len