A naive person might think that [X] is an absolute good thing. Someone smarter than that naive person might realize that [Y] is a strong negative to [X] and desperately needs to be stopped. Someone even smarter than that, to differentiate emself from the second person, might decide [Y] wasn't such a big deal after all.He also gives some examples, like "don't care about Africa / give aid to Africa / don't give aid to Africa", where the third position is the intellectual hipster.
His idea is pretty similar to my beloved conformity theory. One difference is that I conceptualize opinions moving around a circle (i.e. with two axes) whereas he seems to view them as oscillating between two poles.
To me, it doesn't make sense that beliefs would just jump from one point to another. At least subconciously, there has to be some kind of intermediate stance, like "I'm not sure, but I don't consider this issue interesting." The second axes of "caring" or "interestingness" allows the individual to justify holding the belief at every given time point. Otherwise, belief changing would have to happen on a threshold, and there are no thresholds.