maybe it is sound, the description says a different thing. ummmmmmmmm what could help me think about this is figuring out some events and hypotheses say we consider P(noun is noun) or something say we try "i nouned the noun" and got 50% that's one experience then we try "i nouned the verb" and got 0% then we try "i verbed the noun" and got 100% we could say a 100% result implies the classes of the words are all correct and a 0% result implies the classes of the words are all wrong 3 events correspond to 4 different states of experience that could be considered --- thinking about the perception of P(noun is noun) from "i nouned the noun" being 50% . to make this strictly correctly i'd want to consider a population or sample of something