Wittgenstein on the Argument From Analogy

The argument from analogy for the existence of other minds says that I know the connection between my observable traits (physical and behavioral) and my (unobservable) mental state, and thus learn to infer from the observed behavior of another person to a further mental attribution. The reason many including myself consider Wittgenstein to be at least a close relative of behaviorism is his argument that the referents of psychological predicates are necessarily public in the first place. So Wittgenstein rejects the premise that there are unobservable mental things, or at least that these (what many call phenomenal properties) can be spoken about (handled by public language) meaningfully. This is different from the argument that the analogy doesn't work because it is an inference from the smallest possible data set (oneself). That is an argument, but it is not a Wittgensteinian one: on the probability argument, I turn out to lack logical demonstration of something (other minds) that I believe. That is not the same as showing that I cannot reasonably believe in other minds (noncognitivist psychology!). On the Wittgensteinian argument, by contrast, I can neither believe nor disbelieve in other minds, or at least nothing can be said about other minds (or my own) outside of some pragmatic, public context.