Wittgenstein often refers to matters of learning, and there have been efforts to extract a social conception of learning from his writings. In the first half of this article, I look at three such efforts, those of Meredith Williams, Christopher Winch, and David Bakhurst, and I say why I think these efforts fail. As I go on to argue, though, there is a fairly trivial sense in which learning is a social rather than a psychological phenomenon: ordinarily, there are public criteria for whether someone has learned something. Yet, in the second half of the article, I point to an exception to this general rule. Taking a cue from Wittgenstein, I call this ‘intransitive learning’, as it refers to learning experiences where we cannot say what we have learned or where there simply isn't anything in particular that we have learned. This is a use that is not easily accommodated by received definitions of learning. It also represents a genuinely psychological use of the word ‘learn’. In contrast to ordinary cases of learning, claims about intransitive learning function like expressions and are marked by first-person authority.