Diffusing Wenger

While reading Wenger, one question repeatedly arose in my mind, in several forms: Does one need to learn in order to be in a community of practice? In other words (Wenger is fond of this phrase), does practice imply learning? Can communities of practice exist where learning does not? This view is somewhat antithetical to my usual views on learning (that it is ongoing, permanent, omnipresent; that almost anything can be considered "learning"), but Wenger defines learning more in a sociohistorical sense (social theory of learning: that we learn from/through/with/because of others). Let's take Wenger's claims processing example, and say we have the same group of five claims processors who have all been there for a period of time (i.e. no newcomers) who have been working with the same forms for that time (i.e. no revisions or new software). Although the individual claims processed change, one might say these five people have fallen into a "routine," and they all know each other well enough that they have their niches in this community. What are they learning, and how? If they are not learning, are they no longer a community of practice? (I think I know how Wenger would answer these questions, but I'm curious to see how you think and whether you agree with Wenger.)