Sometimes, theoreticians and practitioners alike get into idealisation of a concept. And this is a time in which, sadly, I see it happening to CoPs.
How? Let's see some examples....
One of the most common is, over-reliance on Wenger's definitions , even when nebolous and/or incomplete, as pointed out by Gourlay (though his critique was one-sided on principle it still has some huge merits). Some go far beyond this, thinking that, if Wenger didn't theorize/describe it, then what was added from, by and to the "field" is false/inaccurate.
Another risk I see is overtheorization. Some folks pile definitions on top of definitions and what is said to be informal, dealing with implicit knowledge and not enforceable begins to have more rules and definitions than a Catholic Catechism.
Frequently, you can stumble into nonsense classifications, with people splitting hairs into definitions that aren't operational at all and, often, are all on the same level rather than being branched. The net result of it is a long long list of names, usually "community of something", that aren't characterized well enough to be useful and/or aren't ordered in a hierarchial taxonomy to be applied.
Most frequent thing you can see is the lack of and/or allergy to methods, and by that I mean any method. Now, there is a difference (I hope) between a CoP and a mess, and I'd like to see it characterized more often. Methods for both action learning/research and participatory research were hard to find and get defined, yet having a codified method gave dignity to the discipline and (especially) its findings. Wondering why academics "snob" the field...?
Another problem is the lack of conceptual definitions that could help divulge the findings of the practice. Anedoctal stories could be studied and theorized on much better in presence of some unifying concepts (which by the way would help communication, since it happens through language, which in turn is impossible without codified concepts). Still wondering why academics "snob" the field...?
A huge, almost unexplored part, is the way of "nurturing" a CoP climate, how it works, why, etc., which I believe to be a priority, if we want CoP implementation to be a widespread successful practice. Not addressing THIS part means, de facto, preventing others from reproducing CoP implementation, which means none of our findings can be tested. Again... still wondering why academincs "snob" the field...?
And finally (this is for all, practitioners and theoreticians) there is no one-size-fits-all: even CoPs are inadequate solutions for some situations.