Matthias Melcher (I think that’s his name – had to dig around his x28 blog for a bit) provides an interesting commentary on why he feels connectivism should not be seen as a learning theory. I’ll quote it at length:
IMO, a definition or description would be more appropriate for simpler things that do not suffer when they are isolated and formalized. I think, a complex, emerging concept like connectivism is better understood by its relationships. So, rather than “What is…”, I would prefer something like “How is it related”, or connected, to other ideas, or even, to the world.
Connectivism would, IMO, suffer from restricting definitions such as being a learning theory, which has to obey traditional criteria of an empirically provable but very narrow scope of application. Even though the theory is addressing extensive changes and emancipation, this will not increase the perceived scope of what the theory explains but, instead, the prevailing resistance against such changes will further diminish and restrict the conceded scope.
The whole new view, however, that is enabled by connectivism, extends to much more than learning and schools. Downes’ and Siemens’ discussions shed new light on fundamental concepts, such as rules versus patterns, complicated vs. complex, equivalence vs. similarity, and coping with ambiguity and uncertainty. And these consideration render many entrenched practices of the entire knowledge industry questionable.