there is no unified learning theory

Assertion: There is not yet a unified learning theory of everything
Analogy: Just like there is no unified theory of physics yet
Corollary: Useful things can be learnt from different learning theories

There are a lot of different learning theories! See this page of the learning evolves wiki

I am not advocating even handedness, that all learning theories are equal. Some learning theories are better than others. My personal favourite at the moment is Andy Clark's theory of embodied active cognition.

However, we haven't reached the stage yet where a single learning theory has gobbled up and subsumed all the other learning theories and that the others no longer have anything useful to offer. It might be possible to get to that point one day but we are not there yet.

Some learning theories that are suspect at their core still seem to have some useful things to offer.

There is strong evidence that cognitivism seems to be flawed at its core in suggesting that humans process information (input / output) through hypothesised internal structures (variously called symbols, schema, frames or scripts), ie. that we are creating an elaborate internal model of the outside world. Nevertheless, IMO, supporters of this school of thought still come up with useful research, for example, the finding that hard work, focused effort (effortful study), is the most important factor in acquiring expertise. Another cognitivist, Roger Schank, has produced sublime critiques of current day School based curriculum development and very interesting alternatives based on learning by doing, goal based scenarios and story centred curricula. I feel it would be dogmatic to dismiss ideas such as these because the theoretical base of cognitivism is now under a cloud. It's better in this instance to cherry pick the ideas that seem good, try them out in practice and to keep evaluating both the theory and practice of these different approaches.

I have criticised George Siemens for developing a new learning theory, connectivism, without first thoroughly evaluating already existing learning theories. I've been particularly critical about George's understanding of constructivism and constructionism.

George has responded to this criticism as follows:

I don't agree, however, with your assessment that my ideas have shifted significantly at their core. In the article, I expressed the need for connectivism largely in reaction to the changed climate of learning and knowledge today. Knowledge is growing exponentially, people are using new tools of technology, our information is digital vs. physical, etc. Essentially, we exist in a different space today than we did 30 years ago. The manner in which we encounter information is dramatically different than it was even a decade ago. Or for that matter, with the development of blogs, 5 years ago. With connectivism, I am attempting to address a particular type of knowledge and a type of knowledge need.

So it appears that connectivism is a learning theory for "the act or state of acquiring or possessing actionable knowledge" in a new world where networked digital tools are available to interact with the knowledge explosion.

I still see a problem with connectivism as a niche theory for rapidly changing knowledge, in the context of other learning theories that have not been rigorously evaluated. Either other learning theories have something to say about this type of knowledge or they don't. The claim that connectivism offers better ways to deal with rapidly changing knowledge is not credible unless compared with up to date knowledge of alternative learning theories. For example, I would argue that embodied active cognition presents a well researched and up to date view about how knowledge is processed that is quite adequate for the modern ecology of knowledge. My point is that learning theory has continuted to evolve and that George hasn't checked it out thoroughly. It's not a good idea to create a new theory without a thorough awareness of existing theories.

Of course I can't stop George doing that. But surely a better way forward to improve our knowledge of learning would be to study and evaluate the current knowledge contained in all the many learning theories that already exist.