Is it a new theory of learning? Or is the concept of theory distracting?
Although it can certainly be seen in some sense as a (great) new theory of learning, the concept of a theory is, in fact, distracting.
In today’s climate of scientific certainty, the term “theory” immediately triggers the demand for empirical evidence. I don’t doubt that such evidence could be delivered some day. But in complex environments like knowing and learning (that take place behind the observable input/ output interfaces of the mind), it is very difficult to construct study scenarios that are appropriate for the objects under study, without falling prey to some bias, and without having the results being spoilt by circumstances that unconsciously perpetuate the old-style approaches. Without such careful and patient design, the study would likely end up with the usual “no significant difference” finding which eventually shows up as “There is no …” in some abridging citations (especially by authors who shy at the need for change).
However, Connectivism could very well help inform the design of such studies, because it is a powerful new view of many of the relevant aspects.
What are the weaknesses of connectivism as formulated in this course? What are the strengths?
It is such an exhaustive explanation, that this turns into another potential distraction and vulnerability: The suspicion that it wants to be yet another Grand Unifying Theory, and that it is more philosophical rather than relevant for practice. There are several dimensions where it can be loved and hated for its comprehensiveness.
1. The most superficial reproach is that it wanted to replace all previous learning theories, such as Behaviorism, Cognitivism, and Constructivism. Okay, the new theory asseverates that is has inherited a lot of useful things from its predecessors. But still there is a smack of universality. I think it would be useful to emphasize the complementary character of connectivism, and that it is needed to understand previously unexplained phenomena, and that it coexists with, and sits on top of, its predecessors.
2. The biggest issue is about the three layers Neural, Conceptual, and Social. This is also an issue where our two facilitators disagree. Siemens is content with multiple theories for the multiple levels, while Downes has a single theory for all. (At least that’s what I noted in the Friday Ustream session in week 2.)
For me, the interesting thing is the single metaphor applying to all three layers. This constitutes the principal bracket keeping everything together, and the greatest strength
Particularly the interaction between conceptual level and social/ external level connections is most fruitful to study, without an immediate need for a unifying theory.
3. The toughest part of connectivism as an exhaustive explanation, is the question of representational knowledge. As I understand it, we should not think of knowledge as an internal reproduction of an external world, both of which would possess some absolute, objective, ontological (?) reality. However, I think there are many kinds of simple knowledge where such a simple conception does not hurt.
For many cognitive operations it is just sufficient to think of a knowledge item such as “Paris is the capital of France” as a simple connection between the object of Paris and the object of France, like a database relation which sort-of “snaps in place” when the assertion is learnt, and which helps more assertions about Paris to snap in place when combined. As long as the connection strength equals 1 (certainty or with a neglectable residual doubt about this reality), then I think this special case of knowledge can lightheartedly be viewed in the traditional, non-connectivist way. After all, much public knowledge (accumulated through history, of non-buffs, and dead people) is perceived in this way by the ordinary people who use the term in ordinary language.
If we see it as a border case of connective knowledge like the rectangle is a border case of a parallelogram, then, of course, we could still insist on the dictum that a rectangle is not parallelogram, and that this simple case is not knowledge, but it is not too important, IMO. As I understood it, connectivism allows for the special case, at least it “allows that representational systems – and hence, meaning – exist” (Downes).
So, in the above three dimensions (predecessor theories, layers, and knowledge types), the impression of an exhaustive theory may be distracting; it is not necessary, and a complementary approach suffices.