George Siemens original article: Connectivism: A Learning Theory for the Digital Age
George Siemens blog, wiki, discussion
George Siemens book: Knowing Knowledge pdf, wiki, review (Dave Pollard)

Connectivism: Learning Conceptualized Through the Lens of Today's World (conference abstract)

By George Siemens
Knowledge and learning are forefront in the progress and advancement of humanity. At no other time in history have we stood before as rich a panorama of opportunity as we do today. Our access to information, knowledge, global conversation, research, and the experiences of generations past provide a firm foundation on which to build the society of tomorrow. Yet openness, abundance, and access raise new concerns. The ability to cope with today's knowledge deluge, to engage learners in co-creation of content, and to enlarge classroom walls to include diverse perspectives requires a new conception of learning. Theories of cognition and learning that have served well in the past seem frail, ineffective, and out of touch with the reality of learners and the new context and characteristics of knowledge today.


A challenge to connectivism (conference paper outline by Bill Kerr - separate page)

situating connectivism: relation to existing theories of learning and knowing (part 1)
recent dialogue with George Siemens
Knowing knowledge extracts, with comments by BK interspersed
my intital impressions of connectivism were postive (BK)
Some comments on Plon Verhagen's critique and George reply
Network is not god (critique of George Seimen's analysis of constructivism)


recent dialogue with George Siemens

Recent comments by George clarifying his views:
"stealing from Wittegenstein's language games - context games are our key challenge in dialogue and learning" (KK reference)
"learning is the act or state of acquiring or possessing actionable knowledge..."

"the work Stephen Downes has down with connective knowledge provides the knowledge-based construct of connectivism (i.e. objectivism, pragmativism, interpretivism - where knowledge resides and how we come to acquire it)" - philosophical base
(George Siemens)
dialogue with George at dialogue online thread
(Bill 7Jan07)


Knowing Knowledge extracts:

My comments interspersed in italics - Bill
page 59
We are a bricolage of cognition, emotion, intuition, information consumption, doubt, and belief

The adoption of a particular belief and mindset will hold within it certain accompanying views and logical developments which are less a function of reasoning and logic and more a function of the space we have decided to enter (see Figure 26). The ideas contain in themselves a germinating process.

Comment: Rather than "less reasoning" and "more space" I'd see this as a dialectical relationship between logical reasoning and environment ("space"). In some cases the logic might be dominant, in other cases the environment might be. Example: It is possible for a teacher who spends each day in a classroom setting to think and plan for the overthrow of classroom based environments. But it is much harder for that teacher to consistently stick to that stance, than for a School critic who is outside the classroom, because the teacher is under daily pressure to make that classroom environment work.

The true point of knowing occurs at the stage of creating or adopting an ideology or world view. Once adopted, this view serves as a filter and cognitive off-load tool. If our ideology is strongly empirical, we primarily function within this conceptual domain (though we will hold contradictory view points at times—due to context-games that influence what we think and say at various times). The ideology strongly influences the conclusion. The outcome is not only set in our daily activities, but in the process of selecting a world view.

Comment: Knowing or deepening knowledge occurs through a theory / practice cycle. Practice may impact back onto a world view and change it. To describe ideology as the "true point of knowing" will end up in dogmatism or too many generalisations not linked to practice, which is what I think George is doing

CONTEXT GAMES (pp. 61- )

…we are not about logic
we are not only about our drives and desires…
We are about Context Games.

Our desires and logic are shaped in an orchestra of context: acting and reacting, negotiating and dialoguing.

Comment: the statement, "we are not about logic" is different to "logic (is) shaped in an orchestra of context" ... I'd see the first statement as wrong, it's not a good idea to abandon logic

To define context is to frame the solution.
Context is not as simple as being in a different space…context includes elements like our emotions, recent experiences, beliefs, and the surrounding environment—each element possesses attributes, that when considered in a certain light, informs what is possible in the discussion. The object is tied to the nature of the discussion (framework [or network] of thought). The context-game is the formulation and negotiation of what will be permissible, valued, and the standards to which we will appeal in situations of dispute. The context-game of implementing a new corporate strategy involves individuals, politics, permissible ways of seeing and perceiving, recent events, corporate history, and a multitude of other factors.

Context games are the attempt to clarify and highlight factors that impact our comprehension of a particular situation.

Consider two individuals engaging in a discussion of liberal and conservative politics. The real discussion is not about a particular political issue (for example, the degree to which the government should be involved in social programs). Instead, the real discussion centers on each party attempting to project their deeper views (based on the “pathway” model provided in Figure 26)—namely that by adopting a view, we often walk to its logical outcome.) We do not engage situations neutrally. We engage them based on the manner in which we have crafted our logic or how we have entered the corridors of logic. We do not evaluate a thing only for “what it is.” We evaluate it for how it relates to our defined views and ideologies.

Context games include:
1. What we bring: Our existing viewpoints/ideologies
2. What impacts: The factors that exist and impact the discussion/.knowledge (recent events, news items, networks nodes from which we perceive)
3. What exists: The nature of the topic—it melds with the context and causes ripples of change within the context itself.
4. Space of occurrence: The environment/culture/zeitgeist in which the dialogue/debate occurs
5. Who is involved: Parties with whom we are familiar shape the context; we fill in missing elements based on previous encounters
6. What we possess: Oratory or charismatic traits of the participant
7. What we feel: Emotions
8. What we communicate: The attempt to convey to others the validity of each perspective
9. How we negotiate: How we determine measures of validity and acceptable context (requires give and take)
10. What is the domain, type, state, and level of knowledge?
11. How we debate: The points of logic, emotion, inclusion/exclusion
12. Context breakdown (and archiving for future similar experiences)

Comment: I don't think this is new. It's similar to the notion of situated learning, that learning as it normally occurs is a function of the activity, context and culture in which it occurs (Lave) - again there is not much by way of examples

In a sense, the key area seen as the surface notion of the debate (in our previous example of government involvement in social programs) is not really an issue. It is an opening through which we can express our larger views.

Comment: But often it is an issue. Sometimes the particular is important, at other times, larger issues are important

With regard to knowledge and learning, context influences our capacity to convey our thoughts. If knowledge has been hardened into ideologies, or if new knowledge is seen through ideologies crafted in advance, the outcome of the discussion is essentially set. Debate is largely an attempt to project world views.

Of what value is the act of debate? Debate provides an additional dimension to context that enables individuals to see entities beyond their own worldviews. We categorize and box individuals. This presupposes how they think and act. By understanding context games, we are better able to suspend hard proclamations in advance of understanding the particular concept being expressed by others. If both parties acknowledge context creation activities, the capacity to agree on a particular framework of debate and inclusion of perspectives that may challenge our established ideologies, may be increased. If our debate is less about projecting our world views, and more about exploring what is actually being said, we open our minds to reception of knowledge that is filtered by our opinions.

Comment: Agree with this positive view of debate

We value what is different more than what is known…it pulls on logic toward non-logic directions.

Comment: Another one sided proclamation, sometime we value what is know more than what is different

Existing mental models are not loose enough to allow for new structure to emerge. Mental models (like schema) assume that we are logical and structured in our exposure to knowledge. We are not always logical.

Comment: The first sentence is an over generalisation, that all existing mental models are deficient
Critical comment about schema, is interesting -- explore this more -- this criticism would only apply to cognitivist theories

Schema are similar to framing , eg. "tax relief" - the word "relief" is used to frame tax relief as a liberation from an affliction imposed by the other political party

We are contextually holistic. We act consistent with how we have framed and determined our world. We filter out information we feel is
not important. Our behavior is consistent with our context, though we may at times violate our actions of the past. We do not exclusively subsume, accommodate, or assimilate. We place new knowledge in relation to other knowledge. If similarities exist or revelations occur, the element is connected to our neural structure. .

We connect more than we construct.

Comment: ? not explained in the book


my intitial impressions of connectivism were positive

pipe more important than the contents for mainly a postive interpretation of connectivism (BK)

"the skin is not all that important as a boundary" BF Skinner(1)
The notorious Skinner got that one right. The boundary issue is crucial. In considering the learning process we need to ask: What happens inside our body / brain, what happens outside, in the external environment, and how are the inside and the outside connected? What is the mind, where is it and how does it work? These are core theoretical questions about learning with immense practical significance. The necessary process of formulating a new learning theory ought to incorporate and struggle with a modern synthesis of philosophy, cognitive science (including artifical intelligence research) and the history of learning theory. My critique of George Siemen's Connectivism suggests that a better job could have been done.

(1) Cited in Brainstorms: Philosophical Essays on Mind and Psychology by Daniel Dennett, (1981) Ch 4. Skinner Skinned, p. 55. The original quote is from BF Skinner, "Behaviorism at Fifty". In "Behaviorism and Phenomenology " (1964) Edited by T.W. Wann.

A challenge to connectivism
(old summary of my conference paper - Bill Kerr 26Dec06)

Networks are important but haven't changed learning so much that we need to throw away all of the established learning theories and replace them with a brand new one. How do we test whether a new idea is an interesting speculation or something more substantial? A good learning theory should:
  1. contribute to a theory/practice spiral of curriculum / learning reform,
  2. provide a significant new perspective about how we see learning happening
  3. represent historical alternatives accurately.
Connectivism fails on the first count by using language and slogans that are sometimes “correct” but are too generalised to guide new practice at the level of how learning actually happens.

Connectivisim does contribute to a general world outlook but we already have theories and manifestos for that view (systems theory, chaos theory, network theory, cluetrain manifesto), so we don't need a new -ism in this respect.

Finally, connectivism misrepresents the current state of established alternative learning theories such as constructivism, behaviourism and cognitivism, so this basis for a new theory is also dubious.



1) "... this is not a learning theory, but a pedagogical view on education" (Plon Verhagen)

critique of Siemens theory by Prof. dr. ir. Pløn W. Verhagen, professor Educational Design at the faculty of Behavioural Sciences at the University of Twente
"... this is not a learning theory, but a pedagogical view on education ... The questions that Siemens presents are not to be placed at the instructional level, but at the level of the curriculum. The instructional level deals with how learning takes place, and learning theories are relevant at that level. The level of the curriculum is concerned with what is learned and why ..."

My comment:
Against Verhagen: If you teach learning how to learn skills (eg. how to use google advanced search features) then I think you can argue that you are giving the student the tools to devise their own curriculum (eg. to search for the things of interest to them). Learning how to learn skills integrate the how with the what so the division might not be always that clear as is suggested here.
Against Siemens: However, it could also be argued that if students are just taught "web2.0" tools that they will spontaneously learn how to put games and stories on the web but they won't learn calculus in that way. Alan Kay has a list of non universals that need to be taught explicitly because they won't develop spontaneously. Learning how to use MySpace is not on that list but is also not higher order thinking. So calculus needs to be in the curriculum and a learning theory that informs how to teach calculus needs to be there too. Connectivism doesn't help here.

"A theory should explain phenomena and those explanations should be verifiable ..."
(goes onto say that Siemen's theory is not "sufficiently specific and coherent" to meet this criteria)

In the final section Verhagen critiques Siemens view that "learning may reside in non-human appliances"
Although forms of intelligence (information processing) do exist outside humans and are becoming more important this is not actually new and does not provide a basis for a new learning theory

Learning theory or Pastime for the Self-Amused (Siemens)

Siemens response to the Verhagen critique. This seems to be a restating and updating of Siemen's theory and does not actually respond directly to Verhagen's critique. George seems to be saying that it is not fair that Verhagen focuses his critique on the original 2004 paper because connectivism is a moving target, a dynamic theory, that whenever a new web app comes on line then we become more connected and this is some sort of proof of the correctness of connectivism!! (paras 6,7 and 8).

I don't think this is the way to conduct a proper discussion. George should have addressed the particular criticisms from Verhagen in a shorter and more direct manner. The changes we have seen since 2004 are extrapolations of trends that were clear then. The fact that we now have YouTube and we didn't have it in 2004 doesn't prove anything new at a theoretically level IMO. I would see this as more of a quantitative than qualitative change.

What is fundamental and what is fad?

Alan Kay says that computer science has not developed much in thirty years, that LISP and Smalltalk are currently as good as it gets.
David Wheeler says that the extent of software innovation is over-rated, that we shouldn't confuse faster computers and more storage arising from Moores Law, with innovation, actual new ideas.

In response to Verhagen, George says this:
"The error made in the review is precisely the reason why we need to explore connectivism as a learning theory: static, context-less, content-centric approaches to knowing and understanding are fraught with likelihood of misunderstanding ...This is the danger of product iconization as offered, or explored by prominent theories of learning, thus failing to acknowledge - explicitly - that ongoing changes obsolesce current knowledge" (para 8)

This is an argument that everything is in flux and that we ought to challenge and question everything. ie connectivism is better because it is new. George has to do more work than that and I think his new version is basically a rehash of the old version.


2) "the network is not god" (Bill Kerr)

constructivism vs. connectivism
A rather disappointing analysis by George Siemens (connectivism theory) on the alleged deficiencies of constructivism.
I've said some positive things about George Siemen's new connectivism theory in the past (pipe more important than the contents) but this time around I'm disappointed.

My summary of an earlier interview of George by Teemu Arina was:
"I agree with the general thesis that the pipe, being connected, is becoming more important than the content. I agree that the half life of knowledge is declining and that more and more learning is informal. These changes are corroding schools. Students are different from before and bored with lecture mode. Nevertheless, I'd see the theory of connectivism as sitting alongside the other learning theories, not taking their place."

It seems that others have asked George to think about how connectivism "connects" to constructivism and this has led to this analysis, constructivism vs. connectivism George leads off with this:

"Constructivism, as a model of learning, holds the duality of much promise, and much frustration. On the one hand, it breaks from the structured models of learning that dominated the first half of the last century, giving voice to the "softer" elements of learning (educators often understand this intuitively - we see the lack of direct connection between what we lecture about and what our students actually learn). On the other hand, constructivism has not been well-defined. It can essentially mean anything to anyone. It's an idea without boundaries, a philosophy without root. This vague definition results in everything being labeled as constructivism (see these six paradigms). If anything, my experience with constructivism places it more in the domain of a teaching philosophy, and less in the domain of a theory (consider these attributes of constructivism)." [Emphasis added]

This is too vague, creating a situation where George does not have to do much work in either integrating or refuting constructivism because, after all, it doesn't mean anything. Although it is true that some constructivists are philosophical idealists (that meanings constructed inside the head do not necessarily reflect real things in the external world) it is still quite possible to use a common sense, down to earth interpretation of constructivism. Two points from Piaget stand out:
  • Children build or construct their own intellectual structures
  • Children build on what they know. Piaget's term for children's continual balancing of existing cognitive structures with new experiences is equilibration.

George continues:
"Several individuals have provided excellent guidance in suggesting that I don't try and position connectivism as a replacement for established learning theories (i.e. constructivism, behaviourism, cognitivism). I'm generally supportive of integral thinking, and agree with a matrix posted by Derek Wenmoth on online learning (including a continuum of learning theories)."

George doesn't follow his own advice here. From his article I get a strong feeling that he hasn't read much about constructivism or tried to implement it in practice or even thought much about it. I'm critical of this arrogant attitude, developing a "brand new" learning theory in a vacuum, but without doing the hard work of analysing the good and bad points of existing theories. Theoretical progress does not work that way.

George again:
"Constructivism, for me, fails on two levels: 1) it is not capable of functioning in rapid knowledge growth environments, as it doesn't account for learning that happens in networks and 2) constructivism is a "sometimes" learning habit (we are always connecting, but we only construct in certain situations)."

As a critique this doesn't amount to much. It is just assertion. George has discovered that the network is good, the dominant paradigm of our new age. Network good. Construction not so good. But there is no actual argument here.

"Constructivism, as with other learning theories, assumes that learning happens in our head. In fairness, various flavours of constructivism acknowledge the importance of the social context in which the learning happens, and that learners learn from each other. The act of learning itself is still perceived to be in the head of the individual. Most learning needs today are becoming too complex to be addressed in "our heads". We need to rely on a network of people (and increasingly, technology) to store, access, and retrieve knowledge and motivate its use. The network itself becomes the learning. This is critical today; the rapid development of knowledge means that we need to find new ways of learning and staying current. We cannot increase our capacity for learning ad infinitum. We must begin to conceive learning as socially networked and enhanced by technology (it’s a symbiosis of people and technology that forms our learning networks). We need to acknowledge our learning context not only as an enabler of learning, but as a participant of the learning itself." [Emphasis added]

external image head.gif
In what way am I being perverse to claim that learning still happens inside our head? Richard Dawkins once said that there is such a thing as becoming so open minded that your brains fall out. George seems to be falling into a similar trap, becoming so enamoured with the power of the network, to the point of denying the importance of the individual and the learning that occurs inside "our heads". These thoughts of mine were gathered from various sources, synthesised in my head and then put onto the network. I agree that the network is far more important than it used to be but that we need to do more work in figuring out the correct balance between the learning that occurs inside and outside individuals and at the boundary structures. Radical individualism, the role of a single individual disagreeing, going against the tide is still as important as ever. How do we explain this by network learning theory, aka connectivism? At this point I looked up some quotes from the digital maoism debate and decided not to use them. As far as I am aware I made that decison in my head.

George continues:
"Constructivism is complex. Let your mind wander a bit: My learning is a function of previous life experience, the people around me, the actual environment in which I function, my previous learning experiences (both emotional and cognitive), the nature of group relationships (socially-based), etc. When new information enters the space, I (according to constructivism) construct knowledge of its meaning/relevance against the backdrop of the above mentioned factors. But I can't simply construct - because, to use the molecule metaphor of learning objects (or microcontent), many of the elements that comprise the base of my knowledge come previously constructed (by a discipline, the teacher, the article, etc.). For example, the elements that comprise a new idea come "chunked". I don't construct that entire concept or idea. Instead, I connect it with existing knowledge. If anything, the learning suggested by constructivism is actually in the deconstruction of these packaged elements into smaller pieces of knowledge. A simple example: if someone teaches me the skills of critical thinking, I will largely acquire the elements in "pre-constructed" formats. I will acknowledge that I need to question and validate knowledge sources for authenticity (a concept which can take a lifetime to integrate into practice and habits, and even then I'll still make mistakes). I don't construct anything to make use of this at a basic level. I simply adopt it and try and interrogate new information. My actual learning happens when I deconstruct the knowledge itself (getting deeper into the full meaning of the notion of "validating"). We don't always construct. We are often much more passive in our learning. We read an article and we link it to our existing understanding. We subscribe to a newsletter (or magazine)...we attend certain conferences...we dialogue with certain people/communities. In the end, much of our learning is a connection-forming process (the conduit, not content, is what is king) where we add new elements that augment our capacity to know more. We rely on Google, libraries, friends, social bookmarks/tags, etc. to serve as our personal learning network (we store the knowledge external to ourselves). When we need something, we go to our network (know-where is more important than know-how or know-what)...or we expand our network. In the end, the constant act of connecting in order to stay current is a much more reflective model of learning than constructivism."

The language not used in this long quote illustrates George's paucity of knowledge of some of the main theoreticians and practitioners of constructivism / constructionism. For instance these words are not used: equilibration, assimilation, accomodation (Piaget), constructionism (with a N, Papert), Artificial Intelligence (AI), Society of Mind (Minsky)

The idea that process of deconstruction is some sort of refutation of constructivism is ridiculous. As stated above constructivism is the view that "children build or construct their own intellectual structures." This of course includes intellectual structures that are used for deconstruction. Minsky describes this as reformulating (Society of Mind, 12.2). AI theoreticians have important things to say about learning

I still think that George has come up with some important ideas and insights about learning and the importance of networks in the way we view the world today. However, it simple is not good enough to reject the heritage of constuctivism and for that matter, behaviourism (dennett's creatures, behavourism and the inner environment) in the superficial manner that has been attempted in his constructivism vs. connectivism analysis

Finally, there are some thoughtful comments on constructivism vs. connectivism that have not been responded to. I thought the comments by Christian Spannagel (connecting and constructing are not different processes), Richard Giroday (there is a time and place for each learning theory) and tanbob (activity theory / Engestrom addresses the concerns raised by George) were all thought provoking and deserving of a response.