Sunday, June 17, 2012

Connectivism as Epistemology

Responding to questions from Vance McPherson

1) What is your response to Rita Kop's suggestion that connectivism is a new epistemology but not a new learning theory?
As I understand Rita, she understands the pedagogical aspects of connectivism to have already been present in constructivism, and hence, connectivism is not proposing something new when it comes to giving guidance to instructional staff. There are overlaps to be sure, however:
- criticisms of a teaching practice, which may be grounded if working in a constructivist perspective, are not grounded in a connectivist environment. For example, I responded to criticisms from Heli Nurmi several times in this fashion.
- there is a universalist aspect to constructivism that is not present in connectivism; to be a 'theory' requires statements of general principles of teaching, and connectivism mostly doesn't have these
- and related, constructivism depends on intention in a way connectivism does not - it supposes that people are consciously building or constructing knowledge, whereas in connectivism this is not required

Connectivism is *definitively* a learning theory, or more accurately, incorporates learning theories (specifically, theories about how connections are formed in networks). It suggests some teaching theories (I have capsulized them as 'to teach is to model and demonstrate' and suggested that connectivism argues for the creation of an immersive learning environment).

But all of that said, whether connectivism is a *new* theory of epistemology or pedagogy is irrelevant to me and I don't spend any time worrying about it. I often preface my remarks with the sentence "everything I have to say has been said (often better) by someone else."

2) My understanding of connectivism is currently as both epistemology and learning theory, which presupposes that it has ALWAYS been correct and is not contingent upon modern technological developments to "work." Rather, technology casts light upon the nature of the model.  But many authors are suggesting that the application of the learning theory is primarily to technology-based learning.  What's your take on this?
It has always been correct (insofar as it has components that say anything is 'always' the case). Networks have always learned. Humans have always been a network (at least, since graduating from single-celled organisms).

Other aspects of the theory change over time. At the deepest level, the principles for stable (dynamic, learning) networks - autonomy, diversity, etc - are probably reasonably constant over different types of networks.

But the sort of environments that create learning vary greatly - the sort of environment that produces a modern information-age knowledge-worker for example varies greatly from one which would produce a skilled bow-maker in the middle ages.

Additionally, with the development of technology, new types of networks have come into being. While human commerce has always formed a network, it has been a relatively simple network and certainly a slow-moving one. There's only so much connecting that can take place via personal communication and the Royal Mail. Technology greatly accelerated the size and speed of the human network, producing in a way not previously possible more observable properties of a network (for example, cascade phenomena as an idea or meme propagates through the network).
3) M. K. Dunaway (2011) recently published a paper in Reference Services Review, where she describes connectivism as claiming that knowledge emerges from an individual's learning network as connections are recognized.  If I'm understanding your position correctly, then Dunaway's description is inaccurate in placing the locus of knowledge with the individual learner's recognition patterns, and not in the network itself.  But I may also have not correctly understood you, her, or both.  Could you steer me straight on this?
I would have to read Dunaway to be able to provide a reasonable response - if you could link me to a copy that would be helpful.

That said - 'recognition' is a core thesis of my own theory. To 'know' x is to be capable of *recognizing* 'x'. To recognize 'x' is to assume an appropriate neural configuration when presented with an 'x', where 'appropriate' may be described in any variety of manners. I sometimes talk of 'knowing' 'x' to having the right 'feeling' when represented with 'x', a feeling of recognition. To 'recognize' is a property of a successful network.

Additionally, networks exhibit patterns or regularities. For example, a weather network may exhibit a characteristic 'storm front' or a mumuration of blackbirds may display shapes in the air. In my own work I often use examples like 'facves on a TV screen' or on the surface of Mars. These patterns in a network are phenomena that exist *only* as things that are recognized. To say a pattern 'x' exists in network 'y' requires a perceiver 'P' presented with 'y' and who instantiates an appropriate network state (a 'familiar feeling', a 'habitual reaction', a 'recognition') when presented with (a perspective of) 'x'.
4) I sometimes get the impression that you and George Siemens are not exactly on the same page when it comes to the epistemological aspects of connectivism, which of course would be perfectly fine in the context of a dialogic process, but I wondered if you'd care to comment on this.

George and I have our debates. My sense is that is is much more concerned with the pedagogical aspects of connectivism while I am much more interested in the epistemological aspects. Philosophically, George is a realist while I am more of an idealist - that is to say, he is more likely to say the phenomena we observe (be they chairs or colours or shapes and movements) are 'real' while I (for reasons just stated) say they require a perceiver.
One more thing, something of a comment.  You've described semiotic processes (language, symbols) as epiphenomena of networks, but not essential to them.  This reminds me a lot of Stephen Jay Gould's idea of "spandrels."  I thought it was interesting because one of Bill Kerr's beefs with connectivism seems to be that there is not a good evolutionary / biological explanation for how connectivism is possible.  But I think that, on the contrary, connectivism, if correct, would prove conclusively Gould's spandrel hypothesis, which is widely accepted in evolutionary biology circles.  Just a thought.
I have described the patterns we perceive as supervenient on the phenomena that produce them. So that does make them epiphenomenal in a way.

The whole question of an evolutionary basis for connectivism is one I have not considered. But I think there's a good basis for such an argument. A network is at heart a recognition system; it responds in consistent ways to complex and variable phenomena. It embodies the capacity to adapt to change. The more complex an environment the more likely that a network, rather than a simple innate instinct, would ensure survival.

A language I think emerges quite naturally out of this. Given that humans have the capacity to make noises and gestures, and that these would be consistently produced given certain phenomena, it would not be long before the adaptive advantage of communication ensured its adoption. Most - if not all - of actual language is (in my mind) learned. But there is no question that the networks we are born with at birth are sensitive to the sounds and movements made by people like ourselves.

That said: language (as an entity) is a *social* phenomenon, not a personal phenomenon. Language is stigmergic. As Wittgenstein would say, there is no private language. Not because of some 'private language argument' (I think this is a recreation of Wittgenstein's thought after the fact, and not core to what Wittgenstein had to say) but because the properties of language - specific words (the associated sounds and symbols, and conventional meaning or reference), grammars and syntax, works in literature and art that constitute paradigms, etc. - are physical phenomena, present out there in the world and not in the humans that speak and write it.

Is language a spandral - an accidental artifact of evolution? In one sense no - I think a look at language after the fact shows how important it has been to survival. But in another sense no - it's not an artifact of evolution at all, as it is not a property of individual humans.

But should investigation show a particular innate sensitivity to some aspect of language - a 'mirror neuron for syntax', say, that might be a spandral. That might be a selected preference for a particular aspect of language that *could* have been different (you could have an equally effective language without it) but was the way it was, and was selected for. It might show up in the way, say, an innate preference for the colour red might have - as an aid to identifying dangerous stuff in the world, which in an alternative history could well have been blue or green (think Vulcan) or whatever.

Good questions, interesting discussion, thanks. I will post these to my weblog, if you don't mind.

-- Stephen