Thursday, November 05, 2015

What Else can Work at Scale? and Techniques from Social Media

My contribution to the Networked Learning Conference 'Hot Seat' discussion. You can read the whole discussion here.

# # #

Anything that can be automated works at scale. Anything that cannot be automated must be distributed, and if so, works at scale. Only centralized non-automated things do not work at scale.

# # #

This is a good question:

"Humans scale, and have scaled before any automation existed, so what made us scale? Localized transmission of knowledge (oral tradition), the automated cell/gene replication in combination with adaptations? "

And I think that the answer is inherent in the nature of life: that we are autonomous, and interactive, so we create a distributed network of diverse activities, adapting to local conditions, and scaling naturally.

Life seeks conditions of success. Humans, crickets, birds, plants - we migrate to the places where we flourish and avoid the places we don't, each making our decisions one by one.

Too dense a network and society fails. Too sparse a network and society fails. Autonomy is productive; eliminate it and society fails. But where autonomy is extended to point where it disrupts the network, society fails. (These aren't truisms; they are empirical observations, and subject to verification.)

# # #

Also, comments from the thread titled 'Educational Designers and Techniques from Social Media'.

I have argued that the design for MOOCs should take more from games than from social media (though there are some pretty strong overlaps), including in a talk just last week3. My point was that instead of trying to design learning, which is focused on content, we should create environments in which people can practice.

Social media is a bit like an environment. It is a space (mostly) not bounded by structured presentation of material or decision trees (Facebook's stream is an oft-criticized exception). People are able to try out new ideas and new personals. The problem with social media is that the interaction is (mostly) limited to conversation. I would much rather see people interact by solving problems, figuring out puzzles, playing games, and creating things.

# # #

Again, I have not clearly stated my point.

I use games as a metaphor to talk about MOOCs:

There are two types of games:
- those that depend on programmed design and memorization, and
- that create an environment.where players and objects interact

In the same way, there are two types of MOOC:
- those that depend on programmed design and memorization - xMOOC
- that create an environment.where participants and objects interact - cMOOC

The first type of game was a failure. They could be defeated by mere memorization and were not interesting. They disappeared from the market.

The second type of game was a success, and should be used as a model for MOOCs (and indeed, were a part of the model George and I used when we developed cMOOCs).

So this second type of games is the type of games I am talking about.

When comparing this second type of games and social networks, I agree with you that there are many elements in common. They are both environments, they are based on the interaction between participants, and they can be used to solve problems, negotiate and communicate.

But there are also some important differences:
  • games are inherently about solving problems or responding to challenges, while social networks can be much more passive.
  • games typically involve a wide range of different types of objects (even objects in the physical world) while social networks involve conversational elements only.
This not to say that we must choose between either games or social networks. Both inform the theory of environmentally-based learning, where participants interact in a common space with objects and with each other.

But it is to say that a model based on social networks alone will be insufficient to inform the design of successful MOOCs. The elements of a successful networking environment need to be taken into account.

Because, yes, the connections are of the utmost importance. We cannot learn from each other without connections.

But the manner, organization and structure of those connections must be designed with the intent of creating the most interesting and accessible environment. People will learn from each other, not from the MOOC.

# # #

I don't agree with this: "It's important what the discussion is about, what the goal of the discussion is..." I think there is too much desire on the part of educators to shape the learning of individuals. We should see our function as more supportive than directive.

# # #

You say: "The content - or the concrete problem - should be the center in dialogue for learning." And yet: "not that this should be determined by the instructor/teacher/facilitator."

Two things:
  • First, I don't think both things can be true. If you are going to say something should be the case in learning, then you cannot say that it should not be determined by the instructor/teacher/facilitator.
  • Second, there are many cases wher this need not be the case. Where someone is learning merely for pleasure, for example (which explains how I acquired a knowledge of the Roman Empire). Or where different people are working on different problems, not a common problem.


Analysis and Support of MOOCs

My contributions to the Network Learning 'Hot Seat' addressing the question, "How do we analyse and support the networked interactions of thousands of people learning together?" View the whole discussion here.

# # #

These are two very different questions and I'm not sure it's useful (or even appropriate) to ask them together.

With respect to the analysis of the networked interactions, it's not clear to me that we should even be doing this, at least, not in the sense suggested by the question. If someone had said "how do we analyze the conversations of university students one to another" or "how do we analyze the thousands of phone calls between people" we would quite rightly question the breach of confidence required to undertake such a task, beyond very gross calculations about the numbers of calls and conversations (and even then, we're treading on dangerous grounds). One of the differences between mass courses (at least the way I see them) and more traditional forms of online learning is that students discussions are not automatically located in the LMS not subject to ownership and use by the institution.

I'll leave discussion of the second point until later.

# # #

@jonathanbishop I would be more inclined to say that Etienne Wenger was drawing some the same foundational concepts as George Siemens and I, though I was aware of Wenger's work as far bacl as 2004 (you can hear me mispronounce his name through this entire presentation, the first I ever recorded, from 2004 )

The ideas of connectivism, communities of practice, etc., are derived from a network view of the world, and are rooted in ideas of network theory, 'small pieces loosely joined', the computational theory of connectionism, and the work of people like Francisco Varela, Paul and Patricia Churchland, Stephen Kosslyn and J.J. Gibson.

It has been occasionally suggested that MOOCs have their origin in the work of Ivan Illich. I'm sympathetic with his view, but his work has no influence on the development and design of either connectivism nor the MOOC. If people must credit someone writing specifically in the educational domain, then I would say I was far more influenced by John Holt's How Children Fail. But for me, the origins are from the scientific and philosophical domain, not from educational theorists or social scientists.
# # #
No it isn't. Wellman is interested in the formation and definition of community, and social network theory. This work is interesting, but is not the same as connectivism, which is a theory about knowledge and learning.

What you'll find in connectivism, but not in Wellman:

- 'to know' is to be organized in a certain way, as in a set of connections (such that 'to know x' is to recognize x)
- 'to learn' is to to grow, adapt and change that organizations, ie., to grow and shape connections

Contrast with Wellman, who when he talks about knowledge, talks in information-theoretic terms, such as 'knowledge transfer' or 'knowledge mobilization'. He depicts networks as conduits of knowledge, rather than as instances of knowledge.

Finally, though I am indeed the person who coined and described e-learning 2.0 (based on the earlier concept of web 2.0) I am not trying to "sell" it, or MOOCs, or connectivism. People can decide for themselves, based on the evidence, whether any of these concepts has any merit.

# # #

I see MOOCs as "Closed Educational Resources" for the reason that they
are stand alone, unlike say SCORM packaged learning objects.
This may be true for xMOOCs. It is certainly not true for the cMOOCs George Siemens and I created. Indeed, you could attend one of our MOOCs from beginning to end without even visiting the platform, solely through the use of syndication technologies like RSS. Moreover, our MOOCs were not in one place, like an xMOOC, but were distributed across a web of connections linking students resources, OERs from multiple locations, groups and conferences, and more.


Role for Educators in MOOCs

My comments to the Networked Learning Conference 'hot seat' on the role of educators in MOOCs. Read the full discussion here. 

# # #

At the risk of self-reference, I'd like to contribute to this discussion by pointing to something I wrote on the role of the educator a few years ago - The Role of the Educator - which was based on a few talks, including We Don't Need No Educator.

The point of these presentations is to say two things: first, that as the learning environment is reshaped by technology, it doesn't make sense to think of it as being provided by a single artisan manually performing a wide (and increasing) range of professional tasks; and second, that the role of the educator(s) is far more varied than one might think, even in the age of the MOOC.

After all, it's not as though the one task of the educator was to present content, and it's not as though the educator disappears once a person begins to be able to manage their own learning and find and view resources for themselves. Rather, it means that the same number of professionals (probably more) can now begin to offer specialized services to a much larger number of people.
# # #
@hsp writes, "I have no issue with social constructivism per se, but I do have an issue with the relative and subjective knowledge that it can generate. It leaves room for groups of people to come to different conclusions based on the same facts."

We've digressed a lot from the topic of 'the role for educators' but in an interesting and necessary way. And indeed, many of my discussions of MOOCs begin with a discussion of the nature of knowledge.

Allow me to begin, though, by dividing the discussion into a practical thread and a theoretical thead:
  • the practical matter of whether educators ought to see themselves as conveyors of authoritative knowledge, whether or not such knowledge actually exists, and
  • the nature of truth, and whether there are truths that can be known in an absolute sense, as opposed to the relativism described by @Zerove
With respect to the second matter, I think that the net result of some 2600 years of philosophical enquiry into the nature of knowledge and truth is that knowledge and truth are at best relative to a community, a symbol system, a model, a perceiver, or some such non-global entity.

It does not matter that "a further dialectic process towards a global understanding of facts" is hampered by this. [x] Arguably, this is beyond our reach - we could engage in the process forever, but never know whether we are even getting closer to a conclusion.

We could discuss the state of knowledge and truth at length, but that might best be a separate thread.
The practical question is whether, despite this, and for more pragmatic reasons, teachers out to be seen as conveyors of authoritative facts and knowledge.

Several reasons have been advanced in the discussion this far:
  • it may be the case that students are unable to learn unless facts and knowledge are presented in this way, ie., there are pedagogical reasons
  • it may be the case that there are cultural reasons for presenting knowledge and information this way, ie., there are cultural reasons
  • it may be that the stability and prosperity of society as a whole may depend on a common understanding (or, if you read Rawls, agreement) on certain propositions, ie., there are social reasons
Against these arguments I offer my own proposition that social, pedagogical and cultural issues are better addressed by encouraging learner autonomy than by encouraging teachers to act as sources of authoritative facts and knowledge.
  • first, it is arguable that students learn better if they are able to understand and reconstitute the knowledge for themselves, rather than having it distributed to them as already known and authoritative. There have been some recent discussions1 around the Common Core approach to mathematics that are illustrative of this.
  • second, it is arguable that culture, rather than being harmed by a decline of authority, is actually strengthened by it. This is a long digression which I won't explore here, but a significant topic worthy of discussion.
  • third, society as a whole is more stable if the ondividuals in it are viewed as autonomous, pursuing (as Mill says) "their own good in their own way". There has been a lot of recent discussion showing the quality of decisions and stability of social systems are increased when organized as networks of diverse autonomous members.
One final note. We will not doubt touch on the distinction between 'authority' and 'expert'. The two are very different. The former represents a perspective that is imposed on the learner. The latter represents a perspective that is given weight by being recognized as such by the learner.

Contrary to the perspective of cMOOCs offered in the paper cited by @Vivian below, the position of connectivists (or, at least, of me) is that while teachers should not take on an authoritative role, they can and certainly should function as experts.

Their role is not merely to facilitate - that is a conflation of connectivism with constructivism. Their role is, to put in slogan form, to model and demonstrate.

[x] In the same way the assertion that "we do not have wings" hampers our ability to fly. But no much it hampers our ability to fly, it does not follow that we have wings.

# # #

@Fleur_Prinsen I guess I would ask first of all why it is wrong for people to come to different conclusions based on the same facts. This is a very common phenomenon, and it is indeed this diversity of opinion that strengthens our own cognitive processes and the cognitive processes of a society as a whole.

But also, secondly, I challenge the presupposition that there is such a thing as the "same facts". To my knowledge everybody experiences the world in different ways, and while there may be social and linguistic conventions concerning our shared experiences (eg., the sentence "Paris is the capital of France") even these are experiences by different people in different ways.

What we call 'facts' are theory-laden representations of the world and experience, and are literally different depending on whether one believes in spiritual entities or not, on whether one believes in an underlying reality or not, on whether one interprets the perceptions as particles or waves. They vary from perspective and point of view, in significant and importance and relevance, in salience, and with reference to background knowledge, context or understanding.

So I do not think that it is an objection that people come to different conclusions. I think it is a strength.

# # #
@jeffreykeefer Most of the research on 'theory-laden data' was actually done on research in the hard sciences. This slide show3 outlines some of the foundational work by people like Popper, Kuhn, Lakatos and Feyerabend to establish this. Even the nature of observations themselves, much less theoretical statements, are interpreted differently depending on one's theoretical perspective.

This is important not because it puts astronomy and astrology on the same level - it doesn't - but because it corrects a persistent misunderstanding about the nature of knowledge and scientific theory.

There is no 'foundation', there are no 'basic truths', no particular 'facts of the matter' carry a special status over and above the rest. Knowledge is not an edifice, construct, or canonical representation of the world. It is far more complex (and far more interesting) than that.

A couple of examples, one from Lakatos, and one from my own writing:
  • the Millikan oil drop experiment1 was one of the foundational moments in chemistry and physics. It involved dropping oil drops between two charge plates, measuring the effect of the plates, and deducing the value of the charge of a single electron. Previous efforts had attempted the same with water, but because the results were so fickle, Millikan employed oil instead. "The Millikan–Ehrenhaft contro-versy can open a new window for students, demonstrating how two well-trained scientists can interpretthe same set of data in two different ways."
  • 'basic' mathematics. It has long been held that children should learn 'basic arithmetic', such as addition and subtraction, multiplication and division. But the selection of these as 'foundational' is completely arbitrary (and are not surprisingly challenegd by Collon Core). Mathematics itself may be thought to be based on a foundational set of axioms, such as are represented by Peano arithmetic1, based in the concepts of set theory, identity, and succession. Or perhaps we could employ Mill's methods1, which derive mathematics from as a set operations on a series of pebbles. In view of modern computation (and to help in their later understanding of modal logic) perhaps the core concepts taught ought to be transitivity, substitutivity, and symmetry.

Sunday, November 01, 2015

That Old Canard

In yesterday's New York Times the Arthur C. Brooks argues that there is systemic bias against conservatives in academia. This is my response. 

It must be coming around to election time again, as this old canard (complete with a 1975 study) makes the rounds again.

First of all, it should be no surprise to find more left-leaning people in a public service industry, just as it is no surprise to find that the boardrooms of banks and corporations are staffed almost entirely by people from the right wing.

It should also be no surprise to see people from the left in occupations with a focus on reason and intellect, as generally the more educated a person is, the more left wing they tend to become.

Third, it should be no surprise to see people in academia adopting aa more liberal stance because, as they say, "reality leans left." The right is known for its support for creationism, climate change denialism, anti-vaxxing, an apocalyptic would view, and a host of non-reality-dfriendly positions.

Finally, it's not even clear that the data brought forward by the right actually supports the contention that there is a right wing bias in academic. The cases are carefully selected, picking from research in sociology rather than, say, schools of business or medicine. And they compare apples and oranges; the same data set might *not* lead to the same conclusion when studying poverty and skin rashes, as these are very different phenomena.

This sort of reasoning is reflective of differences in the way the left and the right regard science, just as we saw here in Canada under ten years of conservative rule. On the right, science follows policy. It follows politics. It is manageed to show support for conclusions (and for industries) that have already won political support. But on the left, policy is derived from science. The political positions supported are those, typically, which are supported by evidence and research in the field.

Indeed, the idea that diversity is a virtue in society is a left-wing idea, not typically supported by conservatives. The great movements that created and shaped a more inclusive society - from feminism to anti-racism to aboriginal rights to GLBTQ-friendly policy - are left-wing movements.

It is typical of the right that it would view diversity not as a policy end in itself, but a piece of 'science' that can be taken out of context and used to prop up and rationalize the unsubstantiated conclusion that academia is out to get them. Why would they reason the left would work this way? Because that's what they, the right, would do in a similar position.

I have no objection to the existence of academics and professionals who can articulate and advance the right-wing argument. I personally do not believe such arguments can be sustained. But I don't think that 'diversity' means that we should promote conservatives to positions of academic responsibility simply on the ground that they are conservative. That's not how diversity works. And if the right were concerned with *understanding* diversity, rather than using it to promote their own self-interests, they would understand this.