Saturday 3 October 2020

Escaping the Ego Machine: Rethinking Systems Architecture for Educational Technology

I had an interesting discussion with a friend earlier this week about creating an online space for dialogue about education. The default technical solutions to this all look the same - they are all content management systems in various forms, where "personalisation" is provided in the form of tailoring the presented content to individuals based on their activities. 

Educational platforms like VLEs do not have the sophistication of Facebook's personalisation algorithms, although of course there is a lot of effort going in to trying to make this work. They are cruder CMS systems, displaying content according to particular organisational groups - e.g. modules, classes, programmes, etc. 

But saying all these systems are the same is to say that their basic architecture is the same: there is a server which acts as a repository of stuff contributed by teachers and (sometimes) students. Access to the system is controlled by an individual's account - their username and password. Activities too are tied to the individual. Although some systems allow for group activities in various ways, such group activity is really a different coordination of individual people, and a different presentation of content. 

But what is happening here? When we talk about "content" we are basically talking about "documents". A web-page is a document, a personalised feed in Facebook is a document, an assignment is a document... Who controls the documents when they're on the platform? 

This blog post is a document which I am writing and clearly I control what it says. Most blog posts are not collaborative efforts, they are individual and personal. When individuals publish things, they are mined for whatever information Google might find useful to sell stuff to other individuals. Obviously I have no control over this.

A lot of social media feels like an "Ego machine" - my thoughts, my documents, my desires feed a market of other individuals' thoughts, documents, desires. But to what extent is it an ego machine because of the systems architecture upon which it sits? If we didn't have servers and personal accounts, could it be different?

The question can be represented visually. The "ego machine" model can be represented as a set of interconnected nodes - individuals - with lines representing communicative acts of some sort. This is a "network model" shown below on the left. But social relationships in real life feel more like a cellular model, shown on the right. Cells are not individuals, they are "dialogues" or "codes of communication" - they might even be "communities of practice" - although I would want to dig into that loaded term...

What social media platforms have done is to overlay a "network model" of communication - which, after all was the simplest architecture for computers - onto a cellular model. The problem is that this excludes the detail of what happens in individual cells. 

Cells are not "nodes". They are active processes with a boundary that separates an "environment" which is negotiated and separated from a set of "internal operations". The collective action of the processes engaging in the environment and the processes maintaining internal states is to maintain the boundary - i.e. to maintain homeostasis with the environment. 

Seeing cells as processes with an inside, an outside, and a boundary presents a set of powerful questions which must be addressed. What are the internal processes doing? What are the external processes (those which engage the environment) doing? What happens at the boundary? More powerfully, these questions can be asked of all cells - including those that constitute the natural world. In the light of this, the network diagrams of social media look very one-dimensional - and I think this is reflected in our experience of the technology. 

Once we start asking questions about cells - whether its questions about cells of communication, or cells in our bodies, questions about "ego" take a back seat. We don't know the locus of consciousness and self - but we do know that it sits on a biological substrate, and that this biological substrate is cellular. We can also point to homologies in cellular processes at different levels of organisation. So the homology between cancer and the pathologies of confirmation bias and the corruption of a political system are telling, or the homology between epigenetic mechanisms at a cellular level, and our communicative practice online are also very powerful.

Moreover, once ego takes a back seat, focus falls not on the viability of individual cells, but on the coherence of the whole. 

That is where our technology needs to take us. I don't think a centralised server-based architecture can deliver this. We need something else. Documents can now be created collaboratively and used as vehicles for dialogue. This is opening the way for individuals to become "inter-mediaries" in document production, rather than authors. Distributing documents for diverse collaborative production is something that Microsoft Teams does very well - which is interesting - but then this technology works on a different technical principle to the  CMS. Can a coordination between different groups producing different documents be created so that some groups can naturally split (mitosis/meiosis) or die (apoptosis), all the time maintaining the coherence of the whole? These are the questions we really need to think about if we are to set ourselves straight for the future. 


Rupert Wegerif said...

Can we not design for both dialogues and fixed nodes? Individual learning profiles might be useful to replace high stakes exams in interviews. Docs have degrees of stability. Dialogic short term processes need to issue in longer term knowledge - temporal levels

Nat said...

Hi Mark, I'm wondering if you've heard of underscore protocol ( - in some senses I think they're pointing towards what you're talking about in your last paragraph. The idea is giving collaborative documents git-like tracking that means a given 'document' could evolve over time from various collaborators taking different perspectives on it, adding to it, etc. While not completely divorced from a network structure, it seems like a solid effort towards a more organic way of collaborating on and sharing ideas digitally.