Emergence

This youtube video on emergence cropped up in my YouTube feed.

Provided description

The term pattern formation refers to the process through which a coherent set of associations between element’s states is formed and persists over some period of time, it captures the essence of self-­organization and emergence in all kinds of systems. A primary question we are interested in when studying any pattern is the question of how was it generated or formed? In answering this question, we can make a fundamental distinction between those patterns that were created through order being imposed by some other external organization, or those that were created through the pattern being internally generated.

Identifies that there is two broad ways patterns of organisation occur

Some systems have the ability to self organise.

Apparently, a primary characteristic of emergent pattern formation is the absence of centralised control. Local dynamics - interactions? - influence the system's global behaviour - global pattern.

In an emergent system it's not possible to coordinate/organise the macro level behaviour. Meaning the focus has to be on the small components and their interactions.

Feedback is a central characteristics of internally generated patterns. There's positive feedback and negative feedback. It is through positive feedback that a small change will get amplified into the global pattern. Negative feedback places constraints on pattern formation.

Open source projects are given as an example. The large-scale stable open source projects are those that get positive feedback. IMHO, they also have some commonly identified components/interactions that are known to help.

It had obvious resonance with Col's complexity informed approach to meso-level LA. Also my question about whether Col's design principles had enough explicit connections with complexity thinking.

All this requires energy to go into the system to enable pattern formation. And how the elements of the system are able to intercept and transform the energy is important.

Pattern formation can also be seen as a form of adaptation whereby the system adapts to its environment.

Reinforces the importance of organisation to be able to capture energy/resources to achieve the purpose.

Applying to learning analytics

Most organisational approaches to learning analytics have pre-defined structures/patterns. Structures put in place by the vendor or perhaps the local IT group. Not necessarily a bad thing, but one of the big problems is those structures tend to close down on-going emergence. i.e. they limit positive feedback - in the form of useful on-going interactions (student/student, student/teacher, teacher/teacher, teacher & student/learning analytics etc) - which in turn limits the emergency of new, more appropriate structures/patterns. Structures that better fit the individual variances and complexities of each different need for learning analytics.

On the other hand, tools like OnTask, SRES and EASI (in somewhat different ways) enable on-going interactions which allow varying levels of emergence. EASI achieved it through the EASI crew being deeply emeshed in the CQU context and being able to inject contextual needs into the tool. Both initially and through some evolution. OnTask/SRES achieve this by placing more reliance on the people using the tool to inject their specific contextual needs into the data included into and the use of the tool. This is not easy to do. It typically requires some additional assistance from meso level practitioners to inject that contextual stuff. In the video they talk about this being the energy that's required for emergence.

Aside - What I wonder about OnTask and SRES is the difference in how they've evolved over time. The C in PIRAC. What energy is being injected (and required) to evolve those tools in response to lessons learned during use. My vague impression is that SRES has probably seen more evolution (but perhaps required it). Interestingly, there are three versions of OnTask based on different tech stacks. But it appears that all the recent updates on the core product are to upgrade libraries or patch vulnerabilities.

But the end result, is that it is perhaps the ability for new patterns of analytics to emerge over time that is the key. It's not about designing a fixed system at the start. It's about being able to encourage emergence of new patterns as efficiently as possible.

The challenge for meso-level practitioners (and academics) is that the "energy" they have available to encourage this emergence is limited. Limited by their technical skills, their data analysis skills, their time, the available tools, what they are allowed to do etc. Suggesting perhaps the design principles need to help the meso-level practitioners figure out what energy efficient ways exist for them in their context to encourage this emergence?