A longer than normal title to this post (from Martin Luther King) should not be seen as a modern-day version of Plato’s cave. In this latest post of sidecasting, I want to pick up on the final three categories from the table I published yesterday. I’m still calling it sidecasting by the way, although the lobby for widecasting is increasing so you have an opinion let me know. As the series is coming to an end I also decided it was appropriate for the picture to change a little, next time or more likely the time after we land the fish.
So returning to my extension of Dreborg’s comparison of forecasting and backcasting. In my last post (and I am assuming the content of that) I handled the philosophical and perspective differences between the three approaches. I should emphasis that I see them as complementary, not as competing. I am trying to place them in context. Just as the Cynefin framework seeks to place different and contradictory techniques into context through its five domains, so we should realise that the philosophical views and perspectives determine the applicability of the approaches, methods, and techniques that I will address today.
- The approach differs between forecasting and backcasting in that the former extrapolates trends, measures sensitivity probability, etc. while the latter assumes the definition of a future state and looks to understand the consequences that would allow that future to be manifest. Both (as I have said before) assume causality and a degree of knowability through research or models that creates an evidence base that in turn justifies action or inaction. Sidecasting, as a complexity technique, is in radical contrast with both of these. As a complex system is dispositional, not causal we cannot fall back to conventional approaches to evidence (more on that in the final post) but we can test and understand the evolutionary potential of the system. Critically we need to do things: (i) map the dominant dispositions (hollows in the attached landscape which come from SenseMaker®) and determine their stability, while (ii) paying attention to outliers (the dots in the landscape). In general strong hollows indicate a high degree of consensus so we have to test for coherence to determine if they are authentic or not. Outliers may be coherence (again we have to test) but by definition, they lack consensus so convincing people they are a realistic threat or opportunity will be difficult. Readers may want to reference this post which introduced the concept of balancing coherence and consensus to achieve authenticity, something I explored further in subsequent posts.
- Methods are interesting in Dreborg’s comparison. He contrasts the econometric models of forecasters with the extrapolations of the back casters. To be honest I can’t see much difference other than the direction of travel. Yes, partial and conditional extrapolations will of their nature handle polarities and test limits but then so do models, in particular many of those used by behavioural economics. sidecasting on the other hand has a broad range of techniques available. I will spend most of the next post on one of those (too frequently confused with backcasting) namely The Future Backwards (note to self, having read that link I need to update it). We also have a range of techniques both method-based, workshop-based and software-based which are all designed to increase the numbers of perspectives we bring to bear. In my next post, along with The Future Backwards I will summarise how each of those methods works. For the moment it’s sufficient to be aware that they are about gaining perspective, deploying networks of diverse loosely coupled agents, and are focused on counterfactuals and the use of paradox rather than polarities.
- Techniques for me are difficult to distinguish from methods. They tend to be intertwined. Dreborg lists none for backcasting and a simple reference to algorithms for forecasting. In the pantheon of Cognitive Edge approaches, we use workshops, software as well as managed interaction across cultural and functional boundaries. But these are best described in methods.
So comparisons out of the way, in my next and hopefully penultimate post on this subject I will go through the various methods referenced here indicating how they match the philosophy. In the final post, I will show a workflow for using these ideas in sidecasting and draw conclusions.
Thank by the way for the email, tweets, and other references on this series – its obviously struck a chord. I will take a break from it for a couple of days, however.
Banner picture by McKayla Crump on Unsplash