More on Scenarios and Alternative Futures
Intro to Futures, Week 13
A Taxonomy of Scenario Techniques
Last week I covered scenarios: their consensus definition, how they relate to all the other concepts discussed so far in the Intro class, and a walkthrough of the most common way of building a set of scenarios (the Intuitive Logics 2x2 approach). Even though it greatly outnumbers other techniques for building scenarios, there’s a very broad and diverse set of other options that can be used. There’s an excellent paper from Peter Bishop, Andy Hines, and Terry Collins that systematically lays out the various techniques and creates categories1. Across these techniques, a basic pattern emerges for the creation and use of scenarios: framing the problem, scanning for relevant content, describing alternative futures / creating scenario(s), envisioning a preferred future, doing planning toward it, and then acting on the plan.
The most basic category is judgment techniques. This mostly just uses the human brain’s natural imagination to generate scenarios, whether through “genius forecasting” that is an informed guess about what the future will be like2, something like guided meditation to visualize a different future, having a group role-play within a set situation, or even a more formal activity like sketching out relevant variables and future themes, and then writing down what you think the impact of each theme on each variable will be.
Baseline methods can be used to create a single “expected” future by extrapolating trends. The Manoa method involves identifying three dominant trends, doing a futures wheel3 for each, look at the impacts that each trend might have on the others, and then take all these effects and write a scenario. Systemic Scenarios is a similar method that uses a causal model instead of the looser cross-impact work just described.
Elaboration of fixed scenarios takes the basics of the scenarios as fixed and focuses on imagining the details. This can be done via incasting4, where extreme alternative scenarios are presented and the goal is to describe the impact per domain of interest. Another approach is a more structured SRI matrix, where the rows are a set of relevant domains5 and the columns are a set of four defined futures: what’s expected, best case, worst case, and something really weird (“highly different” is the technical term); filling in the elements of the matrix is the exercise.
Event sequences create branching possibilities for how the future might unfold. This can be done quantitatively with a probability tree, which would allow a calculation of the probability of each final state. Sociovision groups similar branches of the tree together into stories. As a completely different technique, divergence mapping brainstorms a huge list of possible future events, maps each to a set of time horizons, and then links forward related events through time.
Backcasting is an attempt to shake people free from the baggage of the present and imagine something really different. NASA does this via their Horizon Mission Methodology: take some fantastical mission6, assume it happened, decompose into multiple components how it would have been possible, imagine the technology that would be required for each component of the mission to succeed, and see if any work currently underway (or feasible) would contribute to building that technology. Impact of Future Technologies is a more passive version of the same idea: for the scenarios of interest, imagine what breakthroughs would be needed to bring them to pass, and then monitor the environment for signs of those breakthroughs. In future mapping, the work team gets a predefined list of scenarios and events, and maps the events into plausible stories of how the scenarios came to be.
Some approaches directly tackle the dimensions of uncertainty that make it difficult to say useful things about the future. This is where the Intuitive Logics technique for creating a 2x2 matrix of scenarios fits in, as well as the generalized version, Morphological Analysis or Field Anomaly Relaxation, where instead of two drivers with two extremes you consider an arbitrary number of drivers each with an arbitrary number of values. The problem with this method is that the number of scenarios created by this increases exponentially as the drivers and values increase. Option Development and Evaluation is a solution: rank the combinations of features in this hypermatrix by consistency and just develop scenarios for the top n most internally consistent combinations.
Cross-impact methods start with a matrix of events giving the probability of each as well as the probability of each conditioned on each other event occurring. Then a threshold for events is selected and a set of events is selected as having occurred. This could also be accomplished by modeling the possible ranges and interactions and then using a Monte Carlo approach to simulate a large number of outcomes, and then reporting the distributions of the simulations.
Another quantitative set of techniques starts from a single future and generates other possibilities by varying the inputs. Trend Impact Analysis estimates the impact of possible events on current trends. Sensitivity analysis explores the range of outcomes that are possible when the model’s variables, parameters, or inputs are changed. Dynamic models cluster similar events, a causal model is built for each cluster, common elements are brought together, and a range of possible values is explored.
Alternative Futures from Manoa
Clearly the generation of scenarios is a rich arena with a wildly diverse set of techniques. But what if scenarios, at least as previously defined, weren’t the primary goal? A popular alternative way of conceiving futures is called, perhaps not surprisingly, Alternative Futures. It comes from the Manoa School under Jim Dator and represents a slightly different approach to the foresight project. Rather than scoping, it starts with an in-depth review of the past and present of the group/organization/community, to create a common story to jump off from and to get people willing to discuss the future. Some forecasting is done using the familiar elements (trends, emerging issues, etc). Crucially, people get the chance to deeply experience alternative futures. Then preferred futures can be designed, a plan made to move toward it, and finally the commitment to a standing foresight function and capability (and ideally a commitment to repeat the exercise in the future) is established.
What’s so special about the alternative futures in question? The Manoa School has constructed four archetypal scenarios, based on big-picture patterns of civilizational change and abstracted from thousands of cases, then applied to the present as a starting point. The first archetype is Growth: current trends continue (to make it interesting, make them accelerated/exaggerated); this is a world of high technology, urbanization, increasing corporate control, etc, well aligned with the cyberpunk world of Neuromancer, Blade Runner, etc. The second archetype is Collapse: society (or maybe just the organization or community being considered) has faced an extinction- or near-extinction-level event. Notably, the future being described doesn’t take place during the collapse, but afterward; the closest touchstones might be the solarpunk genre, Nausicaä of the Valley of the Wind, or possibly the TV show Station Eleven (from what I’ve heard). The third is Discipline: society has been oriented around sustainability or a focus on fundamental values, usually through the implementation of some form of authoritarianism; lots of dystopian fiction fits here, e.g. The Giver or Divergent. The fourth is Transformation: the concept of humanity is fundamentally changed via transhumanism or some kind of enormous technological and/or spiritual renaissance; Star Trek describes a society transformed by post-scarcity conditions. Of course, if the exercise is to think forward only 20 years, or to study a single organization, the fictional worlds described above may be too big a leap to be credible - move in the intended direction as far as you think people can stretch their minds without breaking the suspension of disbelief.
None of these scenarios should be thought of as inherently better or worse than the others, though each clearly has winners and losers. The goal is to take the group and immerse everyone in at least one of these futures. At minimum, this means sending groups to different corners of the room with a well-written brief, but that only exposes a tiny fraction of what’s possible. If you can get the buy-in (and money), spend time building each future into its own room, with scenery, sounds, smells, artifacts7, even actors that make the experience as vivid as possible - the goal is to make the familiar seem strange, and the strange seem familiar. Then, in groups of maybe about 10, the people who have experienced this future can think and write about their experiences in their experienced future, including their general impressions of what it would be like to live in it (focusing on what they would do to try to thrive), how preferable that future is, and what actions could be taken (right now) to move closer to or farther from it. Then these group reports can be consolidated and shared so people can imagine what the other futures were like, and generally debrief with comments and questions, before moving on to the preferred future work as described above.
Again, there are lots of ways to generate scenarios, and multiple ways to use them. The goal is to expand people’s mind to the fact that the future is likely to be radically different from today, and also from the “official” future of either simple extrapolation of trends or today’s problems being resolved and not creating new ones.
The paper has two other contributions that make it a classic worthy of frequent reference: it does a great job defining and differentiating usually vague terms like project, approach, method, and technique in a futures context, so the concepts being described can take on a more definite shape; and it summarizes Van Notten’s typology of scenarios, including dimensions like normative vs descriptive, long vs short term, conventional vs weird, etc.
This, incidentally, is what I have been intuitively doing for years as the author of a digest about trends in healthcare analytics. Becoming aware that people must be doing something more systematic, and being confident that learning those techniques would make me more effective, was a big part of my entry to futures and foresight.
i.e. neither forecasting (extrapolating from the present to the future) nor backcasting (the reverse), but immersing people in a given scenario and exploring the world it implies.
The example given in the paper is a one-day mission to Jupiter.
One accessible exercise for designing future artifacts is The Thing From The Future.