The idea of Safe-to-Fail Experiments was developed by Dave Snowden as part of his Cynefin framework. The technique is a way to learn about a complex adaptive system without triggering unintended consequences that are out of your control (See the link above.) But the concept of using probes to learn about complex systems is useful in many other contexts, most notably, in social justice advocacy.
Most advocacy is premised on the idea that there are legal constraints on the behavior of target systems, and that these constraints can be used to change the behavior of the system. In other words, advocacy can use procedures repeatedly to create change. Implicitly, we only need to understand the legal constraint under which a system operates and the change procedures (complaints, lawsuits, etc.). We don’t need to understand the politics or history of the system we are trying to change, all of which are, of course, other kinds of constraints.
But we do need to appreciate these aspects of a system before we can hope to successfully change it. This is because even the most apparently logical procedural path of some bureaucratic machine is, as we all know, a little “Peyton Place”, more complex and messier than the bones of the procedure would suggest. Which is to say, all bureaucracies are Complex Adaptive Systems using much of their available energy to prevent disturbance from creating change through forcing them to modify existing constraints.
From inside a bureaucracy (or any large organization, including for-profit corporations), creating change must involve experiments too small to trigger annihilation of the experimenters or the CAS, but enabling you to learn something useful about the systems dispositional trajectory, about its system of constraints.
Safe-To-Fail is also a useful tool for changing that most personal of CAS, yourself.
In addition to the obvious effect of weak constraints on the target system described earlier, we also need to understand weak constraints as fulcrums for coordination, in the same manner, that our bones, joints, and muscles serve as fulcrums for our movement, even the most sophisticated.
If there are no such constraints the system seems freer than it is when these constraints to movement are present. But, this freedom is like that of an amoeba. You can move anywhere but without any sophistication. You are “free” to do much less than you could do if the “constraints” were present. This idea of using constraints as fulcrums for sophisticated advocacy is the key to understanding how we can use the weak constraints (and sometimes the strong constraints) in a system to leverage change. What constraints enable, among other things, is the coordination of our advocacy work to achieve meaningful impact.
Because strong constraints are well defended in target CAS, it can be difficult to change them directly. But the strong constraints still represent fulcrums that the target must respect. So they can be used in much the same way as Aikido or Jiu-jitsu, by channeling the investment in energy that the target CAS must provide in order to prevent damage to itself, into “forced” change. This is different from trying to eliminate or replace strong constraints, which, frankly, almost always ends very badly.
I think frugality drives innovation, just like other constraints do. One of the only ways to get out of a tight box is to invent your way out. -Jeff Bezos
The more constraints one imposes, the more one frees one’s self. And the arbitrariness of the constraint serves only to obtain precision of execution. -Igor Stravinsky
Problems are hidden opportunities, and constraints can actually boost creativity.
So, how do we use weak signals as a basis for changing Complex Adaptive Systems (CAS)? We must look carefully at the weak signal to understand how or if this signal represents a weak constraint, and what the constraint means to the Target CAS.
Earlier I pointed out that weak links buffer the wildness of CAS. This buffering is a form of constraint, and that’s why buffering works. The buffer acts a bit like the banks on a river, constraining the flow of the river without dictating the movement of individual water molecules.
Our usual understanding of system constraints mimics the beliefs of the homeless community and uber-rich communities. Constraints are barriers to the safety or freedom of these communities, and so they are eliminated. Successful elimination of such weak constraints makes those social communities brittle and hyper-responsive to small disturbances.
The image above is a drawing of the effect of the Underground Railroad during and around the Civil War. The Underground Railroad functioned as a weak constraint on the Southern Slave System It was largely ignored when it was small but was attacked (ineffectively) when it expanded and began to operate as a sign of the weakness of that Southern Slave System.
The Underground Railroad was more than a simple barrier. It actively forced the Southern Slave System to respond to it. In the same way, weak constraints do more than provide simple barriers to the system of which they are a part.
Target systems for our advocacy have many weak constraints that are a normal expected part of their day-to-day experience. They are usually ignored or tolerated because the behavior of the weak constraint is a small local cycle that doesn’t threaten the larger system’s normal behaviors. If the weak constraint begins to expand its impact on the larger system, it will trigger a response of some kind from the larger CAS.
In Part Three, I’ll talk more clearly about how we use weak constraints (and sometimes strong constraints) to produce advocated change in CAS.
Bricolage is rightly viewed as one of the “Powers of the Weak”. Elites typically view power as something exerted by a predictable machine of propaganda, sanctions, and punishments, and they view insurgents trying to change this as working to replace their machine of power with some other one.
So, bricolage, used as a tool of subversion, misdirection, or organizing, is hard for elites to see, or target. This is especially true if the bricolage is used to solve a local problem.
The point of using bricolage rather than using the system is to avoid having the problem-solving bricolage subjected to the services logic of the system. This system services logic includes assumptions of:
Spending scarce resources to detect fraud
Using “failure demand” as a tool for managing system work rather than actually providing the service
After an initial period of seeking out persons eligible for the support, gradually turning the point of the system increasingly toward denial of supports.
Bricolage allows a more coherent connection between support purpose and behavior. This coherence is lost once the support is subjected to the support logic of the system.
Tinkering is standard behavior for anyone who is curious. Bricolage is a French word defining tinkering as finding a solution to a problem with whatever is in your immediate environment. Bricolage makes problem-solving local and personal and is more than just playing. Bricolage reliably produces solutions that are inexpensive, easier than managed solutions to implement, and well matched to the actual reality of the problem rather than the “planned” reality of the problem. In fact, in modern life, bricolage is a common response to solutions that are imposed by organized management.
I suspect bricolage was a primary way our hunter-gatherer ancestors engaged the problems of their daily lives. Adequate solutions would become part of a multi-generational exploration of what possibilities these solutions held, a kind of socially organized exaptive process. Bricolage speaks to the personal “engineering” drive we all have.
My father was an extremely capable chemical engineer who worked for Dow Chemical for 45 years. His primary focus over the course of his career was something called “process engineering”. His task was to take a reasonably successful research project and find out if the project had commercial potential. Researchers tend to think that you scale their successful research by simply making it a bigger version of what they used as their research methodology.
In reality, designing and building a commercial pilot that is a million times larger than the research process, respecting the physical environment of seasonal temperature changes, the length of pipes, the delivery of chemical components at the right temperature and with the catalysts and pre-product components at the right time, so the next step in the process can be successfully initiated, and so on. Process Engineering is a particularly large form of bricolage, and the difference between ideology (research) and engineering (bricolage) has many lessons for any attempt to change any CAS.
The quest for certainty blocks the search for meaning. -Erich Fromm
Environmental scanning is not monitoring. It is a deliberate activity designed to increase the possibility of surprise. It assumes you already have a commitment to changing your current view of reality by exposing yourself to what you couldn’t anticipate.
So, having a rigid procedure for environmental scanning won’t work. Over time, you’ll find less and less novelty and more and more repetition. You need an approach that has enough noise in its scan to produce stuff you didn’t expect or even know might exist.
I use a variety of ways of accessing information, including ones that I am uncomfortable with, or frankly disagree with, in order to maximize the possibility of surprise. This approach requires scanning a lot of useless crap. But I’ve gotten faster and more accurate in my scanning for crap over the years, so I still get a fair amount of surprise. I also add and subtract sources regularly to maintain the surprise. I use the criterion that a particular source is no longer surprising to me.
Since anybody’s experience of surprise is conditioned by the personal path they have followed in all its eccentricity and uniqueness, a useful environmental scanning approach will be customized to that anybody. The vagaries and dynamic of our personal purpose and meaning will also influence what we find surprising, and that will change over time as we change. Our ecosystem always includes ourselves.
The way that this kind of environmental scanning helps us detect weak signals is best understood as similar to a kind of process called stochastic resonance. Stochastic resonance happens when you add noise to a weak signal. That part of the noise that matches some part of the signal will boost the “volume” of the signal. That part that doesn’t match will cancel out through interaction with other parts of the noise.
We often try to remove noise if we are dealing with a weak signal because we believe that will make the signal clearer. So it is surprising to find out that noise can help us understand weak signals. This reality is an interesting metaphor/framework for interaction with any CAS.
“I wanted a perfect ending. Now I’ve learned, the hard way, that some poems don’t rhyme, and some stories don’t have a clear beginning, middle, and end. Life is about not knowing, having to change, taking the moment and making the best of it, without knowing what’s going to happen next. Delicious Ambiguity.”
― Gilda Radner
Instead of glossing over surprises as failures of understanding, we should focus on them until we have grasped their novelty and how that novelty needs to change our view of reality. We need to avoid abstracting from surprise to make it only another example of what we already know to be true.
Novel occurrences are novel for us, but they are also typically some “next step” from that with which we are already familiar. They are often called the “adjacent possible” because once they have occurred, it is fairly easy to see how they came about. This is true even if no one anticipates them. It is important to remember that in a Complex Adaptive System, there are always many adjacent possibilities for the future.
There is another common problem that results from rationalizing surprises. We look back on the surprise and try to figure out who accurately anticipated it. We think this will improve our prediction capability in the future.
Looking back does improve our understanding of the current situation. It doesn’t improve our ability to predict any genuinely novel future. If we examine what people thought about the future before the novel occurrence, we will see a very large number of ideas about what might happen. The particular idea about the future that turned out to be accurate had no more or less information about its likelihood than many of the other ideas. The novelty tells us something useful about the current state of the CAS we are in and where it might evolve in the short term. It doesn’t improve our ability to foresee the genuinely new.
Great things are done by a series of small things brought together – Van Gogh
Even the largest avalanche is triggered by small things –VernorVinge
You’ve got to think about big things while you’re doing small things so that all the small things go in the right direction –Alvin Toffler
Men don’t pay attention to small things – Katherine Johnson
Because our usual habits make us ignore weak signals, we need to cultivate new habits that make it more likely we will notice them. These new habits can’t be automatic. They must involve reflective attention-not just attention to something that is there, but consideration of it’s meaning. Below is a list of techniques and concepts that I hope will aid you in seeing what is important, but almost not there:
Surprise Can Point to Weak Signals
Environmental Scanning for Weak Signals
Tinkering and Bricolage to find Weak Signals Right Around You
Learning About Weak Signals Through Safe-to-Fail Experiments
Thinking of Weak Signals as Insurgencies
I’ll try to amplify each of these in the posts below.
First, we have to actually pay attention to them. Our default is to ignore them as unimportant. That means we have to have a way of making them stand out. Most importantly, we have to conserve the meaning in the story of any weak signal instead of homogenizing that meaning or averaging it or abstracting it through ordinary statistical analysis. That is one of the strengths of SenseMaker. Its function is, first of all, to make raw weak signals stand out in a number of ways. We need to do the same.
Then, we have to ask ourselves about the value of the narratives we have acquired to support or undermine positive change. This isn’t simple to do. But our first order goal with these signals is to increase the ones that support positive change and decrease the ones that undermine it. Because these are weak signals, it is feasible for us to try out ways to do both of these in time frames that let us change our approach as we learn which weak signals we can effectively increase and decrease, and when we need to look at different initiatives to produce these outcomes.
The reason why this works at all in trying to change a CAS is that the cycle of experiment and evaluation is short. Such an approach respects the dispositional nature of CAS and doesn’t require us to use prediction and mechanical outcomes as the signs of progress.
Cynefin is a body of knowledge and tools to assist in changing CAS, among other things. Cynefin, as an enterprise intervention, also has developed a “narrative access and analysis tool” called SenseMaker™. Sensemaker allows the intervenors to accurately access raw views by the participants as short narratives without groupthink or homogenization. It is this ability that allows for the detection of weak signals.
Because SenseMaker has developed an app, it is possible for its users to engage huge numbers of people in a very short time. The example that had the most impact on my understanding of its capacities was an effort to work around the unwillingness of local citizens to say what they actually thought to US civil and military personnel in SE Asia.
The system was used to ask children to relate a story from their grandparents about the most important lesson that the grandparents had learned in their lives. Then the children sent the stories using the SenseMaker app. This project got 50,000 stories in four days. There is simply nothing else that supports authentic narrative by real participants with the speed of SenseMaker.
Unfortunately for our community, SenseMaker is an enterprise tool and is priced that way. I have been exploring ways we might be able to use this system in our community, but I am some distance from a genuine solution.
That doesn’t mean that we can’t make use of the idea if we can come up with ways to assure fidelity to SenseMaker’s ability to easily access real raw narratives from participants.
I’ll discuss some ideas for using this general framework to get meaningful narratives in our community in later posts. For now, I hope you can see the importance of weak signals in the development and use of our FutureStrategy.