The importance of having the same vocabulary – part 3/4 – PROBLEM, CAUSE and EFFECT and friends

This is part three of this series about the importance of having the same vocabulary and today we will elaborate on first the word PROBLEM and then the words CAUSE and EFFECT.

Why is problem an important word to bring up?
Art Smalley, a Toyota Production System expert, author and former Toyota employee, states in his recent book “Four types of problems” [1], that problems can be caused/reactive or created/proactive. The reactive ones means that the problems require reactiveness in abnormal conditions, which means that the problems are caused (most of the times by ourselves), having root causes that we need to solve. Problems in normal conditions means proactiveness if we want to go to next level to become better or make a new product, incrementally or by brand new innovations.

This dividing into caused and created problems has the right granularity level, since it does not state anything about context, neither regarding WHAT we are doing nor HOW we are doing it. In his book, Smalley continues to divide the caused and created problems further, but then context come into the picture. And context is king, Dave Snowden repeatedly states, where we have big difference regarding the repetitive manufacturing on one side (WHAT (always simple) equal to HOW) and non-repetitive product development on the other side (WHAT not equal to HOW). So, we leave context out of our picture since we want to be able to handle any problem. But, this first context-free division of problems into caused and created ones, that Smalley has made, is extremely important, because they are very very different. To not mix them up, the reminder of the blog post will call created problems for non-caused problems, since we have not made them ourselves.

If we use Dave Snowden’s well-renowned Cynefin™ Framework, the solution to a non-caused problem, the WHAT (for example any activity when we are designing or manufacturing a product), can quite easily be mapped into one of the domains in the Cynefin™ Framework, when we use the Cynefin™ Framework which can be used in many different ways, as a categorisation model; do we have the knowledge to do it (obvious/simple), can we do some prototypes and get guidance from our experts (complicated), or do we not have the knowledge (complex).

The caused problems on the other hand have cause(s), meaning they are only symptoms of something deeper rooted. And symptoms cannot be solved, only their root causes. This means that we cannot put caused problems directly in any of the four general domains. But, the Cynefin™ Framework has another domain for that, the domain of disorder. And one sub-domain of disorder is inauthentic disorder, where we will be as long as we do not know in which domain our problem is in. And for symptoms we need to ask multiple why? on them in order to finally reach their root cause(s), which are our negated principles. Remember also that principles are universal, and a negation or why question will not change that, so the symptoms and root causes will be in inauthentic disorder too. Finding the root causes can be easily done in a  workshop or similar by collecting the narratives of the organisation, the good and the bad, where the “bad” narratives will be our problems. And preferrably thes narratives are written down individually and an individual yellow notes session for 10-15 minutes will do the job perfectly. Then it is just to start to ask why on the problems (symptoms) to get down to the root cause(s).

Many of the caused problems are normally referred to as wicked problems, or intractable problems as Dave Snowden prefers to call them. And since caused problems are unsolvable symptoms, this means that we can neither solve them by doing safe-to-fail experiments with multiple, oblique or contrary experiments, nor use set-based design, nor focus on extending the problem definitions phase (to avoid premature convergence), nor situational mapping to nudge people in the right directions, nor change the constraint structure, etc., instead we need to ask multiple why to get to the root causes.

If we think about the evolution for a moment, it continuously makes safe-to-fail experiments in parallell, in order to  design for resilience, adaptability and evolvability. But, there is no one controlling these experiments, the eco system is simply a distributed system. BUT, the design that fails in nature, will fail. Nature is not going to repair it. This fail is a symptom that could not be known in advance due to the complexity of nature on earth. But still, symptoms are unsolvable, so we need to find the root causes and fix then. Here comes the cleverness of the evolution, it is safe-to-fail. This means, that nature has already solved all possible root causes in advance by using parallel paths for the evolution of every specie and its co-existence with all other spieces and the environment changes on earth. There is really no other way of solving unresilience or unadaptability within an existing specie, when the environment changes too fast. It is too late. There is also the reason for bigger living creatures to extinct, the number of them are fewer, the number of parallel paths are not only too few, also the feed-back loop are simply also too long for new spiecies to evolve. The extinction of the dinosaurs is one example.
But, as we can understand, the evolution of species are closer to the evolution of software and hardware products, and not for problems within our organisations, sine we there can not only find the organisational root causes, we can also fix them.

And since symptoms are unsolvable, they are also uncategorisable and this indirectly means that we do not know what method to use; we need to find the root cause(s) first and then apply our WHAT (the activity), to be able to categorise and choose context, when we also can chose the appropriate method. And all symptoms together definitely gives a mess, a system of problems, as Dr. Russell Ackoff would have put it.

Trying to categorise too early, lead us to a catch with both Smalley’s and Snowden’s thinking, since they go directly to categorisation of the problems, without analysing if the problem is caused or non-caused. And a caused problem as we have seen above, can never be categorised, since it is only an unsolvable symptom that not even rocket science can solve. Cynefin™ Framework’s inauthentic disorder is of course a good start because we cannot categorise a caused problem, but without asking why? on it (as we need as stated above), we can also never find its root cause(s), meaning we are never close to be able to categorise it. When we instead, by asking multiple why?, finally find the root cause(s) to our caused problem, they are on the other hand simply only activities that need to be solved. This means that the root causes can be complex, complicated or obvious/simple ones and can be put in one of the general domains of the Cynefin™ Framework, in the same way as for non-caused problems.

Why are cause and effect important words to bring up?
Cause and effect are important to bring up, because when we are in the ordered domains, we have linear cause and effect chains in the Obvious domain in the Cynefin™ Framework, making our prediction very high. In the Complicated domain we have some different choices that need to be done, but where we with help from our experts, still with high prediction and some exploiting prototypes can predict when we will be ready, to what cost, resource need and quality.

But, when we are talking about complexity (sometimes called science of uncertainty) which we find in the Complex domain, the words CAUSE and EFFECT are often avoided, since they are associated with linearity, which we only have in the two ordered domains. Instead other words are used as; unintended consequence, intervention, dispositional state, phenomena, etc. when we talk about nonlinearity*.

But the question is if cause and effect cannot be used in the Complex domain, or as a matter of fact should be used in order to have a better knowledge transfer between the complexity theory discipline and other disciplines. A better knowledge transfer and understanding between disciplines on a high level means that we can accelerate transdisciplinary work, many times needed for extraordinary innovations and exaptations.

When we are talking about problem-solving, cause and effect are also synonyms to problem and symptom, where we need to find the root cause to a problem in order to be able to completely solve and all the negative effects. See this blog post series for a thorough elaboration on problem-solving in order to increase our understanding about problems and to increase our ability to solve problems, not only for products, but also organisational problems and their tight connection.

Here are definitions of the words metnitoned above from Cambridge Dictionary (C), Oxford Living dictionaries (Lexico) (O), Wikipedia (W) and Dave Snowden at Cognitive Edge (S):

cause (C) – the reason why something, especially something bad, happens (non necessarily a live system)

root cause (O) – the basic cause of something (not necessarily a live system)

effect (O) – a change which is a result or consequence of an action or other cause (not necessarily a live system)

symptom (C) – any single problem that is caused by and shows a more serious and general problem (not necessarily a live system)

intervention (C) – action taken to intentionally become involved in a difficult situation in order to improve it or prevent it from getting worse (live system)

consequence (C) – a result of an action or situation, esp. (in the plural) a bad result (live system)

unintended consequences (W) – In the social sciences, unintended consequences (sometimes unanticipated consequences or unforeseen consequences) are outcomes that are not the ones foreseen and intended by a purposeful action (live system)

dispositional state (S) [2]- the current view of the system, a set of possibilities and plausibilities in which a future state cannot be predicted or repeated, due to its inherent uncertainty (a situational assessment made, in a crisis or not, will get us a dispositional state) (live system)

As we can see above, on this high level there seems to be very little difference or no difference at all between the different words that are used no matter if we have linearity or nonlinearity; they still mean that actions are taken/happen to get/have a result, which is highly predictable when we have linearity, but only to get feedback from experiments, preferrably parallel and contradictive, in order to gain knowledge when we are in the Complex domain. But, if we instead look into if the system are live or not, we can see some differences, that intervention, consequence, unintended consequences and situational assessment are only for live systems. This is a very important difference, since when we talk about system definitions for products or organisation definitions for organisations, we cannot use these words, since a system/organisation definition is not a live system, it is frankly only the principles building up the system.

And we can also see that EFFECT also can be used for live systems, when referring to some important phenomena [3] in complexity:

The Cobra effect (unintended consequences generated) [3]
In British India a reward was offered for dead cobras in an attempt to reduce the danger to humans.  It worked well for a period but then people started to breed cobras to kill then to collect the reward.

The Butterfly effect (small input result in a large output) [3]
Small changes in the environment combine with other small changes which result in a hurricane. The point being that very small things can result in major outcomes, but its not predictable and the same small changes might not achieve the change in a different context.

The Butterfly Effect means that we cannot make a method or framework, just because 2-3 cases from other companies were successful, since most probably it was a unique combination that will not happen again, not to say all companies that made exactly the same but failed [4].

The Hawthorne effect (the observer/novelty changes output) [3]
Which argues that humans respond well to novelty, but you should not confuse the novel thing with novelty in respect of cause and effect.

So, what is important to emphasize is that there are of course cause and effects in live systems in the Complex domain too, but here there are no linear cause and effect chains, which is very important to understand. Butterfly effect is a good example, where a very small and known cause can generate an enormous effect, but where the effect always is unknown. This means that since system/organisation definitions are not live systems as discussed above, it also means that they are never in the Complex domain, meaning that system definitions and problems regarding it are not under the laws of complexity theory.

This leads to that we have cause and effect chains for products when our system definition is correct, but we by mistake put the wrong component, we are then never entering the Complex domain. But, this also leads to that we have cause and effect chains when our system definition is incorrect, since the components we put in the system cannot fulfil the definition, non-existing components. This we would never do on purpose for our products, since if we fail, we will enter the Complex domain, needing to do experimentation hoping to find a solution and rewrite the system definition. But, for our organisations today, this is common, we indirectly build organisations, that our people (agents) in the system cannot fulfil, i.e. we are violating our organisation definition. The only solution is to rebuild the organisation to fulfil the organisation definition (dissolution of problems), since our people’s cognitive abilities cannot be changed. The good thing on the other hand is that we never pass the Complex domain, since we do not need to do experimentation, we just need to ask why to find the root cause(s) to our organisational problems, the solution is already there. This also means that we are avoiding situational assessments which are not taking care of the organisational problems, and therefore only seeing symptoms, that we no are impossible to solve, or impossible to make continuous improvement on.

A common mistake when the word CAUSE is used, is to mix correlation with causation [5], where 10 or 15 cases are studied and from that is drawing a correlation, even though there is no causation, which is a massive problem in 95% of the management text books. Unfortunately many Agile methodologies is built on this, since the authors are looking for aspects of cases that supports their methodology, “you see only what you expect to see” or confirmation bias, which is not research mentality [6].

As you probably already have figured out, cause and effect can have different “values” like known and unknown, where we also can add a more diffuse “value” like knowable. But that will be the blog post for tomorrow, where we will elaborate on the meaning of “unknown unknowns” and its friends, depending on if we know we have the knowledge or not, and their respective domain affiliation.

C u then.


*In practical terms, this means that a small perturbation** may cause a large effect, a proportional effect, or even no effect at all. In linear systems, effect is always directly proportional to cause. – Wikipedia

**A deviation of a system, moving object, or process from its regular or normal state or path, caused by an outside influence. – Oxford Living dictionaries

***the issue that can be raised here is; what happens if we take away all aspects that differ us humans from animals, like motivation, attitude, intention, identity etc., and only follow rules within our genetics, would it still be the same complex human adaptive system, or something else? something less complex? something less adaptive? a more “termite nest” like system? Definitely something to think about.

[1] Smalley, Art. Four types of problems: from reactive troubleshooting to creative innovation. Lean Enterprise Institute (2018). ISBN 978-1934109557.

[2] Snowden, Dave. Blog posts. Links copied 2019-06-04.

[3] Snowden, Dave. Blog post. Link copied 2019-06-04.

[4] Snowden, Dave. “From Agile to Agility”. Link copied 2019-08-04. at 43:35 minutes

[5] Snowden, Dave. “From Agile to Agility”. Link copied 2019-08-04. at 45:30 minutes

[6] Snowden, Dave. Blog post. Text and link copied 2019-07-29.
“Multiple causes or rather dispositional states with varying modulators are key to understanding complex systems no linear models of single point causality.”