Lights, camera…and ACTION!

Clapper boardMy last post explained the thinking behind the softening of systems thinking – to include the reality of human beings into the mix.

I ended by noting that this naturally leads on to the hugely important question of how interventions into social systems (i.e. attempts at improving them) should be approached

What’s the difference between…?

The word ‘Science’ is a big one! It breaks down into several major branches, which are often set out as the:

  • Natural sciences – the study of natural phenomena;
  • Formal sciences – the study of Mathematics and Logic; and
  • Social sciences – the study of human behaviour, and social patterns.

Natural science can be further broken down into the familiar fields of the Physical sciences (Physics, Chemistry, Earth Science and Astronomy) and the Life sciences (a.k.a Biology).

The aim of scientists working in the natural science domain is to uncover and explain the rules that govern the Universe, and this is done by applying the scientific method (using experimentation1) to their research.

The key to any and every advancement in the Natural sciences is that an experiment that has supposedly added to our ‘body of knowledge’ (i.e. found out something new) must be:

  • Repeatable – you could do it again (and again and again) and get the same result; and
  • Reproduceable – someone else could carry out your method and arrive at the same findings.

This explains why all ‘good science’ must have been subjected to peer review – i.e. robust review by several independent and objective experts in the field in question.

“Erm, okay…thanks for the ‘lecture’…but so what?!”

Well, Social science is different. It involves humans and, as such, is complex.

The Natural science approach to learning (e.g. to set up a hypothesis and then test it experimentally) doesn’t transfer well to the immensely rich and varied reality of humanity.

“In [social science] research you accept the great difficulty of ‘scientific’ experimental work in human situations, since each human situation is not only unique, but changes through time and exhibits multiple conflicting worldviews.” (Checkland)

I’ll try to explain the enormity of this distinction between natural and social scientific learning with some examples, and these will necessarily return to those repeatability and reproducibility tests:

sodium into waterUnique: I’ll start with Chemistry. If you were to line up two beakers of water and (carefully) drop a small piece of sodium into each then you would observe the same explosive reaction…and, even though you could predict what would happen if you did it a third time, you’d still like to do it again 🙂

I was looking for a ‘social’ comparison and, following a comedy coffee conversation with a fellow Dad, the following observation arose: If you are a parent of two or more children, then you’ll know that consistency along the lines of ‘sodium into water’ is a pipe dream. I’ve got two teenage sons (currently 17 and 15 years old) and whenever I think I’ve learned something from bringing up the first one, it usually (and rather quickly) turns out to be mostly the opposite for the second! They are certainly unique.

The same goes for group dynamics – two different groups of people will act and react in different ways…which you won’t be able to fully determine up-front – it will emerge.

Electric circuitChanging over time: Now, over to Physics. If you were to select a  battery, light bulb and resistor combination and then connect them together with cables in a defined pattern (e.g. in series) then you could work out (using good old Ohm’s law) what will happen within the circuit that you’ve just created. Then, you could take it all apart and put it away, safe in the knowledge that it would work in the same predictable way when you got it all out the next time.

However, in our ‘social’ comparison, you can’t expect to do the same with people…because each time you (attempt to) do something to/with them, they change. They attain new interactions, experiences, knowledge and opinions. This means that it is far too simplistic to suggest that “we can always just undo it if we want to” when we are referring to social situations.

Just about every sci-fi movie recognises this fact and comes up with some ingenious device to ‘wipe people’s minds’ such that they conveniently forget what just happened to them – their memories are rewound to a defined point earlier in time. The ‘Men in Black’ use a wand with a bright red light on the end (hence their protective sunglasses…or is that also fashion?).

In reality, because such devices don’t exist (that I’m aware of), people in most organisations suffer from (what I refer to as) ‘Change fatigue’ – they’ve become wary of (what they’ve come to think of as) the current corporate ‘silver bullet’, and act accordingly. This understandably frustrates ‘management’ who often don’t want to see/ understand the fatigue2 and respond with speeches along the lines of “Now, just wipe the past from your mind – pretend none of it happened – and this time around, I want you to act like it’s really worth throwing yourself into 110%!”.

Mmmm, if only they could!

scalpelsConflicting worldviews: Finally, a Biology example. Let’s suppose that you’ve done a couple of lung dissections – one’s pink and spongy, the other is a black oogy mess3. Everyone agrees which one belonged to the 40-a-day-for-life smoker.

However, for our ‘social’ comparison, get a bunch of people into a room and ask them for their opinions on other people and their actions, and you will get wildly differing points of view – just ask a split jury!

The social phenomena of the ‘facts’ are subject to multiple, and changing, interpretations.

After reading the above you might be thinking…

“….so how on earth can we learn when people are involved?”

Kurt LewinThis is where I bring in the foundational work of the psychologist Kurt Lewin (1890 – 1947).

Lewin realised the important difference between natural and social science and came up with a prototype for social research, which he labelled as ‘Action research’.

“The method that [Lewin] evolved was of involving his subjects as active, inquiring participants in the conduct of social experiments about themselves.” (Argyris & Schon)

His reasoning was that:

“People are more likely to accept and act on research findings if they helped to design the research and participate in the gathering and analysis of data.” (Lewin & Grabbe)

Yep, as a fellow human being, I’d wholeheartedly agree with that!

Who’s doing the research?

I hope that you can see that the (potentially grand) title of ‘social research’ doesn’t presume a group of people in white lab coats attached to a University or such like. Rather, applied social research can (and should) be happening every minute of every day within your organisation – it does at Toyota!

Argyris and Schon wrote about two (divergent) methods of attempting to intervene in an organisation. They labelled these as:

  • ‘Spectator – Manipulator’: a distant observer who keeps themselves at arms lengths from the worker, yet frequently disturbs the work with ‘experiments’ to manipulate the environment and observe the response;


  • ‘Agent – Experient’: an actor who locates themselves within the problematic situation (with the people), to appreciate and be guided by it, to facilitate change (in actions and thinking) by better understanding of the situation.

You can see that the first fits well with natural science whilst the second fits with social.

‘The ‘spectator – manipulator’ method also describes rather well the reality of commanding and controlling, through attempting to implement (supposed) ‘best practise’ on people, and then rolling out ever wider.

The nice thing about action research is that the researcher (the agent) and the practitioner (the people doing the work) participate together, meaning that:

“The divide between practitioner and researcher is thus closed down. The two roles become one. All involved are co-workers, co-researchers and co-authors…of the output.” (Flood)

Proper4 action research dissolves the barrier between researcher and participant.

And, as such, ‘Action research’ is now often relabelled as ‘Action learning’…. because that is exactly what the participants are doing.

A note on intervention

InterventionAny intervention into a social system causes change5. Further, the interventionist cannot be ‘separated from the system’ – they will change too!

Argyris and Schon wrote that:

“An inquiry into an actor’s reasons for acting in a certain way is itself an intervention…[which] can and do have powerful effects on the ways in which both inquirer and informant construe the meaning of their interaction, interpret each other’s messages, act towards each other, and perceive each other’s actions. These effects can complicate and often subvert the inquirer’s quest for valid information.

Organisational inquiry is almost inevitably a political process…the attempt to uncover the causes of a systems failure is inevitably a perceived test of loyalty to one’s subgroup and an opportunity to allocate blame or credit…

[We thus focus on] the problem of creating conditions for collaborative inquiry in which people in organisations function as co-researchers rather than as merely subjects.”

You might think that taking a ‘spectator – manipulator’ approach (i.e. remaining distant) removes the problem of unintended consequences from intervening…but this would be the opposite. The more remote you keep yourself then the more concerned the workers will likely be about your motives and intentions….and the less open and expansive their assistance is likely to be.

So, as Argyris and Schon wrote, the best thing for meaningful learning to occur would be to create an appropriate environment – and that would mean gaining people’s trust….and we are back at action learning.

The stages of action research within an organisation

Action research might be described as having three stages6, which are repeated indefinitely. These are:

  1. Discovery;
  2. Measurable action; and
  3. Reflection

Discovery means to study your system, to find out what is really happening, and to drive down to root cause – from events, through patterns of behaviour, to the actual structure of the system (i.e. what fundamentally makes it operate as it does) …and at this point you are likely to be dealing with people’s beliefs.

…and to be crystal clear: the people doing the discovery are not some central corporate function or consultants ‘coming in’ – it’s the people (and perhaps a skilled facilitator) who are working in the system.

Measurable action means to use what you have discovered and, together, take some deliberate experimental action that you (through consensus) believe will move you towards your purpose.

But we aren’t talking about conventional measurement. We’re referring to ‘the right measures, measured right’!

Reflection means to consider what happened, looking from all points of view, and consider the learning within…. leading on to the next loop – starting again with discovery.

This requires an environment that ensures that open and honest reflection will occur. That’s an easy sentence to write, but a much harder thing to achieve – it requires the dismantling of many conventional management instruments. I’m not going to list them – you would need to find them for yourselves…which will only happen once you start your discovery journey.

What it isn’t

Action research isn’t ‘a project’; something to be implemented; best practise; something to be ‘standardised’…(carry on with a list of conventional thinking).

If you want individuals, and the organisation itself, to meaningfully learn then ‘commanding and controlling’ won’t deliver what you desire.

…and finally: A big caveat

warning trianglePeter Checkland adopted action research as the method within his Soft Systems Methodology (SSM) and yet he was highly critical of “the now extensive and rapidly growing literature” on the approach, calling it “poverty stricken”. Here’s why:

“The great issue with action research is obvious: what is its truth criterion? It cannot be the repeatability of natural science, for no human (social) situation ever exactly duplicates another such situation.” (Checkland)

The risk of simply saying “we’re doing action research” is that any account of what you achieved becomes nothing more than plausible story telling. Whilst social research can never be as solid as the repeatable and reproduceable natural science equivalent, there must be an ‘is it reasonable?’ test on the outcomes for it to be meaningful.

There’s already plenty of ‘narrative fallacy’ story telling done within organisations – where virtually every outcome is explained away in a “didn’t we do well” style.

To be able to judge outcomes from action research, Checkland argues that an advanced declaration is required of “what constitutes knowledge about the situation. This helps to draw the distinction between research and novel writing.”

This makes the action research recoverable by anyone interested in subjecting the work to critical scrutiny.

So what does that mean? Well, taking John Seddon’s Vanguard Method7 as an example, the Check stage specifically starts up front with:

  1. Defining the purpose of the system (from the customer’s perspective – ‘outside in’);
  2. Understanding the demands being placed on the system (and so appreciating value from a wide variety of customer points of view); and
  3. Setting out a set of capability measures that would objectively determine whether any subsequent interventions have moved the system towards its purpose.

…and (in meeting Checkland’s point) this is done BEFORE anyone runs off to map any processes etc.

In summary:

We need to appreciate “the role of surprise as a stimulus to new ways of thinking and acting.” (Argyris & Schon).

People should be discovering, doing and seeing for themselves, which will create a learning system.


1. Experiments: If you’d like a clearer understanding of experiments, and some comment on their validity then I wrote about this in a very early post called Shonky Experiments

2. Not wanting to see the change fatigue: This would happen if a manager is feverishly working towards a ‘SMART’ KPI, where this would be exacerbated if there is a bonus attached.

3. Lung dissection: I was searching around for an image of a healthy and then a smoker’s lungs…but thought that not everyone would like to see it for real…so I’ve put up an image with a collection of surgical scalpels – you can imagine for yourself 🙂

4. Proper: see the ‘big caveat’ at the end of the post.

5. Such changes may or may not be intended, and may be considered as positive, negative or benign.

6. Three stages of action research: If I look at the likes of Toyota’s Improvement/Coaching Katas or John Seddon’s Vanguard Method then these three stages can be seen as existing within.

7. The Vanguard Method is based on the foundation of action learning.


Double Trouble

Double troubleThere’s a lovely idea which I’ve known about for some time but which I haven’t yet written about.

The reason for my sluggishness is that the idea sounds so simple…but (as is often the case) there’s a lot more to it. It’s going to ‘mess with my head’ trying to explain – but here goes:

[‘Heads up’: This is one of my long posts]

Learning through feedback

We learn when we (properly) test out a theory, and (appropriately) reflect on what the application of the theory is telling us i.e. we need to test our beliefs against data.

“Theory by itself teaches nothing. Application by itself teaches nothing. Learning is the result of dynamic interplay between the two.” (Scholtes)

Great. So far, so good.

Single-loop learning vs. Double-loop learning

Chris Argyris (1923 – 2013) clarified that there are two levels to this learning, which he explained through the phrases ‘single-loop’ and double-loop’1.

Here are his definitions to start with:

Single-loop learning: learning that changes strategies of action (i.e. the how) …in ways that leave the values of a theory of action unchanged (i.e. the why)

Double-loop learning: learning that results in a change in the values of theory-in-use (i.e. the why), as well as in its strategies and assumptions (i.e. the how).

That’s a bit of a mouthful – and (with no disrespect meant) not much easier to comprehend when you read his book!2

If you look up ‘double loop learning’ on the wonders of Google Images, you will find dozens of (very similar) diagrams3, showing a visualisation of what Argyris was getting at.

Here’s my version4 of such a diagram:

Double loop 1

You can think about this diagram as it relates either to an individual (e.g. yourself) or at an organisational level (how you all work together).

Start at the box on the left. Whether we like it or not, we (at a given point in time) think in a certain way. This thinking comes about from our current beliefs and assumptions about the world (and, for some, what might lie beyond).

Our thinking guides our actions (what we do), and these actions heavily influence5 our performance (what we get).

And so to the ‘error’ bit:

“Organisations [are] continually engaged in transactions with their environments [and, as such] regularly carry out inquiry that takes the form of detection and correction of error.” (Argyris & Schon)

We are continually observing, and inquiring into, our current outcomes – asking ourselves whether we are ‘on track’, or everything is ‘as we would expect’ or perhaps whether we could do better. Such inquiry might range from:

  • subconscious and unstructured (e.g. just part of daily work); right through to
  • deliberate and formal (such as a major review producing a big fat report).

Argyris labels this constant inquiry as the ‘detection of error’. The error is that we aren’t where we would want to be, and the correction is to do something about this.

Okay, so we’ve detected an error and we want to make a corrective change. The easiest thing to do is to revisit our actions (and the strategies that they are derived from), and assess and develop new action strategies whilst keeping our underlying thinking (our beliefs and assumptions) steadfastly constant. This is ‘single-loop’ learning i.e. new actions, borne from the same thinking.

I reflect that the phrase ‘the more things change, the more they stay the same’ fits nicely here:

If the reason for the ‘error’ is within your thinking, then your single-loop learning, and the resultant change, won’t work. Worse, you will re-observe that error as it ‘comes round again’, and probably quicker this time…and so you make another ‘action’ change….and that error keeps on coming around. You have merely been making changes within the system, rather than changing the system.

A previous post called ‘making a wrong thing righter’ demonstrates this loop through the example of short term incentive schemes, and their constant revision “to make them even better”.

So, the final piece of the diagram…that green line. Many ‘errors’ will only be corrected through inquiry into, and modification of, our thinking…and, if this meaningfully occurs, then this would result in ‘double-loop’ learning – you would have changed the system itself.

Right, so that’s me finished explaining the difference between single-loop and double-loop learning…which I hope is clear and makes sense.

You may now be thinking “great, let’s do double-loop learning from now on!”

…because this is how most (if not all) those Google Image diagrams make it look. I mean, now you know about it, why wouldn’t you?

But you can’t!

The bit that’s missing…

Unfortunately, there’s a wall. Worse still, this wall is (currently) invisible. Here’s the diagram again, but altered accordingly:

Double loop 2

Right, I’d better try and explain that wall. Argyris & Schon wrote that:

“People learn collectively to maintain patterns of thoughts and action that inhibit productive learning.”

What are they on about?

Imagine that, through some form of inquiry, an error (as explained above) has been detected and a team of relevant people commence a conversation to talk about it:

  • The hierarchically senior person begins with a ‘take charge’ attitude (assuming responsibility, being persuasive, appealing to larger pre-existing goals);
    • it is typical within organisations that, once goals have been decided, changing them is seen as a sign of weakness.

  • He/she request a ‘constructive dialogue’, thereby stifling the expression of negative (yet real) feelings by themselves, and by everyone else involved…and yet acts as if this is not happening;
    • each person in the group is therefore being asked to suppress their feelings – to experience them privately, censor them from the group, and act as if they are not doing so.”

  • He/she takes a rational approach and asks the group to develop a ‘credible plan’ (which becomes the objective) to respond to the error…and so has skipped the necessary organisational self-reflection for double-loop learning to occur.
    • Coming up with a plan is ‘jumping into solution mode’ before you’ve properly studied the current condition and asked ‘why’.

So how does this affect the group dynamics?

“The participants experience an interest in solving the business problem, but their ways of crafting their conversation, combined with their self-censorship, [will lead] to a dialogue that [is] defensive and self-reinforcing.” (Argyris & Schon)

Given that this approach will hide so much, we can expect lots of private conversations (pre-meetings to prepare for meetings, post-meetings about what was/wasn’t said in meetings, meetings about what meetings aren’t happening…). Does this describe what you sometimes see in your organisation? I think that it is often labelled as ‘politics’…. which would be evidence of that wall.

Taken together, Argyris and Schon label the above as primary inhibitory loops.

Argyris sets out a (non-exhaustive) list of conditions that trigger and, in turn, reinforce, such defensive and dysfunctional behaviour. Here’s the list of conditions, together with how they should be combated:

Condition Corrective response
Vagueness Specify
Ambiguity Clarify
Un-test-ability Make testable
Scattered Information Concert (arrange, co-ordinate)
Information withheld Reveal
Un-discuss-ability Make discussable
Uncertainty Inquire
Inconsistency/ Incompatibility Resolve

”[such] conditions…trigger defensive reactions…these reactions, in turn, reduce the likelihood that individuals will engage in the kind of organisational inquiry that leads to productive learning outcomes.” (Argyris & Schon)

i.e. If you’ve got defensive behaviour, look for these conditions… and work on correcting them. Otherwise you will remain stuck.

Unfortunately, primary loops lead to secondary inhibitory loops. That is, that they lead to second-order consequences, and these become self-reinforcing.

  • Managers begin to (privately) judge their staff poorly, whilst the staff, ahem, ‘return the compliment’, with “both views becoming embedded in the organisational norms that govern relationships between line and staff”;

  • Sensitive issues of inter-group conflict become undiscussable. “Each group sees the other as unmovable, and both see the problem as un-correctable.”
    • A classic example of this is the constant conflict in many organisations between ‘IT’ and ‘The business’.

  • The organisation creates defensive routines intended to protect individuals from experiencing embarrassment or threat”with the unintended side effect that this then prevents the identification of the causes of the embarrassment or threat in order to correct the relevant problems.”

From this we get organisational messages that:

  • are inconsistent (in themselves and/or with other messages);
  • act as if there is no inconsistency; and
  • make the inconsistency undiscussable.

“The message is made undiscussable by the very naturalness with which it is delivered and by the absence of any invitation or disposition to inquire about it.”

Do you receive regular messages from, say, those ‘above you’ in the hierarchy? Perhaps a weekly or monthly Senior Manager communication?

  • How often are you amazed (in an incredulous way) about what they have written or said?
  • Do you feel welcome to point this inconsistency out? Probably not.

We end up with people giving others advice to reinforce the status quo: ‘Be careful what you say’, ‘You’ll get yourself into trouble’, ‘I wouldn’t say that if I were you’, ‘Remember what happened last time’…etc.

In short, there are powerful forces* at work in most organisations that are preventing (or at least seriously impeding) productive learning from taking place, despite the ability and intrinsic desire of those within the organisation to do so.

(* Note: Budgets – as in fixed performance contracts – are a classic ‘single-loop reinforcing’ management instrument. Conversely, Rolling forecasts can be a ‘double-loop’ enabler.)

So what to do instead?

Right, here’s my third (and last) diagram:


Double loop 3

It looks very similar to the last diagram, but this time there’s a ladder! But where do we get one of those from?

“For double-loop learning to occur and persist at any level in the organisation, the self-fuelling processes must be interrupted. In order to interrupt these processes, individual theories-in-use [how we think] must be altered.” (Argyris & Schon)

Oooh, exciting stuff! They go on to write:

“An organisation with a [defensive] learning system is highly unlikely to learn to alter its governing variables, norms and assumptions [i.e. thinking] because this would require organisational inquiry into double-loop issues, and [defensive] systems militate against this…we will have to create a new learning system as a rare event.”

There’s two places to go from here:

  • What would a productive learning system look like? and
  • How might we jolt the system to see the wall, and then attempt to climb the ladder?

If I can begin to tease these two out, then BINGO, this blog post is ready for print. Right, nearly there…

A Productive learning system

Argyris and Schon identify three values necessary for a productive learning system:

  • Valid information;
  • Free and informed choice; and
  • Internal commitment to the choice, including constant monitoring of its implementation.

Sounds lovely…but such a learning system requires the fundamental altering of conventional social virtues that have been taught to us since early in our lives. The following table ‘compares and contrasts’ the conventional with the productive:

Social Virtue: Instead of… Work towards…
Help and Support Giving approval and praise to others, and protecting their feelings Increasing others capacity to confront their own ideas, and to face what they might find.
Respect for others Deferring to others, and avoiding confronting their actions and reasoning. Attributing to others the capacity for self-reflection and self-examination.
Strength Advocating your position in order to ‘win’, and holding firm in the face of advocacy. Advocating your position, whilst encouraging inquiry of it and self-reflection.
Honesty Not telling lies, or

(the opposite) telling others all you think and feel.

Encouraging yourself and others, to reveal what they know yet fear to say. Minimising distortion and cover-up.
Integrity Sticking to your principles, values and beliefs Advocating them in a way that invites enquiry into them. Encouraging others to do likewise.

There’s a HUGE difference between the two.

The consequences will be an enhancement of the conditions necessary for double-loop learning – with current thinking being surfaced, publicly confronted, tested and restructured – and therefore increasing long-term effectiveness.

You’d likely liberate6 a bunch of great people, and create a purpose-seeking organisation.


The first task is for you to see yourself – you have to become aware of the wall…and Argyris & Schon are suggesting that you may (likely) require an intervention (a shake) to do this. Your current defensive learning system is getting in the way.

Let’s be clear on what would make a successful intervention possible, and what would not.

An interventionist would locate themselves in your system and help you (properly) see yourselves…and coach you through contemplating what you see and the new questions that you are now asking…and facilitate you through experimenting with your new thinking and making this the ‘new normal’. This is ‘action learning’.

This ‘new normal’ isn’t version 2 of your current system. It would be a different type of system – one that thinks differently.

Conversely, you will not change the nature of your system if you attempt to ‘get someone in to do it to you’.

Why not?

“Kurt Lewin pointed out many years ago that people are more likely to accept and act on research findings if they helped to design the research, and participate in the gathering and analysis of data.

The method he evolved was that of involving his subjects as active, inquiring participants in the conduct of social experiments about themselves.” (Argyris & Schon)

In short: It can’t be done to you.

That ladder? That would be a skilled interventionist, helping you see and change yourselves through ‘action learning’.

To Close

Next time someone shows you that lovely (as in ‘simple’) double-loop learning diagram, I hope you can tell them about the wall…and the ladder.


1. Chris Argyris is known as one of the co-founders of ‘Organisation Development’ (OD) – the study of successful organizational change and performance. Argyris notes that he borrowed the distinction between single-loop and double-loop from the work of W. Ross Ashby. For blog readers, we met Ashby in an earlier post on requisite variety.

2. Book: ‘Organisational Learning II: Theory, Method, and Practise’ (1996) by Chris Argyris and Donald A. Schon).

3. Diagrams: Many of the diagrams stay true to what Argyris wrote about. Some attempt to build upon it. Others (in my view) bastardise it completely!

4. Language: I should note that Argyris used different language to my diagram. Here’s a table that compares:

My diagram: Argyris and Schon:
Thinking (Our beliefs and assumptions) Values, norms and assumptions
Action Action Strategies
Performance Performance, effectiveness
Defensive learning system Model O – I
Productive learning system Model O – II

5. Influence: I haven’t used the bolder ‘cause’ word because there’s a lot going on that is outside the system (e.g. the external environment).

6. Liberate: You don’t need to bring in ‘new’ people, most of what you need are already with you – they just need liberating from the system that they work within.

7. Kurt Lewin: often referred to as ‘the founder of social psychology’. Much of my writings in this blog are based around Lewin’s equation that Bƒ(PE) or, in plain English, that behaviour is a function of the person in their environment.