antlerboy – Benjamin P Taylor | Systems Community of Inquiry

Last week I shard this

A more useful fly-bottle – systems | cybernetics | complexity in six diagrams

…as soon as you do something like that, you begin to see the problems, have doubts, consider alternatives. What I produced came out in a way that mildly surprised me! So here’s my own critique:

My systems thinking needs to be open to more than agreement

I describe systems | cybernetics | complexity as purposeful engagement with context. That leaves some questions open. Who establishes the purpose? What counts as understanding the context? How do we know when our way of seeing is helping, and who gets to say it isn’t? Some of that is covered but it may not be explicit – it is certainly (and intentionally)

My account brings together framing, different worlds of meaning, work within boundaries, feedback, five core practices, and four dynamics of organisation. I find these connections useful. But putting them into a diagram gives them an appearance of completeness which is erroneous. Saying ‘this is provisional’ doesn’t do much unless I can explain what would lead me to change or abandon part of it.

There are substantial criticisms available from within the traditions I draw on. These are my applications of those criticisms to my own account, rather than reports of what the authors have said about it.

Seeing differently isn’t sufficient

Patrick Hoverstadt’s The Grammar of Systems directs attention towards how arrangements behave, including how connections transmit consequences and how adaptation is distributed. Patterns of Strategy, which he developed with Lucy Loh, examines how relationships between organisations shape their strategic direction. Reading my account against that work, I have to ask whether I’m explaining what produces an outcome or mostly offering another way to describe it. (Major Projects)

Suppose several organisations understand a citizen’s situation perfectly well. They know that their repeated assessments are burdensome. They agree that earlier support would help. But each controls a separate budget, carries different risks, and can improve its own financial position by transferring responsibility elsewhere.

They might already have enough shared understanding to explain why none of them intends to move.

Constructive conversations could expose that arrangement. They could help people negotiate a different one. But the conversation hasn’t changed the arrangement simply because everyone now understands it. Something has to happen to authority, resource flows, obligations, or the conditions under which people make decisions.

My response is to distinguish an orientation to practice from a diagnosis. The five practices aren’t a substitute for investigating the mechanisms maintaining a particular situation. They should help people undertake that investigation and act on it. If ‘clarity’ or ‘learning’ becomes my explanation for every success and its absence my explanation for every failure, I’ve made the framework difficult to contradict and correspondingly less useful.

I also need to resist an overcorrection. Acknowledging the limits of modelling doesn’t mean abandoning models. I want models that help us anticipate, test explanations, expose impossible assumptions, and compare possible actions. They don’t have to contain the whole situation to do useful work.

Nor should the fly-bottle metaphor smuggle back the idea that we inhabit private mental pictures, looking out at an inaccessible reality. That isn’t what I intend. Our ways of seeing are bound up with bodily activity, tools, relationships, institutions, and language. A budget boundary isn’t just someone’s interpretation: it affects who can spend the money. Reframing it might reveal a possibility, but changing it may require a decision, an agreement, and somebody accepting a cost.

Different worlds needn’t converge

Dave Snowden’s argument about ‘requisite conflict’ challenges the assumption that shared understanding and common purpose must precede effective action. He argues for retaining enough difference to explore alternatives without making the conflict destructive. People can agree that an idea is coherent enough to test without agreeing that it is right. (The Cynefin Co)

That is a useful challenge to my five-worlds picture. Citizen, service, management, leadership, and learning and change are ways I distinguish experiences and responsibilities. They aren’t five pieces which, once assembled, produce the complete picture. They aren’t internally uniform, either. There are differences and conflicts within citizen world, within service world, and within each of the others.

I don’t think translation requires consensus. Sometimes its achievement is that we can disagree more accurately. We may discover that what looked like a misunderstanding is a real conflict of interest. We may find a workable accommodation without acquiring a common worldview. We may decide that some activities should remain separate.

That also means withdrawing the claim that most failure is a failure of translation. I don’t have evidence for ‘most’. Translation failure is one mechanism worth investigating. So are inadequate resources, incompatible obligations, deliberate exclusion, and people understanding one another very well but wanting different things.

I still value shared meaning. I just need to be precise about how much is needed, for which activity. Agreeing how to hand over responsibility doesn’t require agreeing on a philosophy of public service. Equally, agreement on a philosophy doesn’t establish that the handover will work.

Complexity has to change the intervention

Some would also press the distinction between our knowledge of a situation and the nature of the processes we are dealing with. I agree ordered and complex conditions can coexist within an organisation and call for different ways of acting. Complexity isn’t simply the feeling of not yet understanding something, though I insist it does supervene on purpose, context, intent, understanding etc.

In any case, I accept the need to distinguish the conditions under which analysis, prediction, experimentation, or immediate action are appropriate. I don’t accept that recognising real constraints and patterns requires me to treat any classification of them as the final description. The classification also needs testing. The practical question is what changes in the work. When causal understanding is weak, I should be considering different, bounded interventions, looking for early consequences, and deciding in advance how to stop or adapt. Questions about what can be changed, what effects can be monitored, and where success can be amplified or failure contained are demanding in precisely that way. Calling a programme ‘adaptive’ while retaining one preferred solution, one success story, and no acceptable route to stopping it wouldn’t meet that test.

I need material explanations too. Suppose a service receives more cases than it can resolve or otherwise close. Its queue grows. A better conversation matters if it helps change those flows, the capacity supporting them, or the arrangements generating the demand. The queue doesn’t become less real because we have developed a richer account of it – but we might have better ideas of what to do about it.

Useful for whom?

Boundary critique makes a more uncomfortable demand than acknowledging that everyone has a perspective. Werner Ulrich points out that boundary judgements select both relevant facts and relevant values. An account is partial because it leaves things out, but also because it benefits some parties more than others. (W. Ulrich)

‘A more useful fly-bottle’ therefore needs an answer to ‘useful for whom?’ A description could help a management team reduce expenditure by making the work transferred to families invisible. It could help a commissioner demonstrate success by excluding people who never gain access to the service. My commitment to human agency doesn’t resolve those questions. I have to examine who currently defines benefit and who ought to; who controls resources; whose knowledge is recognised; and who can challenge the premises without first accepting my vocabulary.

There is a difference between being invited to contribute and being able to change a decision. Ulrich’s emancipatory use of boundary critique is especially relevant here: people affected by professional activity need ways to challenge its claims even when they don’t control the inquiry or possess the professionals’ specialist expertise. (W. Ulrich)

My response has to involve practical arrangements. Who can refuse? Who can appeal? Who can obtain independent advice? Who represents people absent from the discussion? Can someone challenge the purpose, rather than merely suggest a better way of achieving it?

Nor is a wider boundary automatically better. Privacy, independent scrutiny, and limits on authority may depend on maintaining boundaries. I need to distinguish a boundary which obstructs useful work from one which protects people against it.

Viability doesn’t establish legitimacy

Putting ‘viability’ in the centre of the diagram risks making continued existence look like sufficient justification for continued existence. I value the ability of an organisation to sustain worthwhile activity. But an organisation can be adept at preserving arrangements I would want to change. Its survival, its effectiveness, and its legitimacy are different questions.

For me, this is where an ethical commitment has to be stated rather than presented as a finding of systems science. I want public services to support people’s agency and their ability to live worthwhile lives. I don’t think a theorem can relieve me of responsibility for that preference, or settle every conflict arising from it.

Sometimes that commitment will mean strengthening an institution. Sometimes it may mean removing one of its functions, transferring authority, or making the institution unnecessary. I shouldn’t quietly rule out the latter possibilities by defining success as a more capable version of the existing organisation.

When I use the Viable System Model, I also need to specify what I’m examining. The viability of a team, a provider, and a wider arrangement of services aren’t interchangeable. A local improvement may export work or risk elsewhere. Drawing ‘governance’ inside a boundary doesn’t establish whether the people doing the work have the information, resources, and authority they need.

The same discipline applies to variety. I should identify which differences need a response and what responses are available, rather than assuming that richer information or more organisational diversity must help. Collecting a more detailed account of someone’s circumstances while permitting only the same inadequate response may increase frustration rather than capability.

The four dynamics aren’t an answer called ‘balance’

I use segmenting, blending, empowering, and harmonising to notice different activities: developing specialised capability, exchanging across groups, enabling individual contribution, and acting with regard to a wider purpose.

The objection is that I can diagnose any outcome retrospectively as ‘too much’ or ‘too little’ of one of them. Unless I say what I’m observing and what change I expect to help, ‘balance’ explains very little.

My response is that these are dynamics to investigate through activity and experience, not spaces into which people or teams belong. I’m interested in what happens over time. Who gets opportunities to develop expertise? Where do groups actually learn from one another? What discretion can individuals exercise? What happens when their judgement conflicts with a collective commitment?

A symmetrical picture shouldn’t imply equal quantities, a permanent equilibrium, or a compromise halfway between every apparent opposition. Sometimes protecting a dissenting group is more valuable than seeking immediate integration. Sometimes a common standard makes local discretion possible.

There is also an inventive challenge here: perhaps the apparent trade-off can be changed. Could we retain a shared standard while allowing different methods? Could a decision be delegated under clear conditions, with review triggered only when those conditions change? Could specialisation be preserved while removing a damaging handover?

I still find the four dynamics useful. But their value depends on the distinctions and interventions they enable, not on the reassurance of having all four represented.

The five practices could become five new demands

Constructive conversations, clarity, learning, culture shaping, and intent are intended to support work. They could also become an additional compliance layer.

Imagine an organisation requiring evidence that every team holds learning meetings, has completed its clarity templates, and has demonstrated alignment with purpose. Managers might become proficient at producing that evidence while retaining the same decisions, incentives, and constraints. Worse, the time used to demonstrate the practices could reduce the time available for the work they were meant to support.

That is a plausible failure mechanism for my approach. It isn’t adequately answered by saying that the practices were implemented badly. I need to consider how the design invites that implementation. I would look for practices embedded in consequential work. Does a conversation alter an agreement? Does clarification make a decision possible? Does learning change a method, a resource allocation, or a claim about what is happening? What existing activity can stop because the new practice is doing its job?

Culture shaping needs similar discipline. Mindset-based change can that it can treat the effects of people’s experience as if they were independent causes available for direct adjustment. It can also make people’s allegedly deficient attitudes the explanation for a programme’s failure. My account of culture is intended to attend to experience: what happens when someone raises a problem, exercises judgement, makes a mistake, or asks for help; the experience of leadership behaviours, organisational processes and systems, and symbolism and the core gut emotional reactions to them and to the stories told about them (after MacDonald et al and Carolien Taylor). But the label ‘culture shaping’ still risks suggesting that I know the evaluations other people ought to make. I can change arrangements and my own conduct. I can enquire into how those changes are experienced. I can’t legitimately require a favourable interpretation.

‘Constructive’ needs scrutiny too. A refusal, a complaint, or a public challenge may be necessary even when it disrupts the conversation I hoped to have. Surfacing the undiscussable cannot mean that every private experience must become available to management.

Learning isn’t a ladder everyone must climb

I use single-, double-, and triple-loop learning to distinguish changing actions, reconsidering governing assumptions, and reconsidering identity and the ways purposes and assumptions are formed.

I should be explicit that this is the interpretation I use. Tosey, Visser, and Saunders found several competing conceptualisations of triple-loop learning, established that the term doesn’t originate in Argyris and Schön’s published work, and cautioned against assuming that higher-order learning is necessarily preferable. (Sage Journals) That criticism matters in practice. Somebody may need a corrected calculation, a clearer task, or permission to make a decision. They shouldn’t have to undergo an identity transformation to obtain it.

I still want room for learning that changes how we understand our role. A service might reconsider whether its job is to process applications or help people obtain something they need. A consultant might reconsider whether their usefulness lies in providing answers or helping people develop their own capacity. Those changes can be substantial. But they shouldn’t be imposed as a test of maturity or commitment.

The invitation to question assumptions must include questioning the learning process itself. ‘Be autonomous, but arrive at my conclusion’ is not an adequate developmental contract.

I also need to qualify my preference for shorter learning cycles. Earlier detection of harm is valuable, but repeatedly intervening before the effects of previous actions are observable can make things worse. Meadows’ discussion of delays is a reminder that response time has to be considered relative to the processes being influenced. Faster correction isn’t automatically better correction. Sometimes learning means waiting deliberately, with safeguards and a reason for waiting. Sometimes it means maintaining an arrangement because the evidence supports it. I don’t want ‘learning’ to become a permanent demand to change.

The orchestrator can become the bottleneck

There is an Oshry-informed question for my description of leadership as orchestrating practices: am I giving the person already carrying responsibility for everything another elaborate account of why everything depends on them?

Suppose the leader convenes every difficult conversation, adjudicates every ambiguity, and keeps all the learning loops moving. The team may look coherent while becoming more dependent on that person.

My response is to judge the arrangements partly by what others can do without the orchestrator. Who can initiate a conversation? Who can make a consequential decision? Where can people compare their experiences directly? What happens when the leader is absent?

That applies particularly to consulting and facilitation. My presence changes what people say, what receives attention, and what counts as a useful contribution. I can’t remove that influence by announcing that I’m also ‘in the system’. I need to make it available for challenge and notice whether my involvement is increasing or reducing other people’s ability to act.

It also changes how I should think about resistance. An organisation might welcome all five practices while retaining the arrangements that make the old behaviour sensible. If local discretion is withdrawn after the first visible failure because accountability remained elsewhere, the problem may lie in that arrangement rather than in people’s willingness to learn.

Better coordination might be the wrong objective

There is a further objection to the amount of attention I give to translation, boundaries, and working together. Perhaps I’m helping people become better at operating an arrangement which should be redesigned.

Suppose a person has to explain the same situation to six services. Better information sharing might reduce the burden. But perhaps some of those assessments, eligibility distinctions, or handovers shouldn’t exist.

I need to allow for redesign which makes some coordination unnecessary. Otherwise ‘systems convening’ could become a permanent occupation repairing fragmentation that remains commercially or institutionally convenient for the organisations involved. It can help to temporarily stop describing the situation through organisations and roles. Follow the obligations instead. Who acquires a responsibility? Who can discharge it? Who can pass it on? Who is left waiting, and whose time disappears from the account?

That also tests my definition of work. Effort, judgement, discretion, boundaries, time, and purpose give me a useful way of discussing organisational work. Calling it a ‘primitive’ risks claiming more. Care, maintenance, recovery, play, and refusing an imposed purpose shouldn’t have to pass through that definition to matter.

A common vocabulary can conceal differences between methods

I want to work across systems traditions, but I shouldn’t make that sound like assembling compatible pieces of a single theory.

A model used to structure discussion about competing purposes is doing a different job from a model used to test a causal mechanism. A diagnosis of viability doesn’t decide legitimacy. An experiment that produces an encouraging result doesn’t establish why it worked or where else it will work.

The practical issue is the transition between inquiries. What has been established? What remains contested? Does the next step need more evidence, a political decision, a negotiation, a different design, or ordinary technical competence? Sometimes the appropriate response is to stop expanding the discussion. Repair the calculation. Remove the duplicate form. Clarify the delegation. Pay attention to the neglected maintenance.

My interest in reframing mustn’t make competent work inside an adequate frame appear intellectually inferior. Equally, urgency and technical competence mustn’t become reasons for leaving a damaging purpose unquestioned. Deciding between those possibilities remains work; no diagram removes it.

What would count against my account?

I need to separate things I too easily present together. My preference for human agency is an ethical commitment. The five core practices are a proposed repertoire of practice. The four dynamics are a way of investigating and shaping patterns. Claims about what an intervention will do need evidence and explanation appropriate to that situation.

I don’t accept that an introductory account must contain every method or answer every objection. But it should show where its own competence ends, and it mustn’t turn every failure into further evidence that people need more of the framework.

I would count it against the approach if people became more fluent in the language while less able to act; if maintaining its practices displaced worthwhile work; if apparent gains depended on hiding costs elsewhere; or if those affected found it harder to challenge a decision because it had acquired the authority of being ‘systemic’.

I also need to be careful about absorbing every criticism by saying that it was already included in what I meant. Some objections should change a claim. Some should restrict where I use an idea. Some may reveal disagreements I can’t reconcile.

The part I want to preserve is the possibility of developing enough understanding to act with others while remaining answerable for how we understand and what we do. That includes people choosing not to join my inquiry, using a different account, or making a criticism I can’t translate into my preferred language. They shouldn’t have to become systems thinkers before I take them seriously.