[Full link at bottom – paper is not paywalled]
I interrupt my scheduled schedule to immediately review this new paper from Sandra Janoff. If you know me, you’ll know that this paper, and its lovely references, is like crack to me. I also genuinely believe this is going to be useful, important, and foundational, though its pragmatism and light touch may belie that. These are first thoughts, and somewhat fragmentary.
My summary: systems don’t develop because someone explains complexity better. They develop when enough of the relevant system is brought into contact with itself, so its real differences can be seen, held and worked with. The point is not agreement, alignment, or nicer collaboration, but building the capacity to act together before agreement exists.
Official abstract: Whole-System Development focuses on how systems build the capacity to act together on complex challenges. Drawing on open systems thinking and decades of practice, this paper introduces Differentiation/Integration (D/I) as the developmental logic through which systems increase their capacity to learn and act. While many approaches emphasize dialog, alignment, or individual change, this work highlights how the system is brought together, who is present, and what they do. When the relevant parts of a system engage directly, differences become visible and workable rather than divisive. Through the interplay of differentiation and integration, systems can hold multiple perspectives, recognize shared realities, and take coordinated action without requiring agreement. Cases drawn from Future Search practice in business, post-conflict settings, and community development show how this capacity emerges. The paper positions Whole-System Development as a contribution to applied behavioral science and a basis for further inquiry and practice.
Since I’ve said I’m partial, let me test this by my own tests. I say ‘good work’ has a model, intervention method, rich stories with examples of failure, and a learning orientation; and that I’m suspicious of arm-waving, prescriptions, neat solutions and historicist claims.
The paper meets these requirements pretty well. There is a model, there is method, and there is practice. It avoids being merely arm-waving. It does not present as a theory of everything or a panacea. It has a learning orientation. The obvious gap is that the stories of failure are not here, though they are elsewhere in the Future Search literature. That matters because practice literature is often strongest when its failure modes are most explicit: when it says not only ‘this worked’ but ‘this is how we got it wrong, this is where it breaks, and this is what we learned’. I’d still say this paper passes the test, but that is the lacuna I’d mark most clearly.
There is maybe a little bit of a strong claim of solutions, and an organicist if not historicist introduction. If you take the foundational claims as absolutely true and derived from nature, you might scratch your chin. If you take them as useful, practical, and born out in experience, I think you’ll get the value.
This is about Future Search not as just another workshop recipe, but as a theory of conditions.
‘You cannot integrate what you have not yet differentiated’ rings out. In practice, competing realities aren’t argued away. They are brought in to a shared field, differentiated, held, and then organised around common ground sufficient for action. But / and I think it may in fact be the holding of paradox for long enough for a deutero-learning response to be possible. This is compatible because it’s the conditions which enable this.
For me, the heart of this is that Janoff argues that systems develop through the ongoing interplay of differentiation and integration. Development is not ‘more growth’ or ‘more change’. It is increased capacity to address challenges the system could not previously resolve.
Differentiation means bringing forward the differences that matter: role, experience, power, knowledge, identity, responsibility, interest, lived reality – or whatever. Integration means working with those differences without collapsing them into bland agreement. This is the bit that matters. Systems do not need everyone to agree before they act. In complex public service work, for example, if we wait for agreement, we wait forever. What we need is enough shared reality, enough common ground, enough trust, enough clarity, and enough responsibility to act together while continuing to hold real differences.
That is much more demanding than ‘collaboration’, and much more relevant in the real world.
This connects powerfully to Stafford Beer and cybernetic viability, though that tradition is not referenced. ‘Whole system in the room’ can be read as temporarily bringing environment, operations and metasystem into the same learning space. It’s Ashby in human-process garb: variety not just represented, but enacted. And it’s Barry Oshry’s ‘robust systems’, where the Dominant may learn to engage with, appreciate and value the Other in the self-interest of being more robust and more capable of dealing with future challenges.
In this paper and Sandra’s work, ‘whole system’ is not abstract. It means the right sensemaking worlds are present. Not through a stakeholder map as a performance of inclusion, but through careful discernment and sampling of different perspectives, roles, groupings and lived positions.
Facilitator capacity, capability and expertise are not foregrounded here, but they are important. As I’ve been reminded before, it’s the care-filled preparatory work which can enable this perspective to work. The capacity to hold differentiation and integration does not magically arise from invitation design. There is craft, judgement, container-building, boundary work, pacing, courage, humility, and probably luck.
We are left with the practical challenge, regardless of how good this paper is. Who needs to be in the room? Who is missing? Who is present but not powerful? Which differences are being surfaced? What’s being carefully avoided? What common ground is real enough to act on? What structures will keep the learning, the differentiation and integration, alive after the event? Are we better equipped to constructively continue that roiling, energising process within this ‘system’ and in engagement with ‘others’?
Those last two are really critical.
The goal is carefully not idealised: ‘capacity to act together’. I think this is a very useful tag. It’s less warm than ‘relational’, less policy-ish than ‘collaboration’, and less vague than ‘systems leadership’.
It’s clear that the differences and conflicts are not the problem and, critically, because this takes you down dark pathways, are not things to be resolved. The lack of conditions for working with difference is the problem. That, I think, is a better sentence than 80% of the systems change literature. In the IKEA case study given, there is a successful non-resolution of the conflict between profit and sustainability, by integration as interdependent responsibilities.
The point is not to tell people they are part of a system. It is to create the conditions in which the system can encounter itself: the people with authority, resources, expertise, information, and lived experience, working with the whole before rushing to act on the parts.
For me, this connects directly to public services.
In my framing, the citizen world, service world, management world, political world, and the world of learning and change all see different realities. None of them has the whole elephant, and seeing a picture someone, or even we, have made of the whole elephant doesn’t help much. Too often, one world’s partial view, or even collaborative systems mapping, gets turned into a plan, a target, a service specification, or a transformation programme.
You can see this in integrated care, children’s safeguarding, homelessness, place-based prevention, relational public services, and all the other places where everyone already knows, in principle, that no single agency can solve the problem. The question is not whether the system is complex. The question is whether the relevant parts of the system can learn enough together to act differently, without needing the impossible comfort of total agreement.
Whole-system work isn’t workshop theatre, ‘getting buy-in’, or a nice participative wrapper around decisions already made. Done properly, it is an intervention into the system’s capacity to see, learn, decide and act through expressing difference and working with it.
And it connects very naturally to other things I’ve found really valuable: the Centre for Creative Leadership boundary work, systems convening from the Wenger-Trayners, and, interestingly, a topic that has been flickering on my radar recently, why lived experience, impairment, and dependent care can be seen as a gift not just a burden. Not gift in the sentimental or compensatory sense. Gift because these experiences disclose realities, dependencies, exclusions, adaptations and forms of knowledge that dominant systems are structured not to perceive.
But there is still a risk.
‘Whole system’ language can easily become whitewashing. ‘Whole system in the room’ is never literally true. It is a designed fiction. In practice it can be a very useful one, and hugely productive when done well. But it remains a fiction, and the fiction is designed by someone.
The key challenge is boundary judgement. Who says this is the whole system? Who is missing? Who cannot safely be present? Who is present but powerless? What does ‘common ground’ exclude? What is made undiscussable by the future focus?
This is where power has to be metabolised, not just named. Whole-system methods don’t escape power. They exercise power through invitation, framing, sequencing, facilitation, naming, recording, grouping, timing and follow-through. The design group makes boundary judgements, and like all exercise of power those are hopefully visible, contestable etc. Otherwise ‘whole system’ becomes another way for some parts of the system to define reality for the others, with refreshment breaks (that’s me not ChatGPT btw).
Second, ‘common ground’ can become a way of avoiding justice, history and power. Janoff handles this better than many, because she says difference remains. But in public services, especially where citizen voice, race, class, disability, trauma and poverty are involved, ‘focus on the future’ can be used to step around accountability. I think Petruska Clarkson’s intervention priority sequence is relevant here: danger, confusion, conflict, deficit and only then development.
Third, it perhaps needs a clearer account of recursion. A Future Search creates a temporary whole-system container. But systems need ongoing recursive structures for learning and action. Otherwise, the event becomes a beautiful island. Open Systems Theory supporters, and Merelyn Emery, whose partner Fred is another key source here, will say that this is only possible if the holding conditions of group work become the holding conditions for the whole. In any case, at the very least, ongoing practices and attending to differentiation and inclusion dynamics are needed.
I use four: grouping and individual empowerment as forms of differentiation moving towards power and adaptation; blending groups and unity of people as forms of integration moving towards love and stability.
Fourth, ‘capacity to act together’ needs measures, or at least observable signs. Not KPI sludge, but indicators: who now talks to whom, what decisions move faster, what conflicts become discussable, what resources move, what actions are taken without escalation, what citizens experience differently, what the system now senses that it previously ignored.
But these are not criticisms. Nobody said it was going to be easy. And without some move towards shared responsibility, little is possible. Well, not consensually, and that’s another story.
What this paper offers is a superbly well-founded framework and guide, tempered by experience. It gives us a clear and useful phrase for the work many of us are trying to do:
build the capacity to act together.
That is one of the best plain-English definitions of serious systems change I’ve seen for a while.
Systems Community of Inquiry 
You must be logged in to post a comment.