Why Fixing It Sometimes Breaks It Further

On the dynamics of repair — and what becomes possible when you understand them

This is the third of four articles co-authored by Bob Dick and Sebastian Wolf Siebzehnrübl, exploring the foundations of relational practice through the lens of Chris Argyris’s action science. The series builds toward a four-session webinar program, “Repairing Relationship,” offered in July 2026.

In the previous article we looked at the gap between the values we espouse and how we actually behave under pressure. Here we take that one step further — into the territory that Argyris found most surprising, and that Bob Dick has found most important in practice.

It is the territory of the well-intentioned intervention that makes things worse.

Not through carelessness. Not through lack of skill. But through a specific dynamic that most practitioners have never been taught to see — and that, once seen, changes everything about how you approach relational difficulty.

The cycle

Let us begin with the mechanism that Argyris mapped with unusual precision.

In any relationship, each person holds assumptions about the other — particularly about their motives and intentions. Those assumptions are almost never voiced. They are treated as facts. They drive behaviour. And the behaviour they generate tends to confirm the assumption that generated it — because the other person, observing the behaviour without knowing the assumption behind it, responds in a way that seems to validate the original reading.

A has an assumption about B’s motives. That assumption shapes A’s behaviour toward B. B, observing A’s behaviour and having no access to the assumption behind it, develops her own assumption about A. B’s behaviour in response shapes A’s assumptions further. The cycle turns.

What is crucial to understand is that this cycle is self-sealing. Not because either party is acting in bad faith. But because the information that would allow the assumptions to be tested — the actual beliefs, feelings, and intentions of each person — is almost never exchanged. Social taboos against expressing such information are powerful and largely invisible. They feel like tact, or professionalism, or appropriate restraint. They are, in structural terms, the mechanism by which misunderstanding compounds indefinitely.

When the cycle is moving upward — when assumptions are generous and behaviour is open — it is a self-reinforcing engine of trust and genuine contact. When it is moving downward — when assumptions are defensive and behaviour is guarded — it becomes progressively harder to enter or interrupt from the outside.

The intervention problem

Here is where it gets particularly important for practitioners.

Most of us, when we encounter a downward spiral — in a team, a relationship, a group — bring to it our best skills and our genuine care. We design a process. We create conditions for dialogue. We invite honesty. We model the vulnerability we are asking of others.

And sometimes this works. But sometimes — more often than most practitioners will readily acknowledge — it makes things worse. The defensiveness increases. The dynamic that was moving slowly downward begins to move faster. And the practitioner, puzzled and perhaps unsettled, is left working on the surface of something whose depth they cannot quite reach.

Argyris called this skilled incompetence. The more practised the intervention, the more reliably it can trigger the defensive routines it was designed to dissolve. Not because the practitioner is incompetent. Because the very act of intervening — under conditions of stress, ego-threat, or potential exposure — activates the Model I governing values that the intervention is trying to work against.

People don’t like to discover that they don’t practise what they preach. Even a carefully designed, warmly facilitated process that surfaces the gap between espoused values and actual behaviour can feel like an attack on self-perception. And when it does, the defensive response is swift, largely automatic, and largely invisible — even to the person mounting it. They experience themselves as being honest and reasonable. Their behaviour communicates something quite different.

Argyris found this to be true even among highly trained OD practitioners — people who had explicitly studied these dynamics and were committed to different values. The theory-in-use asserts itself under pressure regardless of the espoused theory. That is, precisely, what makes it a theory-in-use.

The coverup of the coverup

There is a further complication that Argyris named with characteristic precision.

When someone’s defensive routine is activated by an intervention, the socially sanctioned response is not to name it. Naming it directly — I notice you’re responding defensively to this — risks deepening the very dynamic it names. So the defensiveness is observed and not named. The observation is covered. And — and this is the move that makes the whole thing structurally self-sealing — the fact that it is being covered is also not named.

The undiscussability of the undiscussable. The coverup of the coverup.

This is not malice. It is the perfectly logical operation of social systems designed around the avoidance of embarrassment and the protection of self-image. But its cumulative effect is to create environments in which the most important dynamics — the ones that are actually generating the difficulty — are the ones that cannot be spoken about directly.

And here is what this means for practitioners: if you try to address the defensive dynamic directly, you risk deepening it. If you don’t address it, you are working around something whose presence shapes everything. There appears to be no way in.

Appears to be.

Real relationship — and what it makes possible

In 1971, the organisational consultant Sheldon Davis was filmed working with a group in conflict. He intervened rarely. But when he did, the effect in the room was immediate and visible. Afterward, when asked to describe his approach, Davis said something that Bob Dick has returned to many times across fifty years of practice:

His goal, Davis said, was to help people experience each other as real people. When that happens, something already present in us — an inbuilt care and concern for one another — becomes available again. What this means, practically, is that the work is not about adding something new to people. It is about removing, carefully and gradually, what is preventing something already there from coming forward.

This is a different orientation from most change and facilitation work, which tends to assume that new capacities need to be installed. Davis — and Argyris, and Dick — are pointing to something more fundamental: that the relational capacities required for repair are already present in most people, most of the time. They are not absent. They are suppressed. By the defensive routines that Model I conditions generate. By the social taboos against honest expression. By the accumulated weight of untested assumptions that have hardened into certainties.

The practitioner’s task, in this frame, is not to teach people how to be more collaborative. It is to create conditions in which collaboration, which is already natural to them, becomes safe enough to re-emerge.

The distinction that makes the difference

There is a distinction that Bob Dick has found, across five decades of practice, to matter more than almost any other.

The distinction between professional relationships and real ones.

Professional relationships — constrained, role-bound, managed within the norms of organisational life — are not without value. In some situations they are a genuine improvement on open conflict. But they carry a structural cost: the more carefully a relationship is managed within professional norms, the harder it becomes to create the quality of contact that actually moves things.

Real relationship — honest, present, genuinely curious about the other person rather than strategically engaged with their role — is the ground from which repair becomes possible. Not as an endpoint to be achieved, but as a quality of contact that can be cultivated, gradually, through specific moves that lower the cost of honest expression.

Those moves have three elements, developed from the work of Viviane Robinson and traceable directly to Argyris:

First, a clear description of the behaviour that occurred — not an interpretation, but the observable thing that happened. When you said X…

Second, an interpretation of the behaviour, clearly labelled as one’s own possibly inaccurate reading. My sense — and I may be wrong — is that…

Third, a genuine invitation for the other person to offer a different interpretation. I’d be genuinely interested in how you experienced that.

This is the dialectical move that Dick describes as creating hard consensus — not the soft consensus of agreement sought from the beginning, and not the adversarial process of competing for correctness, but the genuinely difficult and genuinely productive process of disagreement held within a shared commitment to understanding. Out of that dialectic between opposing views, something new emerges — something that neither party could have arrived at alone.

The crucial quality is tentativeness. Not the false tentativeness of easing in — hedging so carefully that the information is lost. Genuine tentativeness: the willingness to hold your own interpretation lightly enough that you can actually be corrected by the other person’s response.

What you can do today

Think of a situation in your current practice where an intervention has produced less movement than you hoped for — or where you notice yourself working around something rather than with it.

Ask yourself: what is the assumption I am currently holding about why this person or group is responding as they are? Is that assumption voiced, or unvoiced? Is it being treated as a fact?

Now ask: what would it take to voice it as an assumption — clearly, and with a genuine invitation for correction?

Not you clearly don’t trust this process but I’ve been noticing a quality in the room that I’m reading as hesitation, and I’m wondering whether I’m reading it accurately — what’s your experience?

Notice what makes that move difficult. The answer to that question — the specific obstacle you encounter when you try to voice an assumption as an assumption — is the entry point to your own theory-in-use. And it is, therefore, the place where your practice can most genuinely develop.

The spread of the dynamic

One more thing worth naming, because it has direct implications for consultants and OD practitioners working at the systemic level.

The defensive dynamic Argyris mapped does not stay contained to individual relationships. It spreads. Unilateral control — the self-protective mode that stress reliably triggers — moves from person to person, from pair relationship to team, from team to organisation. The governing values of the individuals shape the norms of the system. Those norms then shape the governing values of new individuals who enter it.

The result is what Argyris called a Model I social system — one in which the environment itself is saturated with defensiveness, and in which everyone is both contributing to that environment and being shaped by it, mostly without awareness. Where the most important things cannot be said. Where the coverups are so thoroughly established that they no longer feel like coverups — they feel like how things are.

This is the system that most change programs are attempting to work with. And it is why, as Argyris asked, even correctly diagnosed problems with valid implementation plans and adequate resources so often fall short of expectations. The system’s defences are not located in any one person. They are distributed across every relationship in the organisation, mutually reinforcing, largely invisible, and activated by the very act of trying to change them.

Understanding this does not make the work easier. But it changes its quality. It replaces the practitioner’s frustration — why won’t they just engage with this honestly — with a more useful curiosity: what is the system protecting, and what would need to change for it to feel safe enough to put the protection down?

That curiosity, held steadily and without defensiveness, is the precondition for the kind of intervention that can actually reach the level where things change.

What comes next

The fourth and final article in this series brings the threads together. It asks what becomes possible when this understanding is inhabited as a practice — not applied as a technique — and extends the horizon toward something that matters beyond the professional: the capacity to remain present, honest, and genuinely adaptive in conditions of increasing turbulence and complexity.

It also carries the details of the July 2026 webinar program, and the invitation to the two free orientation sessions that precede it early June.

Curious? Join us in July for Repairing Relationship: Openness, Trust and the Art of Honest Inquiry in a series of four webinars starting on the 7th and continuing on the 14th, 21st and 28th of July 2026: Why do well-intentioned efforts to improve relationships sometimes make them worse? Drawing on Chris Argyris’s action science, this four-session online program offers evaluators, coaches, consultants and OD practitioners precise, learnable tools for rebuilding trust, surfacing untested assumptions, and turning defensive dynamics into genuine inquiry. Maximum 15 participants. Early bird until June 17th at €399 — against a standard fee of €499. More https://www.tickettailor.com/events/contextcollective/2209886

Bob Dick is a Brisbane-based educator, consultant, mentor, and coach with more than fifty years of experience in action science, facilitation, and organisational learning.

Sebastian Wolf Siebzehnrübl is a Berlin-based Systemic Inquiry Practitioner working at the intersection of systems thinking, evaluative inquiry, and relational practice. He is affiliated with context collective, Berlin.