The Picture No One Is Sharing: How Divergent Group Imagoes Make Self-Organisation Impossible

The ceremony that worked and the team that didn't

The sprint board clears every Friday. Velocity is stable. The SM's line manager is satisfied. Retrospectives produce participation rates the coach can point to. By every observable metric, the team is working.

Walk out of standup and watch where each person goes. The senior developer opens a Confluence page for a platform migration he considers the team's real work — the sprint board is a reporting layer on top of what he actually does. The product owner opens her inbox to a queue of stakeholder requests, because in her picture of this team, clearing that queue is what they are there for. The junior developer waits to see what he'll be assigned, because in his picture of this team, someone else makes the calls and his job is to execute well when called on. Three people walk out of the same ceremony and enter three different teams.

None of this appears on the sprint board. The board is a public object — visible, shared, argued over. But each person carries something the board does not capture: a picture of the team itself. What kind of thing it is. What it is for. Where it is going. That picture is never discussed in any ceremony. It shapes everything that happens in them.

The previous article in this series showed that every team carries an unwritten rulebook — a canon — that operates beneath its ceremonies and that working agreement rewrites cannot reach. This article goes one layer deeper. Beneath the canon, each person holds a private mental picture of what this team is and where it is heading. Eric Berne called this the group imago, and in 1963 he wrote something that the field has not fully absorbed in the sixty years since: "the private structure determines the outcome."

When those private pictures diverge significantly — when the SM imagines a self-directing team, the principal engineer imagines a technical delivery unit, and the PO imagines a service function for stakeholder requests — self-organisation becomes structurally impossible. Not because of the methodology, not because of safety levels, and not because the people are unwilling. Because the team's members are not imagining the same team.

What the imago is

The group imago is not team culture. Culture is collective and emergent — it exists between people, in the shared patterns of how a group behaves over time. The imago is individual and internal. It is each person's structured mental model of the group as an object: who is in it, what roles people occupy, what the group is for, what the norms are, what the future looks like. Every person who belongs to a group carries one. It operates before the first word of the meeting is spoken.

Nor is it the same as psychological safety, which is a felt quality of an environment. Or trust, which is a relational assessment of a specific person. The imago is a structured internal picture of the group as a whole. Shapiro and Carr (1991) described it as a mental model "internal to oneself, part of one's inner world, relying upon the inner experiences of my interactions, relations and the activities I engage in, which give rise to images, emotions, values and responses in me." The point is the structure. It is not a feeling about the team. It is a picture of it.

Berne (1963) distinguished the group imago sharply from the public structure — the org chart, the declared purpose, the official roles. The public structure is what the institution announces. The private structure is what each person actually works from. These two can align closely or diverge quietly over months without anyone naming the gap.

The slot structure

Berne conceptualised the imago as having slots — positions for the leader, for peers, for the self, for outsiders. Each slot carries expectations: this person controls this, this person is like that, these people are allies, those are threats. The slots are filled from experience, but not neutrally. They are filled from prior experience of groups that felt similar — previous managers, previous teams, previous organisations with recognisable dynamics.

The coach appears somewhere in every team member's imago, and often in different slots for different people simultaneously. The SM imagines the coach in the mentor slot — someone whose function is to support his development. The senior developer imagines the coach in a management-adjacent slot — another layer of observation with unclear authority. The junior developer imagines the coach in a facilitator slot — the person who runs the retrospectives. Three slots, three coaches, one person. That inconsistency is not a relationship problem. It is a structure problem.

Crucially, the imago forms before experience with this group begins. It is built from prior group history — previous teams, previous managers, previous organisations whose dynamics felt sufficiently familiar that the person's internal model adapted from them. A developer who lived through two failed Agile transformations carries an imago shaped by both. A product owner who spent four years in a command-and-control product function carries that picture into every sprint review. The group is new. The picture is often not. This is what makes the imago both durable and invisible: it predates everything the team has done together.

How imagoes develop: four stages

Petruška Clarkson (1991) described four stages through which a group imago develops as a person accumulates experience with a specific group. The stages matter for the coach because different members of the same team can be at different stages simultaneously — producing divergence as a developmental feature of any working group, not as a pathology to be corrected.

The provisional imago

The provisional imago is the picture a person brings to a group before significant shared experience with it. It is built from prior group history and projected expectations. It feels real — the person experiences it as a picture of this team, not as a projection from somewhere else. In teams assembled under an Agile transformation, almost every member arrives with a provisional imago shaped by the organisation's non-Agile past: the picture is of a hierarchy with sprint labels on it. Ceremonies run inside that provisional picture without updating it. The board is clean every Friday. The provisional imago remains intact.

The adapted imago

As shared experience accumulates, the imago begins to update. But it does not update uniformly across team members, because different people are attending to different signals. The developer who cares most about technical autonomy is watching whether the SM defers on architecture decisions. The PO who cares most about stakeholder responsiveness is watching whether sprint goals can flex mid-sprint. The junior developer is watching whether speaking in retrospective has consequences. Each person updates their imago from the events that matter most to them — which means adapted imagoes are diverging from the first sprint, not converging. The team is accumulating shared history. The pictures are moving apart.

The operative imago

The operative imago is the stabilised internal picture the person is actually using to navigate the group. It is functional enough that there is no pressure to update it. In Agile terms, this is the team in steady-state delivery — working, meeting, reviewing, retrospecting — running from operative imagoes that may not have changed significantly in months. The problem is not that operative imagoes are stable. The problem is that they can be deeply divergent across team members while the ceremonies produce identical-looking outputs. The divergence is invisible to the public structure. The board clears every Friday. The imagoes do not.

The secondarily adjusted imago

A secondarily adjusted imago is a picture that has been revised in light of this team's specific lived history — including its failures, its particular relational texture, and what it has discovered about itself over time. The person is seeing this team, not a prior team with new labels. This stage is rare in young teams and regularly disrupted by composition changes, leadership transitions, or organisational shocks that force the imago back toward provisional.

When some members carry secondarily adjusted imagoes and others are still provisional, the experiential gap produces a specific kind of friction — people who seem to be in different realities, because in the private structure they are. The experienced member who has revised their picture from this team's actual history and the newer member who is still running a provisional picture from a previous organisation are not having a personality conflict. They are working from genuinely different maps.

What the stages show is that imago divergence is not an accident or a sign of dysfunction. It is the predictable result of different people, attending to different signals, updating their internal pictures from different vantage points, at different rates. The team that has been together for eighteen months and looks functionally stable may be running four different operative imagoes that have never been compared or examined. The ceremonies will not surface this. Neither will working agreement workshops. The private structure runs beneath both.

When imagoes diverge

Imago divergence is not disagreement. It is not conflict, disengagement, or a failure of communication. It is a structural condition in which team members are navigating by different maps of the same territory. It coexists with apparent harmony. People can be collegial, ceremonies can go smoothly, nobody is overtly at odds — while each person is working from a private picture of the team that differs significantly from every other person's.

The specific mechanism by which divergent imagoes prevent self-organisation is this: genuine joint decision-making requires a sufficiently shared picture of what the team is trying to do. It does not require agreement on everything. But it does require that the members are navigating by maps that overlap enough in the domains that matter — what the team is for, who holds what kinds of authority, where the team is heading — that a shared direction is possible. Without that overlap, decisions that require collective judgment are quietly resolved by whoever has more formal authority or more persistence.

Consider a technical dependency that emerges mid-sprint and threatens the sprint goal. In the SM's imago, this is a moment for the team to surface trade-offs and propose a path — the event he has been coaching them toward. In the principal engineer's imago, this is a technical decision that is his to make; the sprint review is where he will report it, not where it will be negotiated. In the PO's imago, this requires a stakeholder impact assessment that must happen before anything is decided. Three imagoes, three decision-making processes, one event. What actually happens: the engineer acts, the PO escalates, the SM runs a retrospective item on "improving decision-making clarity." The ceremony produces an output. The imagoes remain divergent.

This produces the distinction between ceremony-shaped outputs and lived self-organisation. Ceremony-shaped outputs are the artifacts the public structure generates: plans, commitments, reviews, retrospective action items. They are real and often useful. Lived self-organisation is something else: a team that, without being directed, surfaces its own problems, negotiates its own trade-offs, and moves from genuine collective judgment. The first can be produced by a team with sharply divergent imagoes. The second cannot.

This also clarifies the relationship between the imago and the canon described in the previous article. The canon — the team's unwritten rulebook — governs behaviour in the public structure: what is done here, what the room does when someone steps outside the norms. Imagoes govern the private structure: each person's picture of what the team fundamentally is. A team can have a fully operative canon — clear norms, consistently enforced — and still have wildly divergent imagoes about what the team is for. They are different layers. Canon-level coaching reaches the rulebook. Imago-level work reaches something deeper.

The collective imago

Berne's (1963) understanding of group imagoes was individual: "there is one major hypothesis to be validated: that each member has a different mental picture of the group, based on his personal feelings." Michael Korpiun (2020) extended this by asking what it would mean for a group to develop something shared — not a single uniform picture imposed on everyone, but a sufficient convergence between individual imagoes that the group could function as a genuine collective agent. His answer was the concept of the collective imago.

Korpiun defines the collective imago as representing "both the shared and discrete attitudes of whomever is aware of an organization regarding self, relationships, groups, organization, and society." The level of shared understanding determines the ability to provide cohesion; the level of discrete understanding determines the force of group dynamics. A team with high convergence has sufficient shared picture to move together. A team with high divergence is driven primarily by the friction between its members' different maps.

What convergence is and is not

Convergence is not consensus. Consensus is a public structure agreement — people have discussed, negotiated, and reached a joint position. Convergence is a private structure overlap — the members' internal pictures of the team share enough common ground that decisions do not require the full negotiation every time. The difference is that consensus can be performed without any underlying convergence. A team can agree in a meeting on a sprint goal while each person's private picture of what the team is for remains unaddressed.

Convergence is also not uniformity. Different roles legitimately produce different imagoes in some dimensions. The engineering lead's picture of the team will emphasise different aspects than the QA engineer's. Full uniformity would be both unrealistic and undesirable — the team benefits from the different vantage points its members hold. Convergence means that the pictures overlap enough in the load-bearing domains that genuine collective action becomes possible. A team with sufficient convergence can hold a significant disagreement about technical direction and still produce a decision that both parties experience as the team's decision. A team with sharp divergence produces formally similar decisions — agreed, documented, moved on — that one party experiences as the team's and another experiences as someone else's imposition, regardless of how the meeting went.

The three load-bearing domains

Not all dimensions of the imago need to converge for self-organisation to become possible. Three domains are load-bearing — the domains where divergence specifically prevents genuine collective action.

The first is primary provision — what the team is fundamentally for. Is it a delivery unit, measured by output against a plan? A learning system, measured by the team's growing capability? A product function, responsible for outcomes in a market? A cross-functional problem-solving entity, assembled to address a specific organisational challenge? When imagoes diverge on this, success criteria diverge with them. The developer measuring success by delivery velocity and the SM measuring it by team autonomy are not disagreeing about methodology. They are operating from different pictures of what the team exists to produce.

The second is authority structure — who controls access to the primary provision, and who holds which kinds of decision-making authority. This maps onto Berne's leader slots. When imagoes diverge on this, every significant decision restarts negotiation from first principles, because each person is working from a different picture of who decides what. The principal engineer who imagines technical decisions as his prerogative and the SM who imagines technical direction as a team conversation are not having a communication failure. They are filling the same slot with different people.

The third is trajectory — where the team is heading and over what timescale. A member whose imago places the team on a long developmental arc toward genuine self-organisation will behave differently from one whose imago places the team as a temporary project vehicle running until the release. Sprint planning looks different through each lens. Investing in team capability looks like overhead through one and like the primary work through the other. A team with convergent imagoes on trajectory can disagree on the right velocity without fracturing. A team divergent on trajectory is in a permanently different conversation about what the work is for.

How convergence happens — and fails to happen

Convergence is not produced by agreement. It is produced by shared experience that has been processed deeply enough that imagoes update in compatible directions. Korpiun (2020) introduces the term imago alignment for this process: "the conscious and purpose-oriented leadership process of helping members of an organization to raise awareness about their individual mental images, sharing them with each other in dialogue and exploring common ground." It is iterative, nonlinear, and takes time — "it may take months or even years for a whole organization to undergo a substantial transformation."

Convergence fails in three specific ways. The first: experiences are shared but not processed. The event happens — a sprint fails, a difficult stakeholder interaction occurs, a surprising success. Everyone is present. No space is created for collective meaning-making. Each person updates their imago privately, from their own vantage point, in their own direction. The team has more shared history. The pictures have moved further apart.

The second: processing is public but not private. The action item is created, the verbal agreement is reached in the retrospective, the process change is announced. But the underlying picture is not updated. The person says "yes, we'll involve QA earlier in the sprint" while their private picture of QA's role in the team — a gate at the end of the pipeline — remains unchanged. The public structure shifts. The private structure does not.

The third: new experience is assimilated into the existing imago rather than updating it. The member fits each new event into their prior picture without revision. A ceremony improves, so the team is becoming better at the methodology. A decision is made by the principal engineer, so senior technical judgment is confirmed as the authority structure. Every new event becomes evidence for the existing picture. The most common version of this is an entrenched provisional imago that has become load-bearing — the person's working model of the team that was formed before significant shared experience and has been reinforced by selective attention ever since. Ceremony-level coaching does not reach this. Neither does the canon. The picture predates both.

Reading divergence without asking

The imago cannot be diagnosed by asking about it directly. Asking produces descriptions of the aspirational team — what people believe the team is or would like it to be. The actual imago is visible in the patterns it produces: who speaks in what ways, what gets escalated and what doesn't, how people respond when something changes. Three diagnostic moves allow the coach to read imago divergence from what is observable in the public structure.

Who speaks in whose name

Watch the pronoun economy. Whose "we" sticks — draws other voices in, advances the conversation, is accepted as reflecting the group? Whose "we" lands as an individual claim — registers briefly and is implicitly reset by the next contribution? In a team with convergent imagoes, "we" claims have broad purchase because people are orienting by sufficiently similar maps. In a team with divergent imagoes, multiple competing "we" claims appear in succession and do not resolve — each is accurate about a different team. The coach listens not just for what is said but for whether the claim has traction. A "we" that the room immediately continues from is convergence-diagnostic. A "we" that produces a moment of quiet before the next person speaks from an entirely different frame is divergence-diagnostic.

What gets escalated and what does not

Routing patterns are imago-legible. A team member whose imago positions the team as a unit under external authority will escalate upward decisions that a member with a self-organising imago would surface to the team. Watch what types of problems are brought to the team space — to retrospectives, to standup, to the team channel — versus what is managed privately, versus what is routed externally to the line manager, the sponsor, or the stakeholder. The pattern is not about risk tolerance or communication preference. It is about what kind of thing the team is imagined to be, and what kinds of decisions that kind of thing is authorised to make. Consistent routing patterns across multiple team members, across multiple events, reveal the structural picture each person is working from.

How change is received

When a change is announced — a sprint goal revision, a new team member joining, a process adjustment — watch how the announcement lands. A team with convergent imagoes produces a shared orienting response: people are all computing what this means for the same team, even if they have different immediate reactions to it. A team with divergent imagoes produces multiple parallel responses, each calibrated to a different picture of the team. One person asks about the impact on the delivery timeline. Another asks about what this means for the team's learning trajectory. Another asks about stakeholder communication implications.

The questions are not wrong. They are calibrated to different primary provisions. The person asking about the delivery timeline is operating from an imago in which the team's primary purpose is delivery against a plan. The person asking about learning trajectory is operating from an imago in which capability development is the point. The person asking about stakeholder communication is operating from an imago in which the team's primary relationship is outward-facing. The same announcement, three teams, one room.

Three moves for working at the imago layer

The coach working at this layer has three moves available — not techniques in a toolkit sense but orientations toward work that differ from ceremony-level coaching. None of them produces convergence directly. They create conditions in which convergence can happen, over enough cycles of shared experience and deliberate processing.

Design for meaning-making

A retrospective that produces action items is processing the public structure. A retrospective that surfaces what a significant event revealed about what kind of team this is — or is trying to be — is processing the private structure. The difference is not in the format. It is in what the facilitator makes space for after the event has been examined. After a sprint failure, after a difficult stakeholder moment, after a success that surprised people, the question that reaches the imago layer is not "what do we do differently next time?" but "what does this tell us about what we're trying to be?" Not as a formal agenda item — as a deliberate facilitation move, held briefly, returned to in subsequent cycles.

The coach also looks for naturally occurring moments of shared experience that have not been processed — a difficult decision that produced tension and was moved past, a significant change that was announced and immediately actioned without examination. These are imago-revision opportunities that pass at ceremony speed. Slowing down is itself an intervention: creating space for the question to land, rather than immediately generating the next artifact.

Make the gap visible — let the team name it

When divergence has been diagnosed, the second move is to surface it without imposing a vocabulary the team does not have. The sentence-completion exercise in the second vignette is one version. Any facilitation move that invites individual expression before collective discussion, and then creates space for seeing the differences, will produce partial imago externalisation. The key is that the coach does not narrate what they observe. They ask: "We have eight different answers here — what do we notice?" When the team names the divergence, they are relating to their own private structure. When the coach names it, they are relating to the coach's interpretation of it.

This move requires reading the room carefully before deploying it. Divergence that becomes suddenly visible can produce defensiveness if the team does not have sufficient experienced safety — the kind that comes from a history of productive difficulty together, not from a safety survey or a working agreement. Sometimes another cycle of the first move — meaning-making from shared events — comes before the gap can be shown. The coach judges the sequence, not just the move.

Protect the processing space

Imago convergence is not produced in a single retrospective. It requires repeated shared experience, repeatedly processed, over enough cycles that the private pictures move in compatible directions. Korpiun (2020) is explicit about the timescale: imago alignment "is a nonlinear, iterative, and evolutionary process." The coach who expects visible convergence from a well-designed workshop is applying a ceremony-level timescale to private-structure work.

This means the coach advocates structurally for what makes sustained processing possible: team composition stability, ceremony regularity, and most critically, the preservation of processing time — retrospectives that are not cut short, discussions that are not replaced by action-item lists, moments where the team can sit with what it has discovered about itself without being immediately productive. In organisations under delivery pressure, these are the moments most readily cut. They are also the moments where imagoes are most available for revision.

The coach also manages their own consistency. If the coach occupies different slots in different team members' imagoes — authority figure for some, peer for others, observer for others still — the coach is part of the divergence. Consistent presence, consistent positioning, consistent facilitation style across members and across sessions allows the coach's slot in the team's imagoes to stabilise. This is itself a small convergence signal. The coach is not just facilitating convergence. They appear in it.

The three layers this series addresses form a connected architecture. The leadership structure — the three types from the first article — determines who sets and maintains the canon. The canon — the team's unwritten rulebook — determines what behaviour is possible within the system. The collective imago determines what the team believes it fundamentally is, and whether its members are imagining the same team at all. Coaching that reaches all three layers is not doing more of the same work. It is working at successive depths of the private structure.

The paradox in this work is the timescale. A team can reach ceremony compliance in two sprints. A coach can produce working agreement improvements in a single retrospective. But imago convergence — the state where the team's members are navigating by maps that overlap enough to support genuine collective judgment — takes months of deliberate processing and continuous shared experience. It is slower work than it appears, and it is the work that determines whether self-organisation is a feature of the sprint board or a feature of the team.

Until the team can see the different pictures in the room, it cannot build a shared one — and self-organisation will remain a feature of the sprint board, not of the team.