The Canon the Team Lives By: Why Coaching the Ceremonies Leaves the Rulebook Untouched

The team that kept returning

Four months of careful coaching. Working agreements produced in the first retrospective: all voices are equal, challenges are welcome, psychological safety is named and described and agreed to. Retro formats refined. Sprint planning genuinely improved. The SM growing in confidence and skill. The manager satisfied. The metrics pointing the right direction.

And yet. In every sprint planning session at the platform team at a retail company, the senior engineer's estimate stands unchallenged. Not because no one disagrees — two developers have raised concerns privately with the coach. But in the room, nothing. The coach raises this in a retrospective. The team agrees it is a problem. They produce a new working agreement: technical decisions are subject to group discussion. In the next sprint planning, the senior engineer's estimate stands unchallenged.

The coach writes a better working agreement. The team agrees to it again. The dynamic persists. At month five, the coach begins to wonder whether the team is resistant, or whether the work is somehow inadequate, or whether there is something here that none of the frameworks they have been using actually reaches.

This article is about the layer beneath the ceremonies. The unwritten rulebook that every team carries, that predates most of its current members, and that standard Agile coaching interventions — working agreements, retrospective formats, facilitation improvement — almost never reach.

What the canon is

Every group develops an actual constitution — the real answer to the question what is allowed here? It is not the document. It is not the working agreements. It is the set of norms that members discover through experience: what happens when someone disagrees openly, what the room does when a sensitive topic appears, what a new member learns in their first two sprints about where the lines actually are. This constitution is real, it is enforced, and it predates most of the people who are currently being governed by it.

What makes it hard to see is that it operates beneath the level of explicit agreement. People follow it without being able to articulate it. They know — without knowing they know — that certain things are done here and others aren't. When asked to describe it, they will often describe the formal working agreements instead. The actual constitution is most visible at the moment someone steps outside it and the group responds.

Berne (1963) called this the group canon and identified two dimensions that together determine its character and its coachability.

Hardness — how clearly the canon is stated

In a hard culture, everyone knows when they are inside or outside the rules. There is clarity about what is and isn't permitted. A new member can learn the norms by observation within a few weeks. The rules may be strict or lenient — hardness describes clarity, not severity. A hard culture can be highly permissive (everyone knows the rules and they're relaxed) or highly restrictive (everyone knows the rules and they're demanding). In either case, people can navigate with confidence.

In a soft culture, the rules are murky. People frequently don't know whether what they're about to do is acceptable until they discover — sometimes unexpectedly, sometimes harshly — that it wasn't. Berne: "In a soft culture he often has doubts, and this causes difficulties in the same way that an inconsistent parent causes difficulties for his family." The effect is not merely confusion. It is a pervasive low-level vigilance that consumes attention and suppresses initiative.

Strictness — how severely the rules are enforced

Berne identified four aspects of strictness that are independent of each other and must be assessed separately: whether the culture is permissive or restrictive (how much freedom members have to act without checking); whether it is remunerative or punitive (whether compliance is rewarded or deviation is punished); whether it is lenient or severe (how harsh the consequences are when rules are broken); and whether it is tolerant or strict (how much deviation is permitted before consequences apply). These four aspects can point in different directions within the same team, producing cultures that are hard to read precisely because they are inconsistent across the dimensions.

A team can be punitive without being severe — deviations are met with real responses, but small ones. Or restrictive without being strict — many rules are stated but rarely enforced, so the culture feels stifling on paper and relaxed in practice. The combinations produce very different working environments, and they require different interventions.

The most destabilising combination is a soft-punitive canon: rules that are murky and unclear, and deviation that is punished — sometimes harshly, sometimes unpredictably. In this environment, the rational survival strategy is to attempt nothing new, because you might discover it was forbidden in a way you couldn't predict. The result is the most rigid surface compliance and the most absent genuine contribution. It is also the combination most frequently misdiagnosed as a psychological safety problem — because the surface symptoms (silence, caution, studied avoidance of certain topics) are identical. The mechanism is different. Improving the safety rhetoric does not change the canon.

Where the canon comes from

The canon is inherited, not chosen. It was formed at the team's founding — under the first manager's working style, in response to the first major crisis, shaped by the first time someone broke a significant unspoken rule and something happened as a result. The current team is running a rulebook that most of its members had no part in writing. They have simply learned it, as all members do: through observation, through small deviations and their consequences, through the stories older members tell about what this team is like.

How changeable the canon is depends largely on the fluidity of the team's leadership structure. Berne identified three degrees:

Fluid leadership

In a fluid leadership structure, the leader is accountable to the members and can be replaced without extraordinary effort. The canon updates when its members change — a new person's arrival can genuinely shift what is permitted, because the group retains the authority to revise its own norms. In Agile terms: a team whose working culture actually updates after retrospectives, where norms evolve visibly over time, is running fluid leadership. The canon is alive.

Viscous leadership

In a viscous structure, the leader is accountable only to a privileged subset of the group — the senior contributors, the long-tenured members, those with specific technical authority. The canon can be changed, but only with the assent of those who hold that authority. New members and majority voice don't reach it. Retro outcomes that lack the endorsement of the viscous authority figures are generated and quietly set aside. In Agile teams, viscous leadership shows up as a senior contributor or technical lead whose implicit endorsement is required before any team-level norm change actually holds.

Frozen leadership

In a frozen structure, the leadership preserves itself regardless of who occupies the formal roles. The canon is self-maintaining. New members are recruited to it rather than shaping it — they learn, within the first two or three sprints, what this team is and what it is not, and they adapt. Successive coaches, working agreement rewrites, and Agile framework changes leave the canon intact because the canon doesn't run through the formal structure. It runs through the private structure, and the private structure determines the outcome.

Sari van Poelje (1995) observed something important about how systems reach this state: they don't begin frozen. Canons freeze through a developmental process that is entirely comprehensible at each step and visible only in retrospect. A founding leader's working style becomes the team's default norm. A crisis produces a particular response that is then preserved as the right way to handle difficulty. A period of turbulence generates protective rules that remain after the turbulence is over. At each step, the move that produced the next degree of rigidity made sense. The team that now has a frozen canon did not choose it. It accumulated it.

This matters for coaching because it reframes the problem entirely. The team with a frozen canon is not broken or resistant or dysfunctional. It is running the rulebook it was given, enforcing the norms that history produced, following the constitution that predates most of its members. Coaching that targets behaviour without addressing the canon is asking the team to deviate from the rules while leaving the enforcement mechanisms intact. The result is predictable: surface compliance during the session and a quiet return to canon-compliant behaviour afterward.

Diagnosing the canon in a real team

The canon cannot be diagnosed by asking about it directly. Asking produces descriptions of the aspirational canon — the working agreements, the stated values, the norms the team believes it has or would like to have. The actual canon is visible in its enforcement: in the patterns that persist across format changes, in the silences that repeat regardless of the facilitator, in what new members learn in their first month that nobody told them explicitly.

Signs of a soft canon — murky rules

Inconsistent responses to similar behaviours: the same action produces different outcomes depending on who performed it, when, or what the room's energy was. People asking "is it OK if I…" about things that should be obvious in an established team. A general atmosphere of uncertainty about what is and isn't permitted that doesn't resolve as the team matures. New members taking much longer than the technical onboarding would explain to find their footing — because the norms they're trying to learn keep shifting. In a soft culture, the rules are discovered retrospectively rather than known in advance.

Signs of an over-hard canon — excessive rule-specification

Working agreements that enumerate specific scenarios rather than principles. Retrospectives that spend significant time re-establishing what everyone already knows. A sense that there is always a rule to invoke — and that invoking the rule is a legitimate substitute for exercising judgment. People complying precisely with the letter of agreements while disregarding their spirit: the retro runs exactly on time, action items are generated with owners and due dates, and completion rates are low because the format was followed but the genuine problem was not reached. An over-hard canon produces bureaucratic compliance without genuine participation.

Signs of a punitive canon — high cost of deviation

Topics that never appear in retrospectives despite being visibly present in the team's working life. Silences filled quickly — by the facilitator, by a senior member, by anyone who can redirect attention before the silence becomes productive. Action items framed carefully: specific enough to appear actionable, vague enough to be undoable without consequence. Escalations that happen in Slack sidechannels rather than in the designated process. People asking "is it safe to say this?" even in nominally safe environments — because the nominally safe environment sits above a canon that doesn't permit the thing they want to say. The working agreements create a declared permission; the canon enforces a different one.

Signs of frozen leadership — a non-updatable canon

The most diagnostic observation: what happens when a new member joins? Does their arrival create genuine space for norm evolution — do they bring something that changes how the team works, even slightly? Or does the team's canon recruit the new member within the first two sprints, teaching them through subtle and direct signals what is done here and what isn't? A canon with frozen leadership shows up most clearly not in the working agreements or the retrospectives but in new member onboarding: specifically, in the informal process through which new members learn the actual rules. If every new member converges on the same behaviours within four weeks regardless of their starting point, the canon is frozen.

Why standard interventions miss the canon

Working agreements, retrospective format improvements, and facilitation skill development all operate at the surface of the system. They describe or invite a different canon. They do not change the actual canon. This is not a failure of the interventions — it is a misalignment between the level at which they operate and the level at which the problem exists.

Working agreements exist in the public structure — they are explicit, visible, agreed to in a meeting. The canon exists in the private structure: the mental picture each member carries of what is actually enforced here, accumulated through experience rather than through agreement. Berne: "The private structure determines the outcome." A working agreement that says "all voices are equal in sprint planning" does not override a canon that says "the senior engineer's technical estimates are not challenged." Both can coexist. The agreement and the canon exist at different levels of the system. The system runs on the canon.

A new retrospective format creates conditions under which people might speak more freely. But if the canon doesn't permit what the format is inviting, the format is filled with canon-compliant content — careful, surface-true, relationship-preserving — while the canon-incompatible content remains unsaid. The format runs; the retro doesn't happen. The coach sees participation rates that look healthy. The team produces the retrospective the format asked for and the retro the canon allowed.

The specific failure mode this produces: a coach can work with a team for twelve months — improving facilitation, growing the SM, producing genuinely good working agreements — and at month thirteen, a new manager arrives, removes the working agreements, and within three sprints the team is running the same canon it had at month one. The ceremonies were built on top of the canon rather than into it. Remove the ceremonies and the canon reasserts itself immediately, because it was never changed.

Three moves that address the canon

Canon-level change requires canon-level intervention. The three moves below operate on the rulebook itself rather than on the behaviours the rulebook produces.

Name the canon, not the behaviour

The coach who says "I notice that technical decisions rarely get challenged in sprint planning" is naming a behaviour. The coach who says "it sounds like there's a rule here — that technical authority belongs to one person, and it doesn't transfer even when the situation is uncertain — and I'm curious whether that's a rule the team has chosen or one it inherited" is naming the canon. The distinction is consequential: people can receive feedback on behaviour without connecting it to the underlying rulebook. Canon-naming invites the team to see the rulebook itself. That is the first step toward being able to revise it.

Canon-naming requires more precision than behaviour-naming. The coach must name the rule, not the outcome. "Nobody challenges the tech lead" describes an outcome. "The rule seems to be that technical authority is vested in the most experienced person and doesn't transfer even when their judgment is uncertain" describes the canon. Naming at this level of precision is uncomfortable because it makes the invisible visible in a way that cannot be unseen. That discomfort is the move. A canon cannot be examined by people who cannot see it.

Distinguish hardness from strictness before intervening

A soft-lenient canon (murky rules, mild enforcement) and a hard-punitive canon (clear rules, harsh punishment) both produce dysfunction, but the appropriate interventions point in opposite directions. Adding clarity to a soft-punitive canon — making the rules more explicit — makes things worse: the rules are now clear, and the punishment for deviation remains real and unpredictable. Relaxing enforcement in a hard-punitive canon while leaving the rules themselves unchanged typically produces a brief period of tentative experimentation followed by a return to the original pattern when enforcement reasserts.

The diagnostic question to ask before any canon-level intervention: is the problem here about clarity or about enforcement? If the team doesn't know what's allowed, the canon needs hardening — clearer statements of what is and isn't permitted, applied consistently. If the team knows the rules and finds the enforcement disproportionate or unpredictable, the canon needs loosening — either in the rules themselves or in the consequences for deviation. Getting this distinction wrong is the most common source of interventions that worsen the dynamic they are trying to address. Berne: "A good leader must be careful to distinguish these two aspects and think about each separately."

Work with the moment of deviation

The canon is most visible — and most available for change — at the moment someone deviates from it and the group responds. This is when the enforcement mechanism activates, which means the rule can be seen operating in real time. It is also the moment the coach has the most leverage: not to adjudicate whether the deviation was right or wrong, but to make the moment available for reflection.

The specific move: name what you notice in the room, without directing the naming at any individual. "Something shifted just now. I'm not sure everyone experienced it the same way — what did people notice?" This invites the team to examine its own enforcement mechanism without requiring anyone to have broken a rule intentionally or to be held responsible for the group's response.

When deviation is survived — when someone steps outside the canon and the enforcement mechanism either doesn't activate or activates mildly and is named — the canon's threshold shifts slightly. Not dramatically. Perceptibly. Over time, through multiple such moments, a viscous or frozen canon can begin to loosen. The canon does not change through discussion alone. It changes through a series of lived moments in which deviation is survived rather than absorbed or punished.

The coach who keeps working on the wrong layer

The coach who cannot read the canon is working on the symptom while the cause runs underneath. This is not a skill failure. It is a conceptual one — the frameworks most Agile practitioners are trained in operate at the ceremony level. They produce ceremony-level results: better retrospectives, cleaner working agreements, improved facilitation. The canon continues unchanged because nothing in the intervention reached it.

Reading the canon requires a different quality of attention than reading the ceremonies. The ceremonies are visible. They are designed to be. The canon is visible only in its enforcement — in the moments when the room responds to something that stepped outside the rules, in the patterns that persist across every format change, in the topics that appear in sidechannels but not in retrospectives. This is the layer beneath the public structure that Berne described as determinative: the private structure, which is never written down, always enforced, and almost always invisible until the coach learns to look for it.

The three layers this series addresses are connected. The leadership structure — the three types from the previous article — determines who sets and maintains the canon. The canon determines what behaviour is possible within the system. The third layer — the collective imago, the team's shared mental picture of what it is and what it could become — determines what the team believes is possible within its canon. Coaching that reaches all three layers is not doing more of the same work. It is working at the level where the system actually runs.

The working agreements are what the team agreed to. The canon is what the team actually lives by. The coach who can only see the first will spend years working on the second without knowing it.