Question 5: What Is the Viable Scope?
The short version
Every system requires a boundary: a distinction between what is inside and what is outside. Without a maintained boundary, there is no system: just a collection of elements dissolving into their environment, unable to self-organise, unable to maintain any characteristic behaviour of their own.
The question of scope is the question of whether that boundary is drawn correctly, whether the system’s extent matches the actual structure of the problem it needs to govern. When scope is wrong, even a well-designed system with correct goals, sufficient variety, complete architecture, and sound feedback will fail. It is solving the wrong problem at the wrong scale.
Scope failures come in two forms, and both are chronic features of the contemporary world. The first is underscoped governance: a problem operates at a scale larger than any existing system can address, leaving every intervention within existing structures short of the root cause. The second is overconnectivity: a drive to integrate, globalise, and connect has dissolved the boundaries that allowed smaller systems to self-organise, destroying the adaptive capacity that no central system can replace.
The fifth diagnostic question is: is the scope of this system matched to the actual scale of the problem it needs to govern, and has overconnectivity dissolved the boundaries that make self-organisation possible at the levels below?
Why this question is non-obvious
The question of scope is easy to overlook because systems come with their boundaries already drawn. An organisation has a legal definition. A government department has a remit. A hospital has a catchment area. A school has an intake. These boundaries feel fixed (given by history, law, or convention) and analysis typically proceeds within them rather than questioning them.
The cybernetic challenge is to treat scope as a variable, not a constant. The relevant question is not “how do we solve this problem within our existing scope?” but “at what scope does this problem actually operate, and is our governance boundary aligned with that scope?”
The two failures are different in character and call for different responses.
In underscoped governance, the problem’s causal structure extends beyond the governance boundary. The regulator can only see and act on the portion of the system that falls within its scope, which is not the portion where the critical feedback loops operate. Interventions within scope produce real effects within scope. They cannot address the root cause because the root cause is outside scope.
In overconnectivity, the drive to connect and integrate has itself destroyed the boundaries that allowed subsystems to govern themselves. The result is a system simultaneously over-governed (everything is coupled to everything, so every local action has non-local consequences) and under-governed (the scale of the whole exceeds any single governing system’s requisite variety).
Both failures are structural. Both are invisible to linear analysis that takes existing scope as given. And both are producing some of the most intractable failures of the current period.
The Markov blanket: a precise definition of boundary
The biophysicist Karl Friston provides the most rigorous current account of what a system boundary is and why it matters.
A Markov blanket is the statistical boundary that separates a system’s internal states from external states, such that internal states are conditionally independent of external states, given the blanket. In plain terms: what happens inside the system is not directly affected by what happens outside it: all influence from outside is mediated through the boundary.
The blanket has two sectors. Sensory states are influenced by external events but not directly by the system’s internal states: they are the system’s inputs. Active states are influenced by internal states but not directly by external events: they are the system’s outputs. Together, they form the closed loop through which the system acts on the world and perceives the consequences.
A Markov blanket is a pattern of conditional independence: a statistical structure, not a physical wall. The skin is a Markov blanket for a biological organism. The cell membrane is a Markov blanket for a cell. A community’s shared norms and local communication patterns form a Markov blanket for that community. An organisation’s internal decision processes form a Markov blanket for the organisation.
What makes this practically important is Friston’s theorem: a system that persists through time (that maintains any characteristic set of states rather than dissolving into its environment) must have a Markov blanket. The blanket is not something a system can choose to have or not have. It is the definition of being a system rather than a collection of unrelated elements.
The corollary: destroy the Markov blanket and you destroy the system’s capacity for self-organisation. What remains is less organised, because the self-organising capacity that existed at the level of the smaller system has been destroyed, and no equivalent capacity exists at the larger level.
Friston’s warning, stated with unusual directness:
“Self-organisation rests upon the sparsity of coupling: the connections that are not there. If you go around globalising and increasing connectivity, you’re destroying Markov blankets. If you destroy Markov blankets, you destroy self-organisation.”
Applying the question
Five diagnostic moves for assessing scope:
1. Draw the actual boundary: not the nominal one.
Identify what is formally inside the scope of governance: the legal remit, the organisational boundary, the policy domain. Then identify what is actually inside the scope of the problem: where do the causal factors originate? Where do the feedback loops close? Where do the consequences land?
If the problem’s causal structure extends beyond the governance boundary, scope is insufficient. If the problem’s consequences land on parties outside the governance boundary (parties who have no voice in the system and whose feedback is not in the loop), scope is distorted.
2. Check whether the critical feedback loops are within scope.
Take the feedback loops identified in Question 4. For each loop: are all the elements of that loop within the scope of the governing system?
If a critical loop closes outside the governance boundary (if the response to the system’s outputs comes from actors or processes the system cannot see or influence), then the system cannot regulate itself on that dimension. It is receiving feedback from an environment it has no mechanism to act on.
3. Check for dissolved Markov blankets below the current scope.
Identify the subsystems that are supposed to self-organise within the larger system. Ask: do they have maintained boundaries? Can they develop their own internal dynamics without those dynamics being directly coupled to every other part of the larger system?
Signs of dissolved Markov blankets: local units have no discretion, because every decision is coupled to central policy. Performance in one unit directly affects the conditions of all others, eliminating the conditional independence that would allow each to adapt locally. Information about internal states of subsystems is routinely collected and transmitted upward, removing the conditional independence between internal states and external observation.
When Markov blankets are dissolved, the subsystems lose their self-organising capacity. They become components rather than systems: executing instructions rather than governing themselves. The variety they once contributed to the whole is lost. The whole system becomes simultaneously more brittle (no local buffers) and harder to govern (more variety has been concentrated at the top, which is exactly the variety mismatch identified in Question 2).
4. Locate where instability is being exported across scope boundaries.
Question 4 identified the export of instability: systems that maintain apparent stability by pushing their instability onto other systems or time periods. Scope analysis locates the direction and destination of that export.
Who is outside the governance boundary but receiving the consequences of decisions made within it? What future period is receiving the instability that cannot be resolved within the current temporal scope? What geography is receiving the environmental cost that is not within the regulatory scope?
Exported instability will return when the receiving system’s capacity to absorb it is exhausted. Scope analysis maps where that return journey begins, and how far along it the situation currently is.
5. Ask whether a new scope needs to be created.
Some problems cannot be addressed by reforming systems that exist at the wrong scope. They require the construction of governance at the appropriate scale: a new Markov blanket at the level where the problem actually operates.
This is a different kind of intervention from anything identified in Questions 1–4. It is not a question of fixing a feedback loop or building a missing function within an existing system. It is the creation of a new system boundary, with the identity, coordination, intelligence, and operational functions that a viable system at that scope requires.
The underscoped governance failure
The most consequential example of underscoped governance in the current period is climate.
The problem (the feedback dynamics that govern global temperature and atmospheric composition) operates at a planetary scale. The carbon cycle does not respect national boundaries. Emissions in one country affect precipitation patterns in another. Deforestation in one region alters the water cycle globally. Permafrost thaw in the Arctic affects monsoon systems in Asia.
The governance systems that exist (international agreements, national carbon pricing, regulatory frameworks) operate at national or sub-national scale. No current governance body has the regulatory scope, the monitoring capacity, or the enforcement mechanism to act at the scale at which the critical feedback loops operate.
Political will is also absent, but the structural scope mismatch would persist even if it were present. The interventions available within existing scope are real interventions: they produce real effects within their scope. They cannot address the root cause because the root cause is the structure of a planetary-scale feedback system, and no planetary-scale governing body exists.
The VSM analysis of this situation (as developed in the CybSoc Cybernetics for Net Zero work) is clear: there is no System 5 at planetary scale: no shared identity, no collective statement of what this system is for and what values govern it. There is no System 4 with authority to act on global environmental intelligence. There is no System 2 coordinating the interactions between national-scale System 1 operations. The architecture required for governance at the appropriate scope does not exist.
Building that architecture is what scope analysis points to: a structural requirement for effective governance. The analysis does not tell you how to build it. It tells you that incremental improvement within existing scope will not close the gap.
The overconnectivity failure
Overconnectivity is the inverse failure, and it has become more acute with each decade of globalisation and digital integration.
The 2008 financial crisis is the paradigm case. The financial system of the early 2000s had become extensively interconnected through instruments (credit default swaps, collateralised debt obligations, repo agreements) that created tight coupling between institutions that had previously operated with significant independence. The Markov blankets between banks, between national financial systems, between asset classes, had been progressively dissolved.
When one part of the system encountered distress (the US subprime mortgage market) the tight coupling transmitted that distress immediately and comprehensively to every other part of the system. There was no conditional independence left to contain the failure. The Markov blankets that would have allowed local failures to be absorbed locally had been destroyed in the pursuit of efficiency and return.
The financial system had become more efficient in its connected state: capital flowed more freely, risk could be distributed more widely, returns could be optimised more precisely. It had also become more fragile, because the sparse coupling that once allowed local failures to remain local had been replaced by dense coupling that ensured any failure became global.
This is the overconnectivity paradox: the same connections that increase efficiency in stable conditions increase fragility in unstable ones. The connections that allow benefits to be shared also ensure that costs are shared: including the cost of failures that would, under sparse coupling, have remained contained.
Overconnectivity in organisational systems
The same dynamics appear in organisational settings, though with less dramatic consequence.
A company that integrates its previously autonomous divisions under a single ERP system, shared service centre, and consolidated management reporting has increased efficiency. It has also created tight coupling: a problem in one division now immediately affects every other division through shared resource pools, shared reporting timelines, and shared management attention. The conditional independence that once allowed divisions to absorb local shocks without system-wide consequence has been dissolved.
A school system that replaces locally developed curricula and locally calibrated assessments with a national standardised framework has increased comparability and, arguably, baseline quality assurance. It has also dissolved the Markov blankets that allowed schools to develop pedagogical approaches suited to their specific populations, communities, and contexts. The local self-organisation that produced diverse, context-adapted practice has been replaced by uniform compliance with a central model that cannot be context-sensitive by design.
Overconnectivity and individual psychology
Friston’s Markov blanket framework applies at the scale of individuals as well as institutions.
A person’s psychological stability (their capacity to act with agency and maintain a coherent sense of self in a changing environment) depends on maintaining a functional boundary between their internal states and the external world. They must be able to form intentions, act on them, and observe the consequences without being directly overwhelmed by every signal the environment produces.
Environments that dissolve this boundary (through constant surveillance, ambient notification, social comparison at scale, or the indefinite extension of work into personal time) produce the chronic anxiety that Friston’s framework predicts: the individual’s internal states are no longer conditionally independent of external observation and judgment. Every action potentially feeds back into a visibility system the individual does not control. The Markov blanket is compromised, and with it the self-organising capacity of the individual.
This is not metaphorical. The mathematical structure is identical whether you are analysing a financial institution, a community, or a person. Destroyed Markov blankets produce loss of self-organisation at the scale of the system whose boundary was destroyed. The consequences (increased anxiety, reduced agency, diminished adaptive capacity) are the mechanical output of the structure, not a cultural or psychological pathology.
The scope mismatch between problem and solution
A specific and common form of scope failure occurs when the intervention is targeted at the wrong level of recursion: when the problem exists at one scale and the solution is applied at another.
Obesity as a public health problem operates at the scale of food environments, economic incentives, urban design, agricultural policy, and the structure of the working day. Interventions targeted at individual behaviour change (dietary advice, calorie labelling, personal responsibility campaigns) are applied at a scale several levels below where the critical feedback loops operate. They produce real effects at their scale: some individuals change behaviour in response to advice. They cannot address the population-level problem because the environmental, economic, and structural forces that shape food behaviour operate at scales the individual-level intervention cannot reach.
Homelessness operates at the intersection of housing markets, mental health provision, addiction services, employment structures, and benefit systems. Interventions that address any one of these systems individually (more social housing, better mental health services, improved employment support) produce real effects within their scope. The people who fall through the intersections between systems are not served by any single-scope intervention, because their situation is defined precisely by the gaps between scopes.
Scope mismatch is not always a reason to escalate to a higher level. Sometimes the appropriate response is to recognise that the problem is at a lower level than the current intervention: that what is being addressed at the level of national policy needs to be addressed at the level of community, neighbourhood, or individual relationship. Over-scoped interventions lose the contextual variety that effective response requires.
The diagnostic question is not “what level should we work at?” in the abstract, but “at what level do the critical feedback loops for this specific problem operate?” The answer to that question locates the appropriate scope.
What the answer tells you
If governance is underscoped relative to the problem: The required intervention is scope construction: building governance capacity at the scale where the problem operates. This involves establishing a System 5 at the appropriate level (a shared identity and set of values that define what the system at that scope is for), building the System 4 intelligence function (environmental scanning and modelling at that scale), and creating the System 2 coordination mechanisms between the lower-level systems that currently operate without coordination at the higher scale.
This is the hardest intervention in the toolkit because it requires the creation of something that does not yet exist: governance structures, legitimacy, and shared purpose at a scale that has not previously been governed. But it is also the only intervention that can work. Incrementally improving governance within the existing scope will not address a problem that operates outside it.
If overconnectivity has dissolved necessary Markov blankets: The intervention is scope restoration: deliberately reducing coupling to preserve or rebuild the conditional independence that allows self-organisation at each scale.
Sparse coupling differs from disconnection. The goal is enough connection for coordination and information flow, without so much coupling that every local event becomes a system-wide event. The design question is: which connections are necessary for the system to function, and which connections are destroying the self-organising capacity of subsystems for marginal gains in efficiency?
Concretely: protect local decision authority for matters that do not affect other units. Maintain information boundaries that prevent the surveillance of internal states from becoming a mechanism of external control. Design resource sharing protocols that allow local units to absorb local shocks without triggering system-wide responses.
If the problem and the intervention are at mismatched levels of recursion: Redirect the intervention to the level where the feedback loops actually operate. This typically means either going higher (building governance capacity at the scale of the environment that is shaping the behaviour you are trying to change) or going lower, restoring the contextual specificity and local self-organisation that system-wide interventions have suppressed.
Using the complete toolkit
Questions 1 through 5 form a complete diagnostic sequence. Used together, they cover the structural dimensions of a failing system:
| Question | What it diagnoses | Intervention type |
|---|---|---|
| 1. What is actually being controlled? | Gap between stated and operational goals | Reattach feedback loops to the right variables |
| 2. Where is the variety mismatch? | Capacity deficit in the control relationship | Amplify regulator variety or attenuate system variety |
| 3. Which function is missing? | Absent or broken architectural component | Build the missing function |
| 4. What is the feedback structure? | Loops amplifying rather than correcting | Attenuate runaway loops; damp oscillation; shorten delays |
| 5. What is the viable scope? | Misalignment between system boundary and problem scale | Construct governance at the right scope; restore sparse coupling |
In practice, the questions are not always applied sequentially. A diagnosis may begin with any question and move to others as the picture develops. A clear scope mismatch (Q5) may reveal why the feedback structure (Q4) cannot be corrected, because the relevant loops close outside the governance boundary. A missing function (Q3) may explain a variety mismatch (Q2), because the function that should have been amplifying regulatory variety was never built.
The five questions are lenses, not steps. Each one illuminates a different dimension of the same system. A complete diagnosis uses all five.
Reference
The primary theoretical foundation for this question is the Free Energy Principle and its account of Markov blankets as system boundaries, developed by Karl Friston.
Key sources:
- Friston, K. (2019). A free energy principle for a particular physics. arXiv preprint.
- The CybSoc presentation: “The Physics of Sentience” (Karl Friston, Cybernetics Society).
Related theoretical frameworks:
- Autopoiesis (Maturana and Varela, Autopoiesis and Cognition, 1980): organisational closure as the defining property of living systems, preceding the FEP formulation
- Viable System Model (Beer): System boundaries as the organisational unit of viable systems; the recursion property as the application of Markov blanket logic to organisational design
- Elinor Ostrom, Governing the Commons (1990): empirical demonstration that communities self-govern when their scope boundaries are maintained; collapse when they are dissolved by external intervention