Last updated: 2026-10-10

U
Undergraduate level

Does an Organisation Have a Self-Model?

An organisational chart is routinely wrong about who actually makes a given decision, who actually gets asked first, and which reporting line people actually use when something urgent happens. It remains, despite being wrong in this way, one of the most causally powerful documents in the building. People defer to it. Budgets get allocated through it. It determines who's treated as authorised to speak for a department in a meeting they weren't invited to. A description that's inaccurate about the organisation and still shapes what the organisation does next is not a passive picture. It's part of the mechanism.

Stafford Beer's Viable System Model FoundationalKnowledge that endures for decades — core principles

Stafford Beer's Viable System Model gives this phenomenon its most rigorous treatment, built from cybernetics rather than from organisational chart-drawing convention. Beer's question is what any organisation needs, structurally, to remain viable — to go on existing as a coherent, self-regulating whole — regardless of what it's actually for1. His answer identifies five necessary subsystems. System 1 is the operational units that do the organisation's actual primary work. System 2 coordinates between those units, damping the oscillation and conflict that would otherwise arise from several operational units acting locally without reference to each other. System 3 regulates the operational units as a whole from the inside, optimising their combined performance, with a specific audit channel — System 3* — that can bypass the normal reporting line to check directly what's actually happening on the ground. System 4 looks outward: scanning the environment, modelling plausible futures, bringing back what the organisation would otherwise have no way of noticing from inside its own operations. System 5 holds the balance between System 3's internal focus and System 4's external one, and carries the organisation's overall identity and purpose — the final arbiter of what the organisation as a whole is for.

The structure is recursive by design, not by accident: each of a viable system's own System 1 operational units is, examined closely enough, a viable system in its own right, containing the same five subsystems at a smaller scale. A department is a viable system inside a division that is a viable system inside a company. Recursion is what lets the model describe an organisation of any size without needing a different theory for each layer.recursion mirrors metacognition

A Self-Model Built Into the Structure, Not Just Described by It FoundationalKnowledge that endures for decades — core principles

System 5, on this account, is not a department or a person — it's a function, and the function is literally self-modelling: holding a representation of what the organisation is and is for, tested continuously against System 3's internal report and System 4's external scan, and used to arbitrate when the two pull in different directions. An organisational chart, a strategy document, a brand statement, a risk register are all partial, visible expressions of whatever is currently occupying that function, whether or not any one of them accurately describes it. That's what makes the opening example more than an isolated quirk: a chart doesn't need to be accurate to be doing System 5's job, because the job isn't description, it's arbitration — and an inaccurate chart still arbitrates, it just arbitrates badly.

The Recursive Formulation FoundationalKnowledge that endures for decades — core principles

Put the two previous sections together and a stronger claim follows, one Beer's own structure supports directly even though this specific formulation is our own: an organisation's self-model is not a report produced about the organisation from outside. It's part of the actual mechanism by which the organisation maintains its boundaries, coordinates its component parts, and continues to count itself as the same organisation over time. Remove System 5's self-modelling function entirely, and System 3 and System 4 have no way of resolving a genuine conflict between internal efficiency and external adaptation — not because nobody's in the room, but because nothing is holding the question of what the organisation is for open and answerable at all.

Narrative Identity: Staying "We" Through Discontinuity FoundationalKnowledge that endures for decades — core principles

Staff leave, buildings change, products disappear, ownership changes, policies reverse — and the organisation keeps saying "we" through all of it. Dan McAdams' account of narrative identity treats this continuity, for an individual person, as something actively constructed rather than simply found: a life story assembled and continually revised from selected episodes, not a complete record replayed on demand2. This site's own page on learning logs and narrative identity already applies that idea carefully at the scale of one learner, distinguishing a kept record from the story told about it afterwards. Extending the same distinction to an organisation is a move we're making, not McAdams' own — his theory is about one person's life story, not an institution's. On that extension, an organisation sustains its own "we" through founding narratives, remembered crises, celebrated successes, institutional heroes and villains, and an account of how it does things here, selected and re-told rather than mechanically preserved. That raises the same questions at organisational scale that narrative identity raises for a person: how much can change before it's no longer the same organisation, who's entitled to narrate its past, whether an organisation can misremember itself, and whether rival subcultures inside one organisation can maintain genuinely incompatible versions of its own story.

Salomon v A Salomon & Co Ltd, the 1897 House of Lords decision that settled the modern principle of corporate separate legal personality, held that a properly incorporated company is a distinct legal entity from the people who own and run it, with its own rights and liabilities regardless of their intentions3. That status is real and consequential — it's what lets a company own property, enter contracts, and be sued in its own name rather than its directors' — but it's a normative and institutional status, not evidence of anything psychological. Incorporation gives an organisation legal capacities. It doesn't, by itself, show that there's a unified experiencing subject corresponding to the legal entity, any more than a sports team being a legal person for tax purposes shows that the team itself feels anything about losing. Keeping this distinction sharp matters directly for a later page in this part, which takes the harder question — whether organisational cognition is accompanied by anything it is like to be that organisation — on its own terms rather than borrowing an answer from company law.

Strongest Objection: the Model Is a Design Target, Not Always a Description FoundationalKnowledge that endures for decades — core principles

Beer's model is explicitly normative — it describes what a viable system needs, which is also a diagnostic tool for finding where a real organisation is failing to provide it, not a claim that every real organisation already has a single, coherent System 5 quietly holding its identity together. Most real organisations have several competing centres that each act as though they're System 5 — a board, a founder, a dominant department, a regulator — without having resolved which one actually is. A later page in this part takes up directly what happens when an organisation's apparent single self-model turns out, on inspection, to be one subculture's version winning out over others rather than a genuinely shared one.

Provisional Conclusion FoundationalKnowledge that endures for decades — core principles

An organisation does have a self-model, in a stronger sense than "a document that describes it": Beer's System 5 function, whatever currently occupies it, actively arbitrates between internal efficiency and external adaptation and is part of the actual mechanism holding the organisation together over time, not a report filed about it afterwards. That self-model is sustained the way a person's is — through a selectively told, continually revised story — and is legally recognised in a way that settles practical questions of ownership and liability without settling the much harder question of whether anything is experienced on the other end of it.

Questions for Further Thought

  • Can an organisation's System 5 function be occupied by more than one, genuinely competing representation at once, or does Beer's model require that one answer eventually wins for the organisation to stay viable at all?
  • If an organisational chart's causal power comes from people deferring to it rather than from its accuracy, what would it take to replace an inaccurate chart without a period of real organisational confusion while the new deference pattern establishes itself?
  • Is there a meaningful difference between an organisation revising its own narrative honestly, in light of new evidence about its past, and an organisation simply retelling its story to suit whoever currently holds System 5?

Further Reading

  • Beer, S. (1981). Brain of the Firm: The Managerial Cybernetics of Organization (2nd ed.). John Wiley & Sons.
  • McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5(2), 100–122.
  • Salomon v A Salomon & Co Ltd [1897] AC 22 (House of Lords).

References

  1. Beer, S. (1981). Brain of the Firm: The Managerial Cybernetics of Organization (2nd ed.). John Wiley & Sons. ↩

  2. McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5(2), 100–122. ↩

  3. Salomon v A Salomon & Co Ltd [1897] AC 22 (House of Lords). ↩