The Mapping We Teach Is Not the Mapping He Wrote

The mapping we repeat

Most of us learned the Viable System Model with a brain attached. The five systems arrived already fitted with anatomy: a system that is the spinal reflex, a system that is the brain stem, a system that looks outward and is therefore, in the shorthand, the eyes. That anatomy is a large part of what makes the model teachable in a morning. It also lends the model an authority it did not earn on its own. Say that System Three is the brain stem and you are offering more than an image. You are implying that a distinction you could otherwise be asked to justify has already been justified by biology.

So the mapping needs an audit, and the audit belongs at the front of this book rather than in an appendix. If the shorthand we teach differs from the mapping Beer set down, there are three possible explanations, and they carry different consequences. Beer may have been wrong about the anatomy. Later teachers may have simplified, substituted or improved his account without recording that they had done so. Or both may have happened, with the second obscuring the first. Argument cannot choose among these. Reading can: the taught text and the written text, put next to each other.

We have not yet done that reading in a documented way. What follows is therefore a charge we cannot yet prove, with the work that would prove or dissolve it. The charge is that the anatomical shorthand circulating in our teaching materials has drifted from the assignments in Beer’s own text, and that nobody has examined the drift because no one in the community had reason to look. The evidence that would settle it is unglamorous: dated captures of the sources we actually teach from, quoted exactly, set beside a paginated reading of the founder’s text. One of those sources is Jon Walker’s guide for co-operatives, written in 1991 and circulated as The Viable Systems Model: A Guide for Co-operatives and Federations (version 2.21, 2001) and now as The Viable Systems Model Guide (3rd ed., version 3.2, 2020), which gives each of the five systems a part of the body or brain. The Metaphorum VSM page gives no such mapping; it says only that the model “is inspired by the way the brain co-ordinates the muscles and organs”.^[1]

The mapping Beer wrote

The reconstruction has to come from Chapters 7 to 14 of the second edition of Brain of the Firm, read as text, and not from the diagrams, slide decks or summaries descended from it. That is stricter than it sounds. Much of what the community treats as Beer’s account reaches us through intermediaries, and an intermediary that compresses a paragraph into a label has already made an editorial decision on our behalf.

Read directly, the assignments are more specific and less tidy than the shorthand suggests, and they are made across several chapters. Chapter 7 sets out a five-level hierarchy on the vertical axis (pp. 97-99), drawn as five control echelons (Figure 14, p. 99) that, Beer says, “model the somatic nervous system of the body” (p. 98), and the chapters that follow attach the system numbers. System One is the lateral axis “which mediates afferent and efferent information” (p. 98; see also pp. 124, 127). System Two is the spinal cord, its vertebral nodes linked by the sympathetic trunk (pp. 124, 129; see also pp. 108, 176). System Three is the medulla and pons, with the mesencephalon folded in (pp. 98, 140-141), and it handles “the parasympathetic information circuits” (p. 176). Systems One to Three together are Beer’s model of autonomic control, “concerned, automatically, with the regulation of internal stability” (p. 76; see also p. 110). System Four is the diencephalon, with the thalami, and the basal ganglia, with the ascending reticular formation as its core (pp. 98, 148). System Five is the cerebral cortex (pp. 98, 144). The receipt at the end of this chapter gives Beer’s wording for each. Page numbers here and below are those of the Wiley second edition (Chichester, 1981).

Two features of the written mapping matter before any verdict. Beer names structures, and he groups them by function: the five levels, he wrote, follow “the major functional differences involved” (p. 98). And his assignments are compound: System Four is a set of structures with a stated core, so a teaching label that names any one of them selects from Beer’s claim instead of summarizing it. The boundaries between the systems also move from chapter to chapter. On p. 132 the sympathetic loop appears in Beer’s account of “The autonomic system, number Three”; on p. 176 the same circuits are “the sympathetic (System Two) circuits”. The hypothalamus is “the highest element in System Three, or the lowest element in System Four” (p. 141), and the ascending reticular formation, “the core of System Four” (p. 148), runs “out through the highest structure of System Three — the mesencephalon” (p. 141). Whether the selection distorts him is exactly what the audit is for.

Beer also said how much the count itself should carry. Introducing the five levels in Chapter 7, he described “what seems to be a five-level hierarchy of systems” and explained the hedge: “I say ‘seems to be’ because the choice of five (rather than some other number) is somewhat arbitrary” (p. 98). On his own account, the anatomy did not settle how many systems there are.

He also said what kind of claim the mapping was, and it was more than an illustration: “Models are more than analogies: they are meant to disclose the key structure of the system under study” (p. 75), and “There can be no interest in analogy at this point” (p. 87). Yet the body was one choice among several. Beer could have used “the amoeba, or a whole animal species” as the model, and “the results are the same” (p. 75). In Decision and Control he called the brain model “a homomorphic mapping of the brain which is by definition a simplification,” and “a simplification of which the limitations are understood,” and he held that a model which leaves out relevant elements “will lose in utility, but it cannot lose in validity.”^[2] The audit therefore has to say which claim it tests. On Beer’s own account an anatomy that fails costs the model usefulness, not validity.

Whether Beer’s neuroanatomy survives current science is a later question, and a different one. The prior question, the one the community has skipped, is what he actually claimed, since we cannot evaluate a mapping we have not first stated correctly. Chapter 2 takes up what happened to System Four, where the compression has gone furthest. Chapter 3 shows the same pattern at work on a borrowed law instead of a borrowed anatomy.

The shelf behind the diagram

Beer’s bibliography is partly system-tagged, which makes it unusually useful evidence: for some entries it tells us which reading he associated with which part of the model. Under the heading “Specific Neurological Systems” in the Select Bibliography (pp. 405-407), the entries by Mitchell and by Eccles, Ito and Szentágothai are listed “On Autonomics (Systems One-Two-Three)”, Brodal “On ‘arousal’ mechanisms (System Four)”, and Burns and Sholl “On the Cortex (System Five)”. General entries such as Netter and Magoun carry no tag, and no source is tagged to System Two or System Three alone. A short account of what each supplied to the mapping would let us reconstruct the materials that were actually on the desk when the mapping was made.

Three things are easy to collapse into one here, and they need to be kept apart: the anatomy Beer explicitly described, the scientific literature he explicitly cited, and the later stories readers have projected backward onto him. The first is a matter of quotation, the second a matter of the bibliography. The third is a matter of our own habits, and we are least equipped to notice it, because a story we have been telling for years feels like something we read.

The bibliography does not establish that Beer read his sources correctly, or that he used them as their authors intended. What it does establish is more modest and more useful: the range of positions available to him, and the vocabulary he had reason to adopt.

The Beer we should not invent

A suspicion comes up often enough in our own conversations: that the model rests on a crude layered-brain story or on a simple reflex arc, and that this is why the anatomy now looks dated. We tested it before repeating it, because criticism aimed at a position the author did not hold is worse than none at all. It discredits the critique and protects the original.

The reflex-arc half fails in Beer’s own words. Chapter 8 prints the arc as Figure 15, a “Familiar simplified account of the reflex arc” (p. 104), and then sets it aside: “Even so, this reflex arc explanation will not do, because it is too simplified” (p. 105). The layered-brain half fails as far as a search of the text can take it. We searched the same edition for MacLean, “triune,” “limbic,” Sherrington, Pavlov and Moruzzi, and found none of them. A negative search is only as good as its record, so the record is part of the result.

• Edition searched: Brain of the Firm, 2nd ed., John Wiley & Sons, Chichester, 1981 (calibre copy, a scan of the printed book)
• Search surface: both: the full text, including glossary, Select Bibliography and index, searched through the scan’s OCR text layer
• Search date: 2026-09-23 (OCR text layer); 2026-09-29 (printed index, read in the page images)
• Search terms and variants: case-insensitive substrings MacLean (also “Mac Lean”), “triune,” “limbic,” Sherrington, Pavlov (covering “Pavlovian”) and Moruzzi, each also as a truncated stem to allow for OCR errors
• Result: MacLean: 0 hits. “Triune”: 0 hits. “Limbic”: 0 hits. Sherrington: 0 hits. Pavlov: 0 hits. Moruzzi: 0 hits. (Magoun, Moruzzi’s co-author of 1949, appears once, as the Select Bibliography entry for The Waking Brain, p. 406. Moruzzi’s absence says little: Beer cites Magoun’s 1958 synthesis and builds on the ascending reticular formation, “the core of System Four” (p. 148).) The printed index (pp. 411-417) was also read page by page in the scan, to catch what the OCR layer might miss: none of the six terms has an entry, and neither has Magoun

The finding is precisely this and no more: these names and formulations do not appear in the stated search surface. Absence in a text is not absence from a mind. Beer worked in a period when several of these ideas were ambient, and he could have absorbed a framework without naming its author. What the search can do is stop us attributing to him a specific doctrine he did not invoke. That aim is narrow, and it is the aim of the whole book: to criticize the mapping he wrote, and leave alone the more convenient mapping assembled for him afterward.

What changed in the brain

Only now does the scientific question become answerable, and it has to be answered element by element. “Outdated” has to be earned for each structure; it cannot be a mood applied to a text of its age. For every anatomical claim we ask whether the named structure is still described substantially as Beer described it; whether its function has been revised, distributed or subdivided; whether current work rejects the assignment outright; whether the claim has survived but become incomplete because later work added mechanisms Beer could not have known; and whether the revision matters for the organizational function assigned to that VSM system.

The last question keeps the audit honest, and it requires holding two verdicts apart: the standing of the neuroscience, and the consequence for the model. A biological analogy can fail while the organizational distinction it illustrated remains sound. Then we have lost an illustration and kept the argument. An anatomical claim can also survive intact without validating the organizational use made of it, since the anatomy never licensed that inference in the first place. Collapsing the two verdicts is how a metaphor acquires the standing of a finding, and it is how we have been reading this mapping. Chapter 9 asks why the collapse has cost us so little for so long.

Receipt: an audit we can publish in months

What this chapter delivers is narrower than another general verdict on the use of metaphor in management cybernetics: a bounded, reproducible audit that a small group could complete in months and that anyone could then check or overturn. It covers all five systems. The current-neuroscience column, and the status codes that depend on it, are to be commissioned from someone competent to read that literature. The commissioned reader starts from one named outside text, Principles of Neural Science (6th ed., 2021), and may confirm or replace it.^[3]

The four status codes are defined before use, and each must be defensible on the row’s own evidence:

• Refuted: current evidence contradicts the anatomical or functional assignment.
• Revised: the assignment retains a defensible core, but its formulation must change.
• Unaffected: the relevant claim remains serviceable at the level of precision Beer used.
• Superseded-with-addition: the claim is not simply false, but later findings add mechanisms or structures that materially change the comparison.

The method used here is the method of the book. Take an inherited claim, recover the founder’s actual text, consult the current science, and specify the smallest piece of work that would settle the matter. Part Two applies it where neighboring literatures have moved on without us; Part Four asks what we lose by the delay; Chapter 28 states what would count against the argument made throughout.

We have taught a compressed mapping, inherited its authority and rarely checked the text beneath it. That is a description of ourselves. The correction begins with an exact account of what Beer wrote, continues with an element-by-element scientific audit, and ends with either a version of the mapping we can defend in front of a neuroscientist or a metaphor we should stop using.