8.1 Function
“System Two is logically necessary to any viable system, since without it System One would be unstable — it would go into an uncontrollable oscillation” (HOE 1979: 177). With this sentence Beer introduces the least familiar of the five subsystems, and he gives it a narrow job. The operational elements of System One are autonomous, but they are not independent. They share customers, suppliers, stocks, facilities and people. When one element changes what it is doing, the others feel the change and adjust, and their adjustments come back to the first element. If each element corrects for the others without any common reference, the corrections can amplify each other, and the system swings back and forth instead of settling. Why not let the metasystem do the damping? Beer asks the question himself and answers it: “Could not the metasystem itself regulate the potential instability of System One, by damping its oscillations? Yes, it could; but in doing so it would necessarily be an autocratic controller, because it would have to deploy high variety on the channel that intervenes in the freedom of System One to manipulate its own horizontal variety” (HOE 1979: 177). The objection rests on variety (see chapter 2.3 on Ashby’s Law). To damp through the command channel, the centre would have to track every interaction between the operational elements. That would flood the command channel and take away the autonomy the model is meant to protect. System Two also intervenes in the freedom of System One, and Beer grants as much. He answers with a limitation: System Two has “the highly specific function of damping oscillations, and nothing else. In fact, then, System Two is a service to System One” (HOE 1979: 177). A barber shows what he means by a service. A haircut removes several possible styles, “I do not however consider that I am being oppressed by my barber, because I gladly expected that he would reduce my hair’s variety in this way. Equally, the viable system engages the services of System Two to cut down the variety of its operational interactions insofar as they are inherently oscillatory — and only to that extent” (HOE 1979: 177). Everything hangs on “only to that extent”. System Two may reduce variety, but only the variety that shows up as oscillation. Beer also keeps System Two apart from power. On the diagram, he concedes, the heavily drawn System Two “looks dominant in the entire model. The cybernetic fact is that it is not dominant; but it is necessary” (HOE 1979: 177). People readily confuse logical necessity with a claim to power, “and that is a mistake. The game of power, if there has to be one, is played on the direct vertical linkages between System One and the metasystem” (ibid.). His standard example is the school timetable. A school with specialist rooms and specialist teachers generates clashes: several classes want the language laboratory in the same period, several forms want the only Russian teacher on the same afternoon. Swaps are requested in a circle, “and so on. The result is: uncontrollable oscillation. Yet somehow or other, schools operate. They have an anti-oscillation device, a System Two. It is called the school timetable” (HOE 1979: 178). Anyone can propose a change against the timetable without setting off a chain of clashes, and if one starts, “the keeper of the school timetable will shortly damp it” (ibid.). Nobody mistakes the timetable keeper for the Head Teacher or the Board of Governors. Those in the System Two role, Beer concludes, are often accused of destroying horizontal variety, “whereas their proper function is merely anti-oscillatory, as the school timetable example makes clear” (HOE 1979: 178).
8.2 Neglected in practice
The timetable is a clean case because its keeper does nothing else: “there is no other function than to be a System Two” (HOE 1979: 179). In most enterprises the function is mixed up with others, and Beer takes this to be the reason it is overlooked. In DTS he puts it bluntly: “The fact is that System Two — the viable system’s anti-oscillatory device for System One is almost totally misunderstood and under-represented in contemporary management technique. It is always present, or the organization would shake itself to pieces” (DTS 1985: 66). He lists cases in which balance is threatened, from East-West détente to “the capital locked into inter-process stocks”, and adds: “All these demand the maintenance of balance in System One” (DTS 1985: 66). Senior Management, he says, does not command the requisite variety to resolve such oscillation “by dictat on the central axis (although many such managements try this impossible trick)” (ibid.). HOE finds the function, unrecognized, in two further places. The first is the personal secretaries of executives. Taken together, “these people between them (which is to say, vertically) constitute an anti-oscillatory device across the horizontal elements of a very powerful kind” (HOE 1979: 179). Unlike the timetable keeper, a secretary may favor her own executive’s engagements. Beer does not claim that executives are run by their secretaries. His claim is that “the firm is employing a damping device, a System Two, of which it is not truly aware — since it has never thought about the necessity for a System Two” (HOE 1979: 179). The second is management accounting. Accountants are resented “down the line” because their constraints look like metasystemic authority. Beer reads them differently: “The fact is that no-one recognizes that a large part of their activity, perhaps eighty percent of it, is purely anti-oscillatory” (HOE 1979: 180). That part gets confused with the small share of their work that imposes mandatory constraint. So Beer divides management accounting into three components that belong to different channels of the model: interventions that constrain variety for legal reasons, interventions for the sake of institutional cohesion (the handling of resources, see chapter 9 on the resource bargain) (HOE 1979: 180), and “(iii) System Two activities, which are purely anti-oscillatory” (HOE 1979: 181). Agreed conventions for the format of information belong in the third group, “Simply because without a large number of agreed conventions, the system would oscillate” (ibid.). So do salary and company-car rules: “Were they not instituted in the name of fair play — that is, precisely to prevent inter-unit oscillations?” (ibid.). His practical proposal is to print the mandatory items on red paper and the System Two items on green, so that everyone can see “that a tiny proportion of the total paperwork machinery is red: and that is the only aspect of the system that genuinely reduces freedom” (HOE 1979: 181). The secretaries and the accountants do what the timetable keeper does, without the title. Their System Two work has to be identified inside existing functions before anyone can design it.
8.3 System Two in BOTF
Brain of the Firm draws System Two more widely (see chapter 4.3): it is “the metasystem subsuming all Systems One” (BOTF 1981: 172). Beer describes its mechanics as “the interlinking of the divisional regulatory centres, and in the corporate regulatory centre”, and adds: “So it would be correct, and even helpful, to think of System Two as an elaborate interface between Systems One and Three. It partakes of both” (BOTF 1981: 172). The neurophysiological analogue is the sympathetic nervous system; BOTF speaks of “the sympathetic (System Two) circuits” (BOTF 1981: 176). The earlier neurophysiological chapters of BOTF draw the line differently. There System Two is “the analogue of the spinal cord” (BOTF 1981: 124), the sympathetic trunk “links together the vertebral nodes of System Two” (BOTF 1981: 129), and in Beer’s example of catching a bus: “The autonomic system, number Three, sits in the middle of the procedure monitoring the effects. The right-hand loop, the sympathetic, supervises the interaction of the organs and the flow of adrenalin” (BOTF 1981: 132). On my reading, the sympathetic system there belongs to System Three and runs alongside System Two, whereas p. 176 identifies it with System Two. Beer does not comment on the difference. BOTF works through an example. Productivity changes in one division, and the divisions that depend on its output are affected. Without a shared, fast signal, they protect themselves. They build up stocks, the financial authorities intervene, and informal rules are adopted locally, until “trust is lost […] and oscillation has set in” (BOTF 1981: 174). System Two is meant to prevent this. The change is relayed “automatically” to the other divisions, and it is “neither fuzzy nor emotive: it is a statistical statement couched in achievement numbers” (BOTF 1981: 174). The corporate regulatory centre receives the same information and reports to System Three where managerial action is needed (BOTF 1981: 175). Here Beer stresses time: “Perhaps the main point of this whole arrangement is its automatic simplicity, and therefore its speed” (BOTF 1981: 175). Under orthodox procedures the matter reaches the corporate level only after the oscillation has set in, and by then the technical control problem has been joined by suspicion and defensiveness. His verdict is sharp. Such oscillatory mechanisms “are a curse of our age — because our age has produced so many large-scale organizations without a System Two. Most of the successful ones I have observed have been entirely unofficial and largely unrecognized” (BOTF 1981: 175). Speed returns in the Chilean work. Beer formulated five principles there, one per subsystem, and the second reads: “The speed of response in an essentially lagged servomechanism is critical: note especially the anti-oscillatory Systems Two” (BOTF 1981: 288). Two versions therefore have to be kept apart. In HOE and DTS, System Two is strictly the anti-oscillatory service to System One, and DTS insists that “System Two is concerned only with the regulation of oscillatory behaviour” (DTS 1985: 144). In BOTF the same term also covers the regulatory centres and the corporate regulatory centre, which filter information upwards to System Three. Both versions share the core function the timetable keeper performs: damping oscillation between the operational elements without turning the damping device into a second command channel. Chapter 9 places System Two alongside the other vertical channels.
8.4 System Two and the metasystem in HOE and DTS
BOTF calls System Two “the metasystem subsuming all Systems One” (BOTF 1981: 172). Neither HOE nor DTS refers to that sentence, to the tall thin box that illustrates it, or to the spinal-cord picture behind it. What both later books do is state for themselves where System Two sits. HOE argues the point in two steps. The first excludes System One: “It is not possible to conceive of System Two as a subsystem of System One” (HOE 1979: 200). A committee in the managerial domain would meet sporadically, while the interactions it would have to damp are continuous; “System Two must be continuously in being” (ibid.). Its criteria, moreover, can be formulated “only by this viable system conceived as an operational unit (part of System One) of the viable system at the next level of recursion” (ibid.). The second step places it: “Thus we have to regard System Two, which is a service to System One, as a subsystem of the metasystem in which System One is embedded” (HOE 1979: 201). Once System Three is named, System Two becomes its subsystem. System Three “is also aware of the anti-oscillatory activity of System Two, since System Two is its own subsystem” (HOE 1979: 202). The recapitulation at the start of Chapter Ten repeats the placement: “We discovered the logical necessity for System Two, outside System One, to damp oscillations internal to System One, and said that it must be a subsystem of the metasystem” (HOE 1979: 251). The same page lists System Two as the first of the complementary channels between System Three and System One (ibid.). Two later passages draw the line more narrowly. “System Two cannot reach into the metasystem: it is dedicated to damping oscillations in System One” (HOE 1979: 259). A few lines on, Beer closes the argument for System Five: “Three-Two-One plus Three-Four-Five is a viable system — where the second group is metasystemic to the first” (ibid.). Discussing two levels of recursion, he writes: “the metasystemic systems Three-Four-Five belong to the management box depicted as a System One at the first level of recursion (and System Two is an appendage to System Three)” (HOE 1979: 316). DTS does not use the word metasystem for System Two at all. It keeps System Two off the command line: “System Two does not lie on the central command axis. Its function is not to command, but to damp oscillations” (DTS 1985: 68). System Three differs from System Two “because it exerts authority on the central command channel”, and it “is responsible for, although it does not conduct, the anti-oscillatory functions of System Two” (DTS 1985: 86). The metasystem, in DTS, is “a metasystem called 3-4-5” (DTS 1985: 132). On my reading, the later books do not take System Two out of the metasystem that BOTF puts it in. HOE says in so many words that it is a subsystem of the metasystem and places it under System Three. What changes is the emphasis. BOTF’s System Two “subsumes” all Systems One; HOE’s is “a service to System One” in the same sentence that makes it metasystemic. HOE is also not uniform: System Two is a subsystem of the metasystem (HOE 1979: 201, 251), yet it “cannot reach into the metasystem” and falls in the group Three-Two-One rather than the metasystemic Three-Four-Five (HOE 1979: 259). A reader can reconcile the two by taking “metasystem” on p. 259 as Systems Three to Five. Beer does not do so himself, and he does not comment on the BOTF formula.
References
- HOE — Beer, Stafford (1979). The Heart of Enterprise. Chichester: John Wiley & Sons.
- BOTF — Beer, Stafford (1981). Brain of the Firm. 2nd ed. Chichester: John Wiley & Sons. (1st ed. London: Allen Lane, 1972.)
- DTS — Beer, Stafford (1985). Diagnosing the System for Organizations. Chichester: John Wiley & Sons.