---
title: "content"
description: "On the diagrams of Diagnosing the System for Organizations, every line that crosses a boundary carries a red diamond: “TRANSDUCTION is the word we need…"
url: "https://wiki.cybernetic-futures.org/pure-vsm/part-02-the-viable-system-model/ch12-channels-transduction-and-the-algedonic-signal/ch12-channels-transduction-and-the-algedonic-signal-content"
modified: "2026-11-12T15:26:26.028Z"
language: "en"
site: "Cybernetic Futures"
---

# content

## 12.1 Channels and transducers
On the diagrams of *Diagnosing the System for Organizations*, every line that crosses a boundary carries a red diamond: "**TRANSDUCTION** is the word we need: 'leading across'. Each red blob is a *transducer*" (DTS 1985: 47). Beer explains why he draws channels and transducers as separate things: "The point is to draw a clear distinction between channel capacity and transduction capacity: they are not at all the same thing. People think too loosely about 'communicating the message', as if any sort of connexion must be able to do the job" (DTS 1985: 47). The preceding chapters described the viable system as subsystems joined by lines. Here the lines themselves are the subject, and Beer gives them requirements of their own in the Principles of Organization of *The Heart of Enterprise*. The First Principle, on requisite variety between management, operation and environment, is listed with the others in chapter 3, and the underlying Law of Requisite Variety is the subject of chapter 2. The Second and Third concern the channels.
The Second Principle requires that the channels between management unit, operation and environment "must each have a higher capacity to transmit a given amount of information relevant to variety selection in a given time than the originating sub-system has to generate it in that time" (HOE 1979: 99). This follows from the variety measure. If variety counts the possible states of a system, a channel that carries fewer distinctions per unit of time than the sending subsystem generates will lose some of those states in transit. The receiver then regulates a simplified image of the sender, and the requisite-variety balance fails even though the two subsystems are formally linked.
The Third Principle adds a separate condition. Every connection crosses a boundary, and "because the 'language' of each subsystem is unique to it, messages have to be 'translated' when a boundary is crossed" (HOE 1979: 101). Beer calls the mechanism at the boundary a transducer, "because it 'leads across'", and formulates the principle: "Wherever the information carried on a channel capable of distinguishing a given variety crosses a boundary, it undergoes transduction; and the variety of the transducer must be at least equivalent to the variety of the channel" (HOE 1979: 101). A channel of ample capacity does not help if the translation at its ends is poorer than the channel. The red diamond in DTS marks the place where that second failure can happen.
Beer applies the principles well beyond the horizontal loops where he first introduced them. The Three–Four homeostat (chapter 11), like every other loop, "MUST OBEY THE FOUR PRINCIPLES OF ORGANIZATION" (DTS 1985: 117). He continues: "Nor can it be over-emphasized that they operate on all **vertical** homeostats (such as Resource Bargaining), as well as on the horizontal ones where they first became familiar" (DTS 1985: 117). In the same passage he warns against a common institutional answer for the Three–Four link: "But, oh please, do not simply constitute those red dotted lines as a bunch of committees. **They will not have Requisite Variety**" (DTS 1985: 117). The Fourth Principle requires that the first three be "cyclically maintained through time, and without hiatus or lags" (HOE 1979: 258). It applies to channels as well, so a channel that meets the capacity requirement only in periodic snapshots does not meet it.
## 12.2 Why an algedonic signal is needed
Every channel on the vertical axis carries filtered information. System Two damps oscillation, System Three aggregates and bargains, and the metasystem receives condensed reports. Without that attenuation System Five would be flooded. Yet the filtering that protects the metasystem from overload can also hide from it what it most needs to know, and Beer takes up this difficulty directly.
In DTS he describes it as a problem of staying awake. Somnolence is "an occupational hazard of System Five", because with all the filters on the main axis the metasystem might hear "the whole organism droning on, and simply 'fall asleep'. For this reason, a special signal (I call it **algedonic**, for pain and pleasure) is always identifiable in viable systems. It *divides* the ascending signal \[...\] coming from System One, and uses its own algedonic filter to decide whether or not TO ALERT SYSTEM FIVE" (DTS 1985: 133). He then gives the range of forms such a signal can take: "The cry is 'wake up — danger!' Informal extreme: the roof has collapsed — phone the boss. Formal extreme: the four-minute missile warning" (DTS 1985: 133). As elsewhere in the workbook, the reader gets a task: to list the organization's algedonic signals and to consider whether informal ones should be formalized (DTS 1985: 133).
The passage fixes two features of the structure. The algedonic signal has no separate source; it is split off from the same ascending stream that feeds the ordinary filters. And it has its own filter with its own criterion: not whether a report should be summarized, but whether it should wake the top.
*Brain of the Firm* places this mechanism in the neurophysiological model. System Three contains an "arousal filter which models the reticular formation of the brain stem"; the collateral fibres feeding the algedonic system are shown in Figure 32 (BOTF 1981: 178). Beer is specific about what this filter must achieve: "Statistical criteria must be established to ensure that ascending algedonic information on the vertical axis, of whatever kind, is not simply absorbed by the P-Q-R-S homeostat in the performance of its own function, for it would then be lost in the upward reports from point P — which, after all, will be about the effective functioning of the homeostat and not about discrete internal events. Those are the ones we seek, algedonically, to monitor" (BOTF 1981: 178–179). Reports on how well the regulator is working are one thing. Signals about particular events the regulator may be failing to handle are another, and they must not be averaged into the first. BOTF adds a second arousal filter in System Four, which "incorporates the new external information in the ascending stream" (BOTF 1981: 183).
## 12.3 Algedonic signals cross recursions
In the final chapter of *The Heart of Enterprise* Beer returns to the algedonic signal and adds the feature that sets it apart from all other channels. He starts from the alarm system of the book's chapter 14 ("Calm and alarm", see 12.4 below) and first separates it from ordinary reporting: the alarm signals "do not belong to the family of management information systems at all. These systems are maintaining calm; but they do so by every kind of aggregation — thereby supressing \[sic\] the signals that point to instability. The alerting system is orthogonal to the calming system, in fact" (HOE 1979: 406). He then names them: "it will be useful to borrow nomenclature from *Brain*, to call these alerting signals **algedonic**" (HOE 1979: 406).
The two systems work on the same data and differ in their filters. "The data are the same; the information is different. It is because the two types of filter act in different ways. The traditional set of (MIS) filters says: 'take no notice, it will all come out right in the end'. \[...\] The algedonic signal screams: 'ouch; it hurts!' No wonder that the enterprise awakes" (HOE 1979: 407).
From one recursion Beer moves to the nest of recursions. Because the lower recursion is autonomous, the higher sees it as a "muddy box, into whose management information system it will not enter" (HOE 1979: 407). If every level reported only its own calm, the whole would appear calm whatever happened below. Beer calls such a "spurious 'metacalm'" an "extremely lethal risk" (HOE 1979: 407), and his answer is the special property of the algedonic channel: "The postulate is that algedonic signals, alone of all the signals in the enterprise, are organized to *break through* recursivity" (HOE 1979: 407). All other channels respect the boundaries of recursion that the model otherwise insists on (chapter 6). The algedonic channel is the one designed exception.
The exception works in both directions. If signals from lower recursions are not dealt with at their own level, "they cause mayhem at the senior management level"; Beer's example is board meetings that discuss tea-breaks (HOE 1979: 407). If, on the other hand, the top did not respond to unblocked algedonic signals, "the viable system would *lose its cohesiveness*" (ibid.). Here Beer ties the algedonic channel to the Law of Cohesion discussed in chapter 13: "In the Law of Cohesiveness we found the variety equation that stops the enterprise from blowing apart. Now we have its mechanism" (HOE 1979: 407). The Law states a variety balance between recursions, and the algedonic signal is the route by which a failure of that balance becomes known at the level that can act on it. Beer at once adds a caution: "just because the mechanism is there, it does not mean that its cybernetic design can be disregarded" (HOE 1979: 407). Chapter 15 returns to the link, because Beer calls the failure of "inter-recursive algedonics" the ultimate pathology of the viable system (HOE 1979: 408).
## 12.4 Calm and alarm
Beer lets the manager sleep. "If the job is being well done by the operation of cybernetic laws, \[...\] the manager may go to sleep. \[...\] What, then, is his job? It is to be alert to incipient instability — and to take action about *that*" (HOE 1979: 276). Calm is the intended state, and the algedonic signal presupposes it. Beer also describes how calm turns dangerous. When aggregates are taken "one *year* with another", at System Five "nothing much will ever appear to happen. Therefore sleep supervenes. If we are not very careful, sleep turns into coma; and coma becomes death. It was to deal with this lethal calm that the measurement of instability was introduced. Calm we must have, but not lethal calm" (HOE 1979: 406).
The usual measures do not detect incipient instability. In the summary of the book's argument Beer writes that most managerial measures are "projections of good health intended for public consumption", whereas the manager "needs to be alerted, instantly alerted, to any evidence of incipient instability in any of the loops that are supposed to display equilibrium. But nobody is measuring stability. It can however be done. The key to the problem is statistical filtration" (HOE 1979: 275). The same summary notes that the manager "must choose both a response and a style of response", which "jointly succeed in suppressing incipient instability within an appropriate time-scale" (HOE 1979: 276).
Statistical filtration goes back to Beer's earlier work. *Management Science* describes it in the context of cost control: "Perfectly ordinary accounting machinery (not to say computers) can be organized not to print the results they compute that lie within the expected variation about the mean" (MS 1967: 47). Only the variances outside statistically set limits are reported, so that "the manager gets one page of cost returns for every twenty he used to get, and the only information that has been suppressed is statistically guaranteed to be of no interest" (MS 1967: 48). In 1967 this is a reporting technique. In HOE the same principle becomes the basis of a signal that interrupts the calm of the metasystem, and in BOTF it defines the criterion of the arousal filter (see 12.2). The indices on which such filtering operated in Chile are described in chapter 16.5.
## 12.5 The algedonic meter
In Chile Beer carried the algedonic principle into a different setting: feedback from citizens to government. In *Brain of the Firm* he describes an "algedonic meter", a simple analogue device with which people could signal their satisfaction without having to rationalise it. He gives the reasoning: "An algedonic loop works on a non-analytic, non-articulated scale of 'pleasure and pain'. It uses the brain as its computer, structured and programmed by individuality and life experience, to produce an output indicating a degree of satisfaction that does not have to be 'rationalized'. This is at last an attempt to provide a metric for Aristotle's eudemony, or 'state of general well being'" (BOTF 1981: 283; figure 45 on p. 281). The device had a knob that changed the proportion of a "happy/unhappy" display, and the person holding it "does not have to explain anything — only to respond algedonically, which people may be observed to do all the time" (BOTF 1981: 283).
Beer reports that a prototype of ten meters, "linked by a single wire in a loop through a large summation meter", was built in England and taken to Chile for experiments. The summation meter "had to be very heavily damped" because each station could move at any time (BOTF 1981: 284).
The meter shows what algedonic means for Beer outside the enterprise. It is a channel with a low-variety signal, a position between satisfaction and disquiet, that bypasses the reporting structure and can reach the top directly. In that respect it does the same work as the algedonic signal in the firm: it keeps the metasystem from mistaking the absence of reports for the absence of trouble, and so from falling asleep over a calm that is only on paper. The Chilean context of the proposal is treated in chapter 16.6; the term algedonic itself is introduced in chapter 2.8.
## 12.6 The model of calm, alarm and response
Chapter 14 of *The Heart of Enterprise*, "Calm and alarm" (HOE 1979: 367–401), gives the full model behind the signal discussed in 12.3 and 12.4. Beer builds it in four figures, all titled "Handling incipient instability" (HOE 1979: 383, 389, 391, 393). He starts from two properties of instability: "The first is that instability may set in anywhere in the behaviour of the organization, and at any time. The second is that there are always pre-symptoms of the fact" (HOE 1979: 374). The ordinary reporting rhythm of a business therefore does not settle the question. What matters is "the interval between a warning and a disaster" (HOE 1979: 375), and Beer calls the task breaking "the time barrier" (HOE 1979: 376). His conclusion: "calm is a function of alertness. It comes from being *poised* to read the signs of incipient instability, in the quiet confidence that instability can be averted" (HOE 1979: 376).
Figure 63 shows stage 1 (HOE 1979: 383). The metasystem sets the criterion for a meter that watches one "point of stability" in the system, a feature judged important to viability. The point may move within a circle of "physiological limits"; if it leaves the circle, the system is incipiently unstable (HOE 1979: 381). The manager sits in his "psychological space" at a "point of calm" (ibid.). The massive variety of the system reaches him through three filters. The *resolution filter* is designed by the metasystem and set for one level of recursion; it "does not know whether John or Mary is operating that machine today" (HOE 1979: 382). The *perceptual filter* is designed, "perhaps from a lack of self-knowledge, by the manager himself", and may attenuate or amplify (ibid.). The *variety filter* is "the formal way of attenuating variety: it houses the bureaucracy" (ibid.). The manager's "general approval" flows back to the system through arrow A (HOE 1979: 384). Each pass through this loop makes him calmer and reinforces the filters, which as learning systems "may easily learn the precisely wrong thing"; calm reached by not being told "is spurious because the viable system is robbed of its metasystemic component — which is a coma" (HOE 1979: 386). A "*special mechanism*" is therefore required, because otherwise "the triple-filtration will take out the vital signal along with all the others that are trivial" (HOE 1979: 384). In practice, Beer observes, that mechanism is the personal network ("I say, Harry, have you heard what's going on down at . . ."), and he lists its costs (HOE 1979: 384–385).
Figure 64, stage 2, replaces the network with "a device, operating in real-time, that monitors chosen points of stability by statistical methods" (HOE 1979: 386; figure on p. 389). When the meter detects instability, ideally "*before* the point actually leaves its circle", it generates "the *alarm signal*, which impinges on the manager's calm" (HOE 1979: 386). The manager then follows a three-phase trajectory. "The trajectory T1 is the trajectory of judgment": he chooses between a range of responses (ibid.). "The trajectory T2 is the manager's *choice of style* as to the projection of his judgment into the system" (HOE 1979: 387). The "effectiveness trajectory is a product of T1 and T2"; if it returns the point to its circle, the alarm stops and he returns on T3 "to his (newly defined) point of calm" (ibid.). An outer loop adds learning: a comparator records the deviation and "*how long it took*" to restore stability, and a feedback modifier uses this to redesign the variety filter, so that the bureaucracy could become "more and more, instead of less and less, effective to its supposed purpose of informing management" (ibid.).
Beer then qualifies the separation of T1 and T2. Judgment and style are not independent: a draconian measure attenuates the variety of available style, and a benign style chosen only for its own sake makes draconian measures ineffective (HOE 1979: 387–388). Figure 65, stage 3, draws these mutual inhibitions as a web between choice of response and choice of style, and adds a second comparator whose readings should reveal the inadequacies of the perceptual filter (HOE 1979: 388; figure on p. 391). From the whole picture Beer draws two parameters for classifying managerial situations: variety and relaxation time (HOE 1979: 390). On the second: "if the system has not returned to stability before another shock drives it once more towards instability, then regulation is impossible. So is learning impossible, so adaptation, so evolution" (ibid.).
Figure 66, the final stage, is a procedure given in words (HOE 1979: 393). Its criterion of effectiveness is that a point becoming incipiently unstable "must be returned to stability (within acknowledged physiological limits) within a time-scale that does not allow immanent disaster to become actual", recognized by the return of the manager to his "POINT OF CALM" (ibid.). The procedure eliminates actions that cannot meet the criterion in time, estimates the relaxation time of the rest, matches each to a style, drops those whose effective style is unavailable or implausible, and weighs the stress of an unnatural style against the speed of response. If that fails, the situation is recast; if that fails too, the last line reads "RESIGN. (And do not fail to do so.)" (ibid.). Beer names the parameters of effectiveness as "judgment, style, and relaxation time" and says they derive from the model of any viable system (HOE 1979: 392). He ties the model to closure: a senior manager who legislates for "his own ambiance of calm and alarm" and classifies his decisions by "the tripartite model here set out" may "operate closure in the sense defined" (HOE 1979: 394). The chapter ends with a bar dialogue (HOE 1979: 395–401).
On my reading, this model is the source of the algedonic signal in 12.3: the alarm signal of Figure 64 is what the final chapter of HOE renames algedonic (HOE 1979: 406), and the feedback modifier is Beer's answer to the calming bias of the filters described in 12.2.

## References

- **MS** — Beer, Stafford (1967). *Management Science: The Business Use of Operations Research*. Garden City, NY: Doubleday, in association with Aldus Books, London.
- **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.