Marian Mazur, Cybernetyka a zarządzanie (‘Cybernetics and Management’), Warsaw 1969, first edition (Training and Publications Department of the MSW (Polish Ministry of Internal Affairs), print run marked as intended for internal use), 181 pages, 46 figures, a bibliography (25 items) and a subject index.
The book is a course in cybernetics for practitioners of management. In it Mazur presents, in simplified form and with many examples from everyday life and anecdotes, the basic concepts of cybernetics — some of which (couplings, information, the autonomous system) he had previously set out rigorously in Cybernetyczna teoria układów samodzielnych (‘A Cybernetic Theory of Autonomous Systems’), published in 1966 — and combines them with two practical threads: a survey of the decision theory of the day (operational research, linear programming, game theory, network diagrams) and his own thesis about the role of the manager.
The book falls into two parts. Chapters 1–5 provide the apparatus: the most general concepts (coupling, information, control) and a model of the human being as an autonomous system; each of Chapters 2–5 also contains a practical conclusion for anyone who is in charge of people. Chapters 6–8 apply this apparatus to action, decisions and leadership. The computational passages are set in small type — the editors' preface treats this typeface as a mark of difficulty, not of lesser importance — and the preface cautions that the discussion of leading people necessarily simplifies reality: it offers a sense of direction among principles, not a prescription for direct use.
The site's dictionary contains no definitions from this book (it covers books from 1966, 1970 and 1976). The concepts Mazur discusses in this book do, however, have entries drawn from his later or earlier works — for each chapter we give links wherever the same concept is meant. The definitions in the dictionary come from those books, and their wording sometimes differs from the wording here.
Chapter 1. On cybernetics (pp. 7–10)
This short introductory chapter establishes what cybernetics deals with: control, in every form it takes. Mazur arrives at this concise formula by shortening the title of Wiener's 1948 book, dropping the elements he regards as superfluous. He sees the source of its strength in its perspective: cybernetics describes functioning and leaves the material aside, so its concepts apply in every field, from technology to social life. The history of the word itself — from the Greek term for the art of directing to nineteenth-century attempts to make it the name of a science of governing (Ampère, Trentowski) — and a roll of forerunners serve as background. The author presents Chapters 2 and 3, on coupling and information, as the source of the methods that cybernetics can offer management today — a field that in the old sense of the word was its entire subject and is now one of its many topics.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: control.
Chapter 2. Coupling (pp. 11–28)
The chapter's most important idea is a practical one: anyone who recognises the kind of coupling can foresee how a phenomenon will unfold without any calculation, and in dealings with people the direction in which mutual influence gathers momentum depends largely on how it is started. Since the other side's response cannot be predicted, Mazur considers all the possible reactions and shows that a well-disposed start cannot end badly, nor an ill-disposed one well. The illustration is an extended scene between a superior and a subordinate in two variants, in which the same situation develops in opposite directions, and in both the superior takes for an accurate forecast what he himself has brought about.
The chapter's apparatus is simple. An influence running in one direction only is simple coupling; when the effect comes back and changes its own cause, the loop closes and we have feedback. The author divides the elements of the loop into two kinds: if more at the input gives more at the output, he calls the transducer an affirmant, and if it gives less, a negant; hence positive and negative feedback, and within each of them divergent, critical and convergent courses. The calculations, on simple equations, are printed in small type, graphs show six characteristic courses, and alongside them stand numerous examples from technology, the economy and everyday life.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: none (the dictionary has no entry for ‘feedback’; the related entry control loop is given under Chapter 4, where this book discusses the feedback between the controlled system and the controlling system).
Chapter 3. Information (pp. 29–55)
The chapter answers the question of what information is, and translates into a practitioner's language the concept that Mazur had set out rigorously in the 1966 book. The key is a relation: if images stand in the same relation to one another as the originals they correspond to, they can replace those originals — and it is precisely this shared relation within a set that Mazur calls information, while he calls the relation linking an original with an image a code. It follows that a single message carries no information. The guiding example is a map, on which the ratios of distances can be seen without knowing the scale, whereas kilometres can be read only with it.
The author reaches this definition from a charge against information theory: it has learnt to measure the amount of information well, but has not defined the concept itself (here he sets side by side the positions of Shannon, Couffignal and Ashby). The quantitative part — the base-2 logarithm, the bit and the principle of asking questions so that each one splits the remaining possibilities in half — he presents as that theory's achievement. Mazur also applies the concept of the amount of information to regulations: a set of loose, frequently amended details burdens their addressee with a far greater amount of information than a few simple principles from which those details could be derived. The remaining parts of the chapter sort out the cases: they distinguish the transmission of information from its processing, and useful information from redundant information and from noise, against which redundancy is precisely the protection; finally they introduce information throughput, that is, the amount of information received per unit of time.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur (here: from Jakościowa teoria informacji (‘A Qualitative Theory of Information’), 1970): information, message, original, image, code, useful information, redundant information.
Chapter 4. Control (pp. 56–64)
The chapter's thesis, which matters to every manager: every instance of control is a loop, even when one part of it remains invisible. Mazur regards the technical distinction between ‘open-loop’ and ‘closed-loop’ control as misleading, because even in the former the circuit closes — through the human being who designed the device or operates it. Anyone who treats management as a one-way influence fails to see half of the circuit; the book returns to this matter in Chapter 8.
The background is the definition of control as purposeful influence on the course of a phenomenon, together with a set of basic terms in which the name of a device is decided by the function it performs, not by its structure. Using several examples from technology, everyday life and the economy, the author shows a recurring pattern: the controlled element of the loop acts as an affirmant and the controlling one as a negant, and together they produce negative feedback; the generalisation is a scheme with receptors, a correlator and effectors around the controlled system (Fig. 18). Every control process can therefore also be described as the transmission of information — an observation that links this chapter with the previous one.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: control, receptor, effector, correlator, controlled system, controlling system (here Mazur still writes układ sterowany and układ sterujący, the older Polish terms), control process, control loop (the feedback between the controlled system and the controlling one).
Chapter 5. Basics of a cybernetic theory of thinking (pp. 65–81)
The chapter is a condensed version of the theory from Cybernetyczna teoria układów samodzielnych (1966; the book is listed in the bibliography) and leads up to a translation of the words of psychology — sensation, memorising, representation, emotion, reflection, consciousness, thinking — into a description of control processes. In this account thinking turns out to be one of the varieties of correlation already described — the one that runs in the reflection circuit; Mazur points out that the choice of which process to give this name is a matter of linguistic usage, not of discovery. For management this yields a warning: an unpleasant experience can inhibit the reaction even to a strong stimulus, so a manager who relies solely on orders and disregards people's vital interests and their earlier experiences is using only one of the three available channels of influence.
The author justifies his method by arguing that the question ‘what is thinking’ was badly posed: the names came from everyday speech before anyone had established what phenomena they denote. He therefore starts from the phenomena and chooses the names last, and he regards his description as necessary rather than as one of many hypotheses — all information processing is a physical phenomenon, so it is governed by a difference of potentials and the energy balance. He assembles the model itself gradually, adding one organ after another until a system emerges that, thanks to the homeostat, controls itself in its own interest; the scheme in Fig. 20 gathers it into two paths, an information path and an energy path. Mazur counts all organisms among autonomous systems, but also families, enterprises and states, and he calls machines of this kind, not yet built at the time, autonoms. Inside the correlator, a decision is the crossing of a potential threshold, and a memory trace is an increase in the conductance of a path along which energy has already flowed.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: autonomous system (in this book the Polish term is układ samodzielny), effector, alimentators, accumulator, receptor, correlator, homeostat, functional equilibrium, registrator, estimator, correlational energy, correlational power, correlational potential, correlational medium, correlational conductance, registration potential, estimation potential, decision potential, decision, reflection potential, sensation, representation, emotion, reflection, consciousness, thinking. In this book Mazur treats consciousness, thinking and sensation more simply than in the dictionary definitions (from 1976).
Chapter 6. Praxiology (pp. 82–96)
The chapter sets cybernetics alongside praxiology, the science of efficient action: Mazur credits praxiology with a rich body of observation, but cybernetics with the advantage of method, because it derives cases from a model instead of collecting them; at the same time he does not regard the two sciences as rivals, since their fields overlap only in part. From this approach Mazur also draws a more general warning: the result of the whole is decided by the mutual influence of its parts, so the impeccable work of each executor taken separately may, in sum, do harm. The background is a survey of what praxiology has achieved, from proverbs to Tadeusz Kotarbiński's Traktat o dobrej robocie (‘A Treatise on Good Work’) and Taylor's scientific management, and the outcome is a critique of its tools: of nine criteria for assessing action, which largely overlap, Mazur leaves only effectiveness and economy standing as independent, and he objects that the five-stage scheme of action fails when the situation develops differently from what the plan assumed.
The core of the chapter is a dispute about the sources of errors. Working from the model of the autonomous system, Mazur arrives at twelve of them, because each of the six organs can fail for an internal or an external reason, and he states that the praxiologists' list covers only half. In Kotarbiński's five categories, from sluggishness to poor discernment, he sees a view through simple coupling: external difficulties are treated in it as given, so it is always the executor who turns out to be to blame.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: the organs described in Chapter 5 — effector, receptor, correlator, alimentators, accumulator, homeostat. Praxiological concepts have no entries in the dictionary.
Chapter 7. Decision-making (pp. 97–159)
The longest chapter — more than a third of the book — arranges decisions according to how much information they draw on, and from this scale derives the role of the manager: with the most carefully prepared decisions, the manager's task is no longer calculation but approval of the result and — what is increasingly difficult — posing the problem correctly. For Mazur, the decision proper in an institution is approval, that is, the endorsement of the result by someone entitled to give it; everything that precedes it, even a calculation carried out with the latest methods, is only preparation. This conclusion follows from reconciling two meanings of the word ‘decision’, the physical one from Chapter 5 and the everyday one: since a decision is only what sets action in motion, it cannot be withdrawn — one can only take another one.
The scale has four rungs: self-generated decisions, based solely on what the decision-maker carries within (experience, current stimuli and attitude); prognostic decisions, which draw on data from the environment in order to foresee how things will develop; trial decisions, which gain information through action; and optimisation decisions, prepared by specialists. The lowest rung, self-generated decisions, has the poorest stock of information, and yet — as Mazur observes — management rested on it for centuries; this is why the author devotes considerable space to it. He judges the method of trial and error severely: it requires each failure to suggest the direction of the next trial, and even then it has six faults, from slowness and cost to a tendency to stop at a solution that is the best only in its immediate neighbourhood.
The greater part of the chapter is taken up by a survey of the operational research methods of the day — linear and dynamic programming, inventory and queueing calculations, game theory, network diagrams with the critical path — with numerical examples. Above technique, however, Mazur places two reservations: the optimum is not the maximum, and an optimum is always calculated within the limits that someone has set for the task, and outside them it may be harmful. The chapter closes with a humorous anecdote that once more recalls the reservation about criteria: a flawless calculation based on a badly chosen measure leads to absurdity.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: decision, decision potential, estimation potential, registration potential, correlational conductance, reflection potential. The kinds of decision (self-generated, prognostic, trial, optimisation and their varieties) have no entries in the dictionary.
Chapter 8. Leadership (pp. 160–175)
The chapter, and with it the book, ends with the thesis that the manager in the old sense is losing their reason for existing. An order is to justify itself by an accuracy that can be demonstrated, not by the mere authority of the person who issues it, and leadership thereby shifts from compulsion to persuasion. What really relieves managers is only handing decisions over to the workers, and their proper job is preventing disturbances, not removing them. Multiplying levels does not help: each level processes orders and reports, introducing noise, and the defence against it — the redundancy of reminders and chasers — swallows up the time that the multi-level structure was supposed to save the manager.
Mazur reduces the organisational side of the chapter to a dispute between two charts. He regards the functional chart, made up of workstations and what circulates between them (information, materials, money), as the proper picture of an institution operating normally; the ordinary pyramid of subordination, called the ‘organisational chart’, makes sense in his view mainly where unforeseen events are part of everyday life, because in a well-organised institution the manager in any case deals almost exclusively with emergencies. The justification is cybernetic: the manager and the worker are autonomous systems of equal standing, each with a homeostat, so each side can give orders to the other, enforce them and demand reports (Fig. 41). Traditional leadership keeps only one of these six interactions — the order from the top down — and thereby reduces management to simple coupling.
Concepts discussed in the chapter — dictionary entries with definitions from other books by Mazur: autonomous system (the autonom as a machine with a homeostat), homeostat, useful information, redundant information.
How to read it
It is best to start with Chapters 2 and 8, which can be read without preparation and which show most clearly what cybernetics is for a manager: the first teaches one to recognise couplings in everyday situations, the second draws conclusions from this about organisation. Chapter 3 requires attention, because in it Mazur sets out accessibly his own concept of information — the same one he had already formulated in Cybernetyczna teoria układów samodzielnych — and Chapter 5 is the most difficult and the most condensed, and is worth reading alongside that book. In Chapter 7 the numerical examples printed in small type (linear programming, the inventory problem, dynamic programming) can be skimmed on a first reading: the part on self-generated and trial decisions grows out of Mazur's own theory, while the rest surveys methods from the operational research literature of the day.
Concepts from this book in the dictionary
The dictionary contains no definitions taken from this book (number of entries: 0). The concepts mentioned above have entries with definitions from other works by Mazur — see the domains cybernetics, qualitative information theory and theory of character.
Our own account, working version; definitions of concepts are in the dictionary, with their sources.