Marian Mazur began work on the theory that later grew into the theory of autonomous systems in 1942. In the preface to his 1966 book he notes that he did not then know of the new field of science that was emerging, since Wiener's work was still little known (Cybernetyczna teoria układów samodzielnych, p. 7). When he returned to this work in 1956, the writings of Wiener and Shannon were already widely known; according to Lechowski, this did not take away the novelty of the concept of the autonomous system (Lechowski 1987).
Mazur does not dispute the definition of cybernetics. He holds that Wiener gave it himself in the title of his book, and notes its agreement with Ashby's view that cybernetics is the science of control in general (Cybernetyczna teoria układów samodzielnych, p. 11). He defended this position in his polemic ‘O cybernetyce i cybernetykach’ (‘On Cybernetics and Cyberneticists’, 1965).
In our reading, the main disagreement concerns information. Mazur acknowledges that Shannon introduced the concept of the amount of information in a strictly defined way, but stresses that his theory is about the amount of information, not information itself, and that Shannon himself stated that he leaves aside the content of messages (Jakościowa teoria informacji, pp. 15–22; Cybernetyka i charakter, pp. 118–119). He considers Wiener's characterisations of information insufficient (Jakościowa teoria informacji, pp. 19–20), and cites Wiener's statement that information is neither matter nor energy as a trace of dualism (Cybernetyka i charakter, p. 128). He builds his own theory starting from an analysis of the control process; in it, the number of items of identifying information turns out to equal the amount of information given by Shannon's formula (Jakościowa teoria informacji, pp. 203–206). From Ashby, Mazur took the name ‘homeostat’, giving it, however — in our reading — a different scope (Cybernetyczna teoria układów samodzielnych, p. 58).
How to read the comparison
We chose the concepts on which the largest number of definitions in the dictionary rest, and the main concepts of Mazur's cybernetics. Each is compared with Mazur's definition as given on its entry page and assigned one of three classes:
- equivalent — the same concept, often under a different name;
- related concept — a concept with a different scope; the sentence in the ‘Difference’ column says what the difference is;
- no counterpart — a concept specific to Mazur; the sentence says what makes it specific.
The comparison covers only the works listed at the end of this page. We quote nothing from the Western works: we give the name of the concept and its place — the section (or the introduction) in Shannon, the section in Ashby, the chapter and the page of the Polish edition in Wiener (for Cybernetyka i społeczeństwo, the chapter only), and the chapter and the page in Bertalanffy. The class ‘no counterpart’ means only that we found no counterpart in these works, not that none exists anywhere. Concepts for which the sources examined did not allow a class to be assigned with confidence are not included.
Mazur refers to Western authors explicitly in few places — the most important of them are given in the introduction. The choice of counterpart, the class and the description of the difference are therefore usually our own comparison. Each such place is marked with the phrase ‘in our reading’, including in the section ‘What the comparison shows’. Sentences without this phrase report what the sources say at the places indicated.
The comparison covers 39 concepts: equivalent — 5, related concept — 27, no counterpart — 7.
Cybernetics
| Mazur's concept | Polish term | Western counterpart | Class | Difference | Source |
|---|---|---|---|---|---|
| control | sterowanie | control, regulation (Ashby); control (Wiener) | related concept | For Mazur, any behaviour of one system that leads to specified changes in another is control. In our reading, the closest notions in Ashby are two: control as the ability to force an outcome whatever the opponent does, and regulation as keeping essential variables within permitted limits despite disturbances. | Ashby 1956, §§ 10/6, 11/3; Wiener 1948, Introduction (Polish ed. p. 35) |
| controlling system | system sterujący | regulator (Ashby) | related concept | In our reading, the difference is this: Ashby's regulator is defined by what it does — it keeps the variety coming from disturbances from reaching the essential variables — whereas for Mazur the controlling system is a role within a pair of systems; using a furnace with a thermostat, he shows that which of them controls is settled only by someone outside who wants one of them to behave in a given way. | Ashby 1956, §§ 10/5–10/6; Mazur, Jakościowa teoria informacji, pp. 29–30 |
| controlled system | system sterowany | essential variables (Ashby) | related concept | In our reading, its counterpart in Ashby is the essential variables: what is to be controlled is described in his work by variables that must stay within a set of permitted states. Mazur defines the controlled system by the fact that the action of another system brings about the desired (specified) changes in it — with no set of permitted states and no limits within which anything must stay. | Ashby 1956, §§ 10/4–10/6; Mazur, Jakościowa teoria informacji, p. 28; Cybernetyka i charakter, p. 112 |
| control path | tor sterowniczy | channel (Shannon) | related concept | Shannon's channel is merely the medium carrying the signal from transmitter to receiver, in one direction. Mazur's control path is a system through which one system acts on the other — the controlling system on the controlled one, or the controlled system on the controlling one; a control loop has two such paths. In our reading, the difference is that a control path is always a link in a feedback loop, whereas the channel in Shannon's diagram is a link in communication running in one direction. | Shannon 1948, Introduction (diagram of a communication system); Mazur, Jakościowa teoria informacji, pp. 28–29; Cybernetyka i charakter, pp. 112, 119 |
| control loop | obwód sterowniczy | feedback loop (Wiener); feedback (Ashby) | equivalent | The same concept — a closed circuit of actions. Mazur lists its components (two systems and the paths joining them); in our reading, Wiener stresses above all stabilising negative feedback, and Ashby the circularity of action between parts of a system as such. | Wiener 1948, ch. IV (Polish ed. pp. 131–132); Ashby 1956, § 4/11; Ashby 1952, § 3/11 |
| autonomous system | system autonomiczny | ultrastable system (Ashby); open system (Bertalanffy) | related concept | Ashby's ultrastable system rejects one field after another until it reaches one that does not lead to critical states; energy is not discussed in it. Bertalanffy's open system exchanges matter with its surroundings and maintains a steady state, but has no control organs. In our reading, Mazur's autonomous system differs from both in having informational as well as energetic organs, and it is distinguished by its ability to counteract the loss of its ability to control itself. | Ashby 1952, §§ 8/4–8/5; Bertalanffy 1952, ch. 4, p. 125 |
| homeostat | homeostat | homeostat (Ashby) | related concept | Mazur took the name from Ashby and says so himself. In Ashby the homeostat is a whole machine of four coupled units, built to demonstrate ultrastability. In Mazur it is one of the six organs of the autonomous system — the one that counteracts flows of information and energy which weaken the system's capacity to act on its environment. In our reading, it is therefore a related concept, not an equivalent: the same name denotes the whole for one author and a part for the other. | Ashby 1952, § 8/8; Ashby 1956, § 5/14; Mazur, Cybernetyczna teoria układów samodzielnych, p. 58 (footnote) |
| functional equilibrium | równowaga funkcjonalna | keeping essential variables within physiological limits (Ashby); homeostasis (Wiener, after Cannon) | related concept | Ashby and Wiener speak of keeping variables within the range in which the organism lives. Mazur singles out one state within that range: the one farthest from both limits beyond which the system loses its ability to control. In our reading, these are related concepts; Mazur himself calls Ashby's homeostat a model of the restoration of functional equilibrium. | Ashby 1952, §§ 3/14, 5/3; Wiener 1948, ch. IV (Polish ed. pp. 152–153); Mazur, Cybernetyczna teoria układów samodzielnych, p. 58 (footnote) |
| receptors | receptory | receptors (Wiener, Ashby) | equivalent | The same function and the same name: organs taking in information from the environment. In Mazur they are one of the six organs of every autonomous system. | Wiener 1948, ch. I (Polish ed. p. 71); Ashby 1952, § 3/11 |
| effectors | efektory | effectors (Wiener) | equivalent | The same function and the same name: the organs through which the system acts on its environment. | Wiener 1948, ch. I (Polish ed. p. 71) |
| correlator | korelator | central control system between receptors and effectors (Wiener) | related concept | In Wiener the central system processes sensations, stores information (memory) and itself produces the response of the effectors. Mazur's correlator also processes and stores information; in our reading, what sets it apart is that in Mazur decisions and reactions are also influenced by the homeostat coupled with the correlator (although a reflex decision can be taken before the homeostat acts). | Wiener 1948, ch. I (Polish ed. p. 71); Mazur, Cybernetyka i charakter, pp. 194, 219 |
| accumulator | akumulator | — | no counterpart | The organ that converts and stores energy. Ashby takes energy as given and leaves it out of his considerations (in his view, cybernetics can be defined as the study of systems open to energy but closed to information), and Wiener mentions the flow of energy between an automaton and the world; in our reading, neither of them assigns a separate organ to energy. | Ashby 1956, § 1/5; Wiener 1948, ch. I (Polish ed. p. 71) |
| alimentators | alimentatory | — | no counterpart | Organs taking energy from the environment — the system's energetic input beside its informational one (the receptors). In our reading, a control system in the works examined has no separate organ of supply. | examined: Wiener 1948, 1950; Shannon 1948; Ashby 1952, 1956 (see § 1/5) |
| sociological power | moc socjologiczna | — | no counterpart | Power converted in the environment of an autonomous system, as distinct from the power converted within it. In our reading, the works examined do not divide power into that converted within the system and that converted in its environment; Bertalanffy speaks only of the energy an organism needs for its activities and for maintaining its steady state. | examined: Wiener 1948, 1950; Shannon 1948; Ashby 1952, 1956; Bertalanffy 1952 (ch. 4) |
Qualitative information theory
| Mazur's concept | Polish term | Western counterpart | Class | Difference | Source |
|---|---|---|---|---|---|
| source of action | źródło oddziaływania | information source (Shannon) | related concept | Shannon's source produces messages with given statistics. For Mazur the source is any system acting on another system in a control loop — in our reading, with no assumptions about probability, since Mazur's definition contains none. | Shannon 1948, Introduction (part 1 of the diagram), § 2 |
| receiver of action | odbiornik oddziaływania | receiver, destination (Shannon) | related concept | Shannon separates the receiver, which reconstructs the message from the signal, from the destination for which the message is intended. Mazur has one concept: the system acted on by another system in a control loop; in our reading, it is related to both of these roles. | Shannon 1948, Introduction (parts 4–5 of the diagram) |
| message | komunikat | message and signal (Shannon); message (Ashby) | related concept | Shannon's message is an element chosen from a set at the source and reproduced at the destination, distinct from the signal in the channel. Mazur's message is a physical state at any point of the path, so in our reading the concept covers both message and signal. | Shannon 1948, Introduction; Ashby 1956, § 7/5 |
| transverse set of messages | zbiór poprzeczny komunikatów | set of possible messages (Shannon); set of possibilities and its variety (Ashby) | related concept | Shannon and Ashby speak of a set of possible messages from which one is selected; Ashby measures it by its variety. Mazur's transverse set is the messages at one, arbitrary point of the path — in our reading, with no probabilities of selection, for which Mazur's definition makes no provision. | Shannon 1948, Introduction; Ashby 1956, §§ 7/5–7/6 |
| original | oryginał | message produced by the source (Shannon) | related concept | In Shannon the transmitted message is to be reproduced exactly or approximately. Mazur's original is a message at the output of the source of action; in our reading, the difference is that what is to be preserved in the images is not the original itself but the information between originals. | Shannon 1948, Introduction |
| image | obraz | message reconstructed for the destination (Shannon) | related concept | In Shannon the receiver reconstructs the message from the signal. In our reading, Mazur's image is a related but broader concept: it may physically bear no resemblance to the original, and what matters is whether the information between images is the same as between originals. | Shannon 1948, Introduction (part 4 of the diagram) |
| primary message | komunikat pierwotny | operand (Ashby) | equivalent | In our reading, the same concept: that which is acted on by a change. Ashby speaks of states of any system, Mazur of messages in a control path; we regard this difference of scope as secondary to their common function. | Ashby 1956, § 2/2 |
| secondary message | komunikat wtórny | transform (Ashby) | equivalent | In our reading, the same concept: what a change turns its object into. As with the primary message, we regard the difference of scope (states of any system in Ashby, messages in a path in Mazur) as secondary. | Ashby 1956, § 2/2 |
| transformation | transformacja | transformation (Ashby) | related concept | Ashby defines a transformation as a set of transitions over the whole set of operands — what changes into what — without going into the cause of the change. Mazur's transformation concerns a single association: it is the process to which one message must be subjected to obtain the other, and at the same time the relation between them; in our reading, it is closer to a single transition in Ashby. | Ashby 1956, §§ 2/2–2/3; Mazur, Jakościowa teoria informacji, p. 42 |
| operation | operacja | operator (Ashby) | related concept | Ashby's operator is the factor that causes a change. Mazur's operation is one of the elementary processes of which a transformation consists. In our reading, the comparison is a loose one: it pairs the agent of a change with a component of a process. | Ashby 1956, § 2/2 |
| code | kod | coding as a transformation (Ashby); encoding by the transmitter (Shannon) | related concept | Ashby too defines coding by a transformation, but one of a whole set of messages into their coded forms; in Shannon encoding is done by the transmitter. Mazur's code is the change of one message into another along the path, framed — in our reading — for a single association rather than for a whole set of messages. That coding takes place everywhere, including in the senses and the nerves, is something Ashby also notes. | Ashby 1956, §§ 8/2, 8/4; Shannon 1948, Introduction (part 2 of the diagram), § 8 |
| coding chain | łańcuch kodowy | transducers connected in tandem (Shannon) | related concept | Shannon connects transducers in series and shows that the whole is again a transducer. In our reading, the difference is that Mazur's coding chain is the sequence of messages along the path linked by codes — considered as a whole from original to image, not as a combination of devices. | Shannon 1948, § 8 |
| information | informacja | amount of information as a measure of choice (Shannon); amount of information as negative entropy, a measure of organisation (Wiener) | related concept | Shannon and Wiener define the amount of information, not information itself; Wiener also gives a descriptive characterisation through ‘content’. For Mazur information is the transformation of one message into another at one point of the path — in our reading, a relation between messages, not a measure. | Shannon 1948, Introduction, § 6; Wiener 1948, Introduction (Polish ed. pp. 34–35), ch. III (p. 96); Wiener 1950, ch. I |
| identifying information | informacja identyfikująca | amount of information, bit (Shannon) | related concept | Shannon gives a measure. Mazur singles out the items of information needed to identify a message in a chain and shows that their number corresponds to the amount of information given by Shannon's formula. | Shannon 1948, Introduction, § 6; Mazur, Jakościowa teoria informacji, pp. 203–206 |
| describing information | informacja opisująca | variety, counted as the number of distinct elements (Ashby) | related concept | The information needed to determine any message in a chain, not only a distinguished one. The number of such items equals the number of distinct messages, that is — in our reading — Ashby's variety in the first of its two senses; Ashby, however, speaks of a property of a set, Mazur of the items of information themselves. According to Mazur and Lechowski, Shannon's theory does not cover them. | Ashby 1956, §§ 7/6–7/7; Mazur, Jakościowa teoria informacji, pp. 204–206; Lechowski 1987 |
| redundant information | informacja redundancyjna | redundancy (Shannon, Ashby) | related concept | In our reading, the difference is this: redundancy in Shannon and Ashby is a numerical property of a source or a set — what part of the possibilities is superfluous — whereas Mazur's redundant information is a single item of information that follows from other information in the same chain. | Shannon 1948, § 7; Ashby 1956, § 9/16 |
| parasitic information | informacja pasożytnicza | noise (Shannon, Ashby) | related concept | Shannon's noise is a chance variable that perturbs the signal in the channel; Ashby stresses that noise is variety unwanted by a given recipient. In our reading, the difference is that Mazur defines parasitic information by its origin outside the given control process, with no reference to randomness. | Shannon 1948, § 11; Ashby 1956, § 9/19 |
| informing | informowanie | communication as reproducing a message at another point (Shannon) | related concept | In Shannon the task of communication is to reproduce a message at another point. In Mazur informing carries information from the chain of originals to the chain of images: the messages themselves may change, and Mazur distinguishes kinds of informing by whether the information has been preserved and by the structure of the coding chains — shared messages, incomplete chains, paramessages. In our reading, it is a concept related to communication in Shannon, not its equivalent. | Shannon 1948, Introduction; Mazur, Jakościowa teoria informacji, pp. 82–84 |
| pseudoinforming | pseudoinformowanie | non-invertible distortion (Shannon) | related concept | Shannon calls it distortion when what arrives is fully determined by what was sent, and notes that it can be undone when no two different transmitted signals give the same received one. In Mazur, messages shared by several coding chains define pseudoinforming, whose varieties Mazur then distinguishes; in our reading, Shannon's distortion corresponds to only one of them — a common image of different originals. | Shannon 1948, § 11 |
| disinforming | dezinformowanie | noise and equivocation (Shannon) | related concept | In our reading, the closest notions to disinforming in Shannon are noise and equivocation. Shannon measures the average uncertainty about the transmitted signal once the signal has been received (equivocation). Mazur does not measure but distinguishes gaps in coding chains: an image with no corresponding original, and an original with no image. | Shannon 1948, §§ 11–12 |
| parainformation | parainformacja | — | no counterpart | Information in an association one of whose messages belongs to an informational chain but to no coding chain. The distinction follows from the division into transverse and longitudinal sets, which — in our reading — the works examined do not make as a named and consistently applied distinction. | examined: Shannon 1948; Ashby 1956, ch. 7–9 |
| metainformation | metainformacja | — | no counterpart | Information about the relation between the information of two different control paths, used in a metacontrol path. In our reading, the works examined have no concept of information arising from comparing two paths. | examined: Shannon 1948; Wiener 1948; Ashby 1956 |
Theory of character
| Mazur's concept | Polish term | Western counterpart | Class | Difference | Source |
|---|---|---|---|---|---|
| motivation | motywacja | affective tone (Wiener) | related concept | Wiener sketches a feedback mechanism in which a scale from pain to pleasure reinforces or inhibits the nervous processes under way, and adds that the conditioned reflex may, though need not, work in this way. For Mazur motivation is the dependence of the effector potential on the action of the homeostat, that is, of decision on emotion and reflection: approval supports a decision, aversion blocks it. In our reading, the difference lies in Mazur assigning this role to one organ of the autonomous system. | Wiener 1948, ch. V (Polish ed. pp. 168–169); Mazur, Cybernetyka i charakter, pp. 202–203, 208 |
| character | charakter | — | no counterpart | The set of rigid control properties of an autonomous system; Mazur develops the concept for human beings. The theory of character, built on parameters of information processing (level) and of energy processing (dynamism), has, in our reading, no counterpart in the works examined. | examined: Wiener 1948, 1950; Ashby 1952, 1956; Mazur, Cybernetyka i charakter, pp. 272, 289 |
| character dynamism | dynamizm charakteru | — | no counterpart | A parameter of character computed from how fast the system expands compared with how fast it ages (the logarithm of the ratio of the two coefficients). Derived from the energetics of the autonomous system; in our reading, specific to Mazur. | examined: Wiener 1948, 1950; Ashby 1952, 1956 |
What the comparison shows
In our reading, Mazur shares a basic vocabulary with Wiener: a closed circuit of actions, receptors, effectors and a centre that processes information. The difference appears, in our reading, with energy. Wiener notes that an automaton is linked to the world by a flow of energy as well, but describes it from the side of information. Mazur, in our reading, does the opposite: he gives energy its own organs and quantities — the accumulator, the alimentators and the kinds of power — and from them later derives the energetic character parameters, above all dynamism.
Formally, in our reading, the closest comparison is with Ashby. The operand and the transform correspond — in our reading — to the primary and the secondary message, and coding is for both authors a kind of transformation. In our reading, however, two things set them apart: Ashby deliberately deals with systems open to energy but closed to information, and his homeostat is a whole machine, built to demonstrate ultrastability, whereas Mazur's is one of the organs of the autonomous system.
In our reading, Shannon and Mazur answer different questions: Shannon measures how much choice a message involves and deliberately leaves its meaning aside, whereas Mazur asks what information is and answers with a relation between messages. Shannon's measure does not disappear from this theory, since it corresponds to the number of items of identifying information. The named and consistently applied distinction between transverse and longitudinal sets — and with it para- and metainformation — has no counterpart in the works examined; in our reading, this is the most specific part of the comparison. The number of items of describing information, on the other hand, has, in our reading, a close counterpart in Ashby's variety counted as the number of distinct elements.
Works compared
- Wiener 1948 — N. Wiener, Cybernetics: or Control and Communication in the Animal and the Machine, 1948 (2nd ed. 1961); pages from the Polish edition: Cybernetyka, czyli sterowanie i komunikacja w zwierzęciu i maszynie, PWN, Warsaw 1971.
- Wiener 1950 — N. Wiener, The Human Use of Human Beings, 1950 (2nd revised ed. 1954); chapters from the Polish edition, translated from the revised edition: Cybernetyka i społeczeństwo, Książka i Wiedza, Warsaw 1960.
- Shannon 1948 — C. E. Shannon, ‘A Mathematical Theory of Communication’, Bell System Technical Journal 27 (1948); references to sections.
- Ashby 1952 — W. R. Ashby, Design for a Brain, 1952 (reprint with corrections: Wiley, New York 1954); references to sections, e.g. § 8/8.
- Ashby 1956 — W. R. Ashby, An Introduction to Cybernetics, Wiley, New York 1956; references to sections.
- Bertalanffy 1952 — L. von Bertalanffy, Problems of Life, Watts, London 1952 (German edition 1949).
Works by Mazur: Cybernetyczna teoria układów samodzielnych (‘A Cybernetic Theory of Autonomous Systems’, 1966), Cybernetyka a zarządzanie (‘Cybernetics and Management’, 1969), Jakościowa teoria informacji (‘A Qualitative Theory of Information’, 1970), Cybernetyka i charakter (‘Cybernetics and Character’, 1976), the article ‘O cybernetyce i cybernetykach’ (‘On Cybernetics and Cyberneticists’, 1965). On Mazur: J. Lechowski, ‘Życiorys naukowy profesora Mariana Mazura (1909–1983)’ (‘The Scientific Biography of Professor Marian Mazur (1909–1983)’), Postępy Cybernetyki, vol. 10, no. 2, 1987, pp. 7–20.