One recurring myth in “pop” complexity is the notion of circular causality. This is a poor expression: because it overlaps with the term feedback; because of the confusion it creates around the word causality; and because of the unprecise use of the adjective circular, since what is actually meant here is recursion/iteration and not circularity.
Unfortunately, the expression was used in the title of one the Macy’s Conferences that contributed to launch the cybernetics (a.k.a. systems science) approach. Heinz von Foerster, who was learning English, was nominated secretary of the eighth conference and assigned as curator of the minutes, which were eventually published as “Cybernetics: Circular Causal and Feedback Mechanisms” in Biological and Social Systems: Transactions of the Eighth Conference, Edited by Heinz von Foerster. New York: Josiah Macy, Jr. Foundation, 1952.
The “circular causality” contained in this unfortunate title hasn’t had much success in the scientific field: to date, it only sticks in a few minor and rather naive psychology and neurology publications. But it has enjoyed a great fortune in pop complexity literature (e.g., “complexity theories” for management), where it is employed as a synonym of feedback, with an added cheap “philosophical” spin.
Feedback is what occurs when the temperature in the room falls below a certain value and a thermostat orders your air conditioner to stop cooling: information concerning the output of the system or its environment (room temperature) is fed back to the the system itself (the conditioner) in order to modify its behavior.
In other occasions, it is not just information being fed back, but a piece of the output signal itself: in electronics, for example, a portion of the output signal is extracted and then subtracted from the input signal before it reaches the system/processor. This is done for output stabilization purposes.
From Encyclopedia Britannica, [feedback]:
In biology, a response within a system (molecule, cell, organism, or population) that influences the continued activity or productivity of that system. In essence, it is the control of a biological reaction by the end products of that reaction. Similar usage prevails in mathematics, particularly in several areas of communications theory. In every instance, part of the output is fed back as new input to modify and improve the subsequent output of a system. See also cybernetics.
Feedback was studied systematically for the first time in the 1920’s, in both living organisms and electric circuits. In the latter field, the pioneers were physicist Harry Nyquist and electronic engineer Harold Black, of Bell Labs. In biology, pyschology, neurophysiology and economics feedback was addressed, among others, by A. Bogdanov, P. Anokhin, S. Odobleja, I. P. Pavlov and A. R. Wagner.
In 1943 A. Rosenblueth and N. Wiener argued that feedback in technological systems was analogous to that in living systems: in both cases feedback is a strategy employed for stabilization, self-regulation and ultimately survival. Cybernetics, the science they were founding, was to study feedback processes in nature in order to abstract ideas and strategies to be employed in artificial systems as well.
Feedback brings recursion and self-regulation in the affected processes.
Think of the air conditioning case: the system interacts with its enviroment and adjusts to it, because information about the environment is fed as input into the system which is meant to regulate the environment itself. In the electronic amplifier case, the input signal to be processed depends somehow on the output signal (the result of the processing) as well, since a portion of the output is fed in as input.
This recursion, known as feedback loop in engineering circles, is what gets the minds of cheap philosophers excited, and makes them talk about circular causality.
Linear and circular
People who use the expression circular causality to refer to the feedback loop are misusing the “linear” and “circular” adjectives and are confused about system evolution over time.
By saying circular causality, what the naive writer is trying to imply is that between event A and event B there isn’t a simple, straight cause-effect relationship, but rather a more convoluted one, of the kind: A affects B but is also affected by it.
This situation is more appropriately labeled by terms such as recursion or iteration, to preserve the progressive aspects of the process. If the process were “circular”, it would endlessly loop with no further progress, as it is implied by the notion of a closed line, the circle.
As an example, the curve produced by an ideal pendulum in phase space is indeed a circumference. But a system that changes dramatically over time, as opposed to just bounce back and forth between the same states over and over like an idealized pendulum, produces open curves in phase space: the real pendulum, whose energy is declining due to attrition, produces a spiral in phase space, not a circle.
In general, a [non trivial] process with feedback proceeds from a state to a different one and, as the accurate Britannica definition clarifies, feedback occurs after feedforward: “[…] part of the output is fed back as new input to modify and improve the subsequent output of a system.” (The boldface is mine). Graphically, this situation is described not by a circle but rather by a cycloid, which is a loop with progress.
Crippled feedback and causality
In psychoneurology, a variant of the circular causality expression has made its way in minor portions of the literature, in more or less the following form:
«Feedback is one form of nonlinear causation. A second form, termed circular causality […], describes bidirectional causation between different levels of a system. A coherent, higher-order form or function causes a particular pattern of coupling among lower-order elements, while this pattern simultaneously causes the higher-order form. The top-down flow of causation may be considered an emergent constraint (by the system as a whole) on the actions of the parts.»
This definition, taking us in the realm of feedback between a whole system and some of its parts, is severely flawed and reveals great confusion around the notion of system, the term causality and the term linear. I will not waste my precious time 🙂 in a confutation; just see for example Bakker.
Also notice that this definition preserves the logical confusion I was referring to in the preceeding section: a confusion between system states at different times. It is true that output may affect input (or that whole may affect components, if you so wish), but it does that in a subsequent iteration of the process. I.e., state A affects state B and may be affected by it, but the latter happens at a subsequent iteration. B–>A means that state A(t+1) is influenced by state B(t).
(NB: In particle physics, situations have been observed where causality seems violated in the sense that effect E happens in time before its cause B, and information is transmitted at a speed greater than light’s. However, this is not the kind of situation to which pop complexity authors are referring to when they speak of circular causality.)
Circularity and Francisco Varela
Perhaps it is not surprising that the only scientific fields where the expression “circular causality” is still used (although not much) be areas of life sciences, such as psychology and neurophysiology.
These are fields where “hard” scientific methods and terminology have started to penetrate only in the past decades (thanks to cybernetics, by the way) and where their application is tougher, given the complexity of the subject matter. It is therefore simply natural that, in these fields, some mathematical terminology may be imperfect or that systems science concepts may be slightly abused.
One notable exponent of the use of “circularity” in human sciences was Francisco Varela (F. Varela, “The Creative Circle: Sketches on the Natural History of Circularity”, in The Invented Reality, edited by P. Watzlavick, Norton Publishing, New York 1984), the creator of autopoiesis, of which I will write in a separate piece.
Varela was a great man and a valued thinker, although with a pronounced tendency to fall in the traps of his own beautiful and elegant metaphors. Witness his application of the liar paradox to the dynamical context of molecular biology. The circle vs. the cycloid…