Nonlinear causality is a form of causation where cause and effect can flow in a bidirectional fashion between two or more elements or systems. The essential characteristic of nonlinear causality is the idea of feedback that an effect can create a cause, but equally, this cause can then feedback to create an effect in the first system. Nonlinear causality can be contrasted with linear causality where the direction of affect flows in a unique direction. Nonlinear causation leads to a number of important outcomes that are not possible when considering more simple circumstances of linear causality. Nonlinear causality can lead to self-reinforcing or self-amplifying processes through feedback, thus allowing for disproportionality between initial cause and final effect. The second outcome to nonlinear causality is the bidirectional flow of causation between the macro and micro levels within a system, thus enabling downward causation. Thirdly it can allow for reverse causality in time; that set future goals can feedback to affect current events. Likewise, nonlinear causation implies the action of many variables in creating a cause or vise verse, which can lead to equifinality, the idea that some end effect may be created or reached through a number of different pathways. Finally, in contrast to linear causality that creates a deterministic vision of the world, nonlinear causality may lead to the indetermination of outcomes.
Feedback processes are a central characteristic of nonlinear causation. Events do not happen in isolation but instead feedback to effect their source. Thus nonlinear causality can have the characteristic of being self-perpetuating and self-referential. For example, when looking at some relationship dynamic between a parent and child we might note that the child is uninterested in her parents because her parents showed no interest in her, but equally, her parents showed no interest in her because she appeared uninterested. This is a self-reinforcing loop enabled by nonlinear causation. Likewise, the formation of hurricanes, financial crisis, animal stampedes, or the development of a culture would be other examples of processes driven by self-reinforcing cause and effect.
Whereas linear causality is defined by a degree of proportionality between a given cause and its effect, nonlinear causality driven by feedback can enable a disproportionality between cause and effect; what is called the butterfly effect. Because of feedback loops, some small event can get compounded through each causal feedback iteration enabling a rapid change in proportionality between the input and output. The result of this can be widely divergent outcomes to some situation depending on only small changes in input values, what is called sensitivity to initial conditions.
Whereas linear causation is predicated on being able to isolate a single or small amount of variables causing a given effect, nonlinear causation does not try to reduce the number of reasons for a given effect or vice versa, the number of effects derived from a cause. For example, if we took the linear reductionist paradigm and asked the question why does an airplane fly? This approach would try to reduce the cause to a limited number of direct physical interactions. In which case, the answer would be traced back to the dynamics of the airflow around the wing as described by a few variables within the Navier-Stokes equations of fluid dynamics.
Inversely a nonlinear causal description may involve many different explanations to this question. We might say the flight is caused by the pilot directing the plane to its destination. Or the flight is caused by the fact that it was chartered to fly at a particular time. Or that it is flying because the company can make a profit from putting on that flight etc. Without any of these factors, the flight would not be happening.1 When this non-reductive, more holistic interpretation to causality is taken, it is possible to see that some effects are the product of an almost infinite number of interacting factors, and it stops making sense to talk about a direct effect. Instead, the language switches to that of emergence. Asking how many different factors interact in a specific fashion to create a given outcome, with emergent phenomena creating higher level patterns that then feedback to exert an effect on lower level more elementary parts.
Whereas linear causality implies that causality flows from the bottom-up but not in the reverse direction, nonlinear causality and the idea of emergence enable the interpretation of events as both upwardly caused and downwardly caused, with causation flowing bidirectionally from the micro to the macro and back again. To illustrate this, we might think about a polar bear and ask why is it white? The reason it is white one might say is because of its genotype, which is a bottom-up answer. But if we ask why its genes are such we would discover that they are a product of evolution, which has selected for the color best suited to that environment. The polar environment is white so the genes that produce a white bear have been selected for, if it were in the Canadian forests it would be a brown color. Thus we see downward caution acting, with the “cause” coming from the environment to affect the state of the individual.
Likewise, the specific atomic nuclear interaction in the interior of a star at any time is determined by where that reaction is taking place within the overall star. Thus the overall structure of the system is affecting the specific phenomenon in a downward fashion. But again the atomic interactions affect the whole, creating a nonlinear causal relationship between the system’s micro and macro level with feedback. These patterns have downward causal efficacy in that they can affect which causal powers of their constituents are activated, and this has significant implications for our conception of determinism.2 Both top-down and bottom-up causation can occur at the same time3
Whereas linear causality and upward causality leads to the vision of a deterministic world. Nonlinear causality leads to greater capacity for indeterminism. Indeterminism is the concept that events are not caused, or not wholly caused deterministically by prior incidents.4 With linear causality, a past cause creates a current effect in a direct fashion. With nonlinear causality, a cause may create an effect but because of the top-down and bottom-up bidirectional flow to causality, the context and conditions of this low-level interaction are conditioned by higher level phenomena. Meaning the overall outcome is a product of this more complex interaction between low-level cause and effect and the upper-level organization that sets the context, thus enabling a much greater opportunity for indeterminism.
Philosophy Robert Van Gulick describes this phenomenon as such “A given physical constituent may have many causal powers, but only some subsets of them will be active in a given situation. The larger context (i.e. the pattern) of which it is a part may affect which of its causal powers get activated. . . . Thus the whole is not any simple function of its parts, since the whole at least partially determines what contributions are made by its parts”5 Thus with nonlinear causality, the cause of events are not directly determined by preceding events. But more emerge out of the bi-directional exchange between the conditions set by the overall system and the local interactions.
With nonlinear causality, a single cause can have many effects, such as a nerve cell sending out many impulses, or inversely many causes can have a single effect such as a hurricane being the product of temperature, pressure, humidity, etc. This leads to the idea of equifinality, which is the principle that in open systems a given end state can be reached by many potential means.6 The term and concept are derived from Hans Driesch, the developmental biologist and later applied by Ludwig von Bertalanffy, the founder of general systems theory.
Some systems have more than one pathway or process for achieving a given goal. This increases the likelihood that the system will achieve its goal under varying environments and circumstance. If one subsystem is damaged, or if environmental circumstances change significantly, the presence of multiple mechanisms or pathways thereby increases the likelihood that the varying conditions can be adapted to or overcome. For example, the human immune system has both a pre-existing component and induced antibody component to respond to foreign “invaders.” Some organisms have multiple pathways and utilize them in different environmental circumstances. These Redundant systems accomplish the same end but do so with more than one similar or equivalent channel.7
Our conception of time is closely connected to our understanding of causality. A linear conception of causality leads to a unidirectionality to cause and effect in time, and this is a fundamental component in structuring our understanding of past, present, and future. When considering only matter and energy within a purely physical system this unidirectionality to causality with respect to time may hold. However when we introduce information into the model it now becomes possible for future events to feedback to affect current events, thus enabling reverse causation.
Information encoding and processing is an essential feature of biological systems that differentiates them from purely physical systems and allows their departure from deterministic physical causality. Many entities have control systems that enable them to process information; examples include animals, people, social institutions and various kinds of technology. With a control system the structure and initial conditions may not matter, what matters is the goal, future goals can determine current actions. Here we can note causality reversing as it goes from some projected future event, back to affect the present state. The human body is acting out this process virtually every moving second, in that we typically formulate some goal before initiating any action, such as desiring the future state of eating a meal before acting out the process of cooking. The physical structure and initial conditions determine the outcomes as governed by equations, past effects cause current events going forward but with a control system the information defines a system’s desired behavior or response and thus cause and effect are contingent on higher level information whose cause is derived from some future projection.8