(Note: I just got back from a three-week vacation. While I catch up with my reading and gather my thoughts for new posts, here’s an excerpt from a 2022 post, How to Break a Causal Chain)
Exploring causality, one baby step at a time…
“A root cause is an initiating cause of either a condition or a causal chain that leads to an outcome or effect of interest.” Wikipedia
“A causal chain is a sequence of conditions or events that results in a particular effect.” APA Dictionary of Psychology
What does it mean that “X caused Y”? Some possibilities: 1). X is a necessary and/or sufficient condition of Y; 2) If X had not occurred, Y would not have occurred; 3) There is a causal mechanism leading from the occurrence of X to the occurrence of Y. (Paraphrase from Understanding Society/University of Michigan-Dearborn.)
“Causal processes, causal interactions, and causal laws provide the mechanisms by which the world works; to understand why certain things happen, we need to see how they are produced by these mechanisms. Causal relations cannot be directly inferred from facts about association among variables.” Understanding Society/University of Michigan-Dearborn
There are three broad approaches to causal mechanisms: individual-level choice, social influence, and structural or institutional effects. An example of the latter would be how a country’s tax system influences tax compliance. (Paraphrase from Understanding Society/University of Michigan-Dearborn.)
“A deterministic interpretation of causation means that if A causes B, then A must always be followed by B… On the other hand, a probabilistic interpretation simply means that causes raise the probability of their effects.” Wikipedia
But X may increase the probability of Y without being a cause of Y, just as a falling barometer increases the probability of a storm coming without causing the storm. To establish causation you need an external intervention to change X. If Y still happens, then X does not cause Y.
“…the crucial piece missing from both mechanism and covariation accounts of causal reasoning is the notion of intervention. Specifically, researchers have suggested that knowing that X directly causes Y means knowing that, all else being equal, intervening to change X can change Y.” Schulz, Gopnik, & Glymour/Preschool children learn about causal structure from conditional interventions
“As long as not all relevant factors involved in an investigated causal structure are controlled for in the set-up of a pertinent study, corresponding data tends to be confounded by hidden variables and is, hence, likely not to unambiguously reflect underlying causal structures.” Baumgartner/The Causal Chain Problem
Effects alter the causal chain through feedback and feedforward mechanisms. Outputs generate data that become inputs for further causal processes and interactions. Causal links may be neutralized, reinforced or weakened by ongoing inputs and outputs.
Structures and systems undermine and reinforce each other. Nothing lasts forever but some things can be propped up for a while.
X may feed Y more than other sources of influence, but that doesn’t mean that getting rid of X is the only or best way to get rid of Y. Y may be overdetermined, the product of multiple causal pathways, some independent of the others.
Sometimes small interventions have large ripple effects, as causal links across connected pathways are recalibrated.
A proximate cause may be the effect of multiple causes, each in turn the effect of multiple causes, each cause contributing its bit.
The cause may not be a thing but a quantity of a thing - a dose. The dose effect depends, in part, on the characteristics of its recipient. Moderators can bump up a dose to the point of generating an effect. Fixing, or at least managing, a problem may be a matter of changing the dose of its cause. Or it may be a matter of building resilience in recipients so they’re protected from ill-effects. Often the best interventions go after both the dose and resilience. (Think climate change mitigation and adaptation).
The root cause of a bad thing may be the same root cause of a good thing. Trees put out more than one root.
A root cause is not fixed: its effects are not fixed. What gets the ball rolling may not keep it going. What keeps it going may change. What keeps the ball rolling may lose potency with repetition, or be worn down by contrary forces. In other words, causal pathways are subject to decay.