Expecting Normal Accidents in Interpersonal Relationships
Making a second blog post to go deeper into this portion of the previous blog: While we tend to fixate on a specific trigger event, catastrophic outcomes are rarely the result of individual negligence– they are inherent to the architecture of the system itself.
Here’s a good link to the terminology I will be using in reference to “Normal Accidents” by Charles Perrow.
The 2 main axes in Perrow’s book on “high-risk” systems are: Interactive Complexity and Tight Coupling. I find it intuitive to correlate Perrow’s failure points in nuclear systems to failure points in interpersonal relationships.
“Interactive Complexity” is the presence of invisible, non-linear feedback loops. In a romantic relationship, this is the “baggage” we don’t realize is connected to the present. EX: when a comment about the dishes triggers a sub-system of childhood neglect, which in turn activates a defense mechanism that the other partner interprets as hostility. Because the interactions are hidden and complex, the “accident” seems to come from nowhere. It is, however, “normal”.
“Tight Coupling” in this context would best be put as the absence of a buffer between lives. EX: In a family, it is when a parent’s professional exhaustion at 5:00 PM becomes the child’s anxiety at 6:00 PM. When the emotional stability of a household is tethered entirely to its most volatile member, the system is operating with zero margin for error. There is no insulation to prevent one person’s friction from igniting the entire structure.
When a system is both complex and tightly coupled, it is highly unreliable and difficult to to establish fault tolerance.
It is rational to stop viewing the “crash” as a moral failing and start seeing it as a systemic certainty. The question is not how to avoid the accident, but how to design a system that where the impacted sub-systems can degrade gracefully so that the entire system doesn’t explode.
I don’t have answers on how to design a flawless system, but the markers of a reliable and resilient system can always be monitored. Decay in sub-systems can be observed and be updated to degrade gracefully. We just have to find the right metrics.