When Things Go Wrong: Complexity, Timing and Human Systems

When things go wrong, we usually look for the thing that went wrong.

A person made a mistake. An institution failed. A government chose badly. A technology produced an unintended consequence. Somewhere, we assume, there must be a defective part that can be identified and corrected. Define the problem clearly enough, locate its cause, intervene, and things should improve.

Sometimes they do. But this familiar logic becomes increasingly unreliable as the systems involved become more complex.

A society is not a machine in which every part waits patiently for us to identify it. It consists of processes operating at different rates, across different scales, with different histories and dependencies. Markets can move in milliseconds. Political decisions take months or years. Infrastructure lasts for decades. Institutions carry habits accumulated over generations. Ecological systems operate across seasons, centuries and longer. Human lives contain still other rhythms.

These differences in timing are not defects. They are part of how complex systems exist. Complexity depends upon processes remaining different while continuing to accommodate one another. The objective cannot be perfect synchronisation. A system in which everything moved together would lose much of the difference through which it adapts.

Problems arise when processes that depend upon one another can no longer accommodate those differences. A response arrives too late. Another arrives too quickly. One process accelerates while something upon which it depends cannot. A correction overshoots because circumstances have changed before it takes effect. Each component can behave reasonably according to its own circumstances while their interaction produces an increasingly unreasonable result.

The failure exists between them.

This creates a serious problem for the way modern societies attempt to govern themselves. We repeatedly assume that sufficiently simple definitions will allow us to intervene successfully in enormously complicated systems. We define unemployment, productivity, security, disadvantage, efficiency, risk, growth or educational achievement, attach measurements to those definitions, and organise interventions around improving the numbers.

The definition is necessary. It is not the world.

Once a definition becomes operational, institutions begin reorganising themselves around it. Resources move. Behaviour changes. People adapt. Incentives appear. Costs migrate. The intervention becomes part of the system it was intended to control.

The original problem can then reappear somewhere else, at another time, in another form, often carrying additional costs created by the attempted solution.

This is one reason apparently successful policies can coexist with persistent social failure. A department improves its processing time while transferring complexity to citizens. A business reduces costs while transferring them to workers, customers or the environment. A government suppresses an immediate problem while accumulating a larger one for whoever governs ten years later. Each intervention can be defended within the boundary used to measure it.

The boundary is doing much of the intellectual work.

This changes what efficiency means. Making one component faster does not necessarily make the system faster. If one process accelerates beyond the capacity of connected processes to respond, the result can be congestion, instability and repeated correction. What appears efficient locally can become expensive everywhere else.

The same applies to what appears to be waste. Spare capacity, redundancy, waiting, reflection and apparently unused resources provide room for adjustment. A system in which every component has been optimised to its measurable limit can become extraordinarily brittle because there is nowhere left for disturbance to go.

Complex problems are not impossible to influence. But intervention requires more humility than modern managerial culture usually permits. We need to ask not merely whether an action produces its intended immediate effect, but what changes elsewhere because the action occurred, which dependencies have been altered, which delays have been introduced, and where the consequences will become visible later.

This is where systems thinking becomes a human question. Displaced consequences do not disappear simply because they fall outside the boundary of a policy, institution or spreadsheet. Someone encounters them. Someone waits longer, works harder, absorbs the risk, loses the service, inherits the debt or lives with the environmental damage. A system can record an improvement while making somebody’s life worse.

A humanistic philosophy for a technological civilisation cannot simply oppose humanity to technology. Nor should it romanticise slowness, complexity or inefficiency. The point is not that everything is complicated and nothing can be done. That would merely replace simplistic confidence with sophisticated paralysis.

The point is that action must remain answerable to consequence.

We need institutions capable of discovering that their solutions are reproducing their problems. Measures must remain open to question rather than becoming substitutes for what they were created to describe. Above all, intervention must remain revisable when the world responds differently from the way our definitions predicted.

This is not weakness. It is what acting intelligently inside a complex system requires.

The alternative is familiar. Define the problem narrowly. Announce the solution. Measure the intended result. Declare success. Discover years later that the cost has migrated elsewhere. Define the new problem narrowly and begin again.

A civilisation can spend an enormous amount of energy solving the consequences of its previous solutions.

Differential timing helps explain how this happens. Causes and consequences do not necessarily arrive together. An intervention can produce immediate benefits and delayed damage. The institution responsible for the benefit may never encounter the institution responsible for the eventual cost. By the time the consequences arrive, the original decision has disappeared into administrative history. A new causal story is constructed around the new problem, and another intervention begins.

A humanistic philosophy informed by advanced systems thinking starts somewhere else. It accepts that knowledge will always be partial, that consequences propagate beyond the boundaries of our definitions, and that different processes must remain different while somehow continuing to coexist. It treats intervention as participation rather than command.

That does not prevent decisive action. It changes what intelligent action means.

The question is no longer simply: Did we fix the problem?

We must also ask: What did the solution change, where did the problem go, when will its consequences arrive, and who will eventually pay for them?

The task is not to discover a sufficiently simple description of the world and force reality to conform to it.

It is to learn how to act within a world that will always exceed our description of it.