Complexity theory

According to complexity theory, performance is the result of complex interactions and relationships. It is seen as an emergent property. The performance can however not overlap with was stakeholders like to see in accident investigations. They prefer to blame individuals when their system fails. It is this narrow-minded and easy way of thinking that is being criticized here.

Newtonian science

The Newtonian way of thinking is very appealing, because it is simple, coherent, seems to be complete and consistent with common sense. The most famous principle is that of analysis of reductionism. The entire system can be explained by combining all the separate elements. To find out the cause of an error, they rely on the defenses-in-depth metaphor, which breaks down the system in a linear way to find the broken layers or parts. The goal is to analyze the basic components and find out where it is failing.

The following are important aspects of the Newtonian science:

• Error causation. According to Newtonian science, everything has a definitive and identifiable cause and a definitive effect. Finding what caused a failure is the goal of the accident investigation. It assumes that physical effects have physical causes.

• Newton also said that the future can be predicted with complete certainty, if its state at any point was known in all details. So if somebody can be shown to have known, or should have known, the initial positions and movement of the components, then this person could have predicted the failure in the future.

• Newtonian science can also be used to investigate a trajectory backwards. Evolution can be reversed to reconstruct an earlier state.

• All the laws of the world can be discovered and all knowledge can be gathered. The more facts are collected, the more realistic is the representation of what is being investigated. It is therefore also possible to just have one version of the truth, which will be the ultimate, complete, perfect truth. This is called ‘the true story’. ‘The truest story’ is the one with the smallest gap between external events and internal representations.

Complexity theory

Complex behavior is the consequence of interactions between the components of a system. The focus in this approach lies not on the individual components, but on their interaction. The complexity is therefore not embedded within a certain component, but is generated by the interactions in the system to react to changing conditions in the environment. Because the knowledge of each component is limited and lays within that component, there cannot be one single component with enough capacity to represent the complexity of the entire system.

Complex systems are formed by separated local relationships. Not one of the components knows the behavior of the system as a whole. The components respond locally to the information that they are given. Complexity is created by huge webs of relationships and interactions, vague boundaries and interdependencies.

Errors, failures and accidents happen as a consequence of relationships between components and not because of individual components. It is the interactive complexity of the system which gives rise to conditions that help cause an accident. Think about a slow loss of control or automation leading to carelessness.

Assymetry and foreseeability

Assymetry, or non-linearity, means that a tiny change in the starting conditions can lead to massive differences later on. This plays a big role in the blaming debate. Decisions that are being made at a certain time can be completely rational given the circumstances in which they were made. They can be made with all the goals, knowledge and attention of the decision maker. However, the interactive complexity of the system makes it impossible to predict the outcome of the system. The relationship of that single decision to the outcome is complex and non-linear and impossible to predict.

Irreversibility and incompleteness of knowledge

Unlike the Newtonian approach, the complexity theory says it is not possible to reconstruct the past. Because the system after the failure is not the same as the system before the failure. Complex systems are constantly changing because of evolving relationships. They constantly need to change with their changing environment. The ‘causes’ of the failure are embedded in many relationships, unwritten routines, implicit expectations and professional judgements.

Also, unlike the Newtonian science, the complexity theory says that we can never obtain all the knowledge and acquire one ‘truth’. Because the observer creates the truth based upon the inputs he is receiving. Different observers will interpret these inputs differently, or might even notice completely different inputs. It is impossible to determine whose view is right.

How to analyze failures?

The writers introduce a post-Newtonian analysis of failure in complex systems. An investigator should try to gather as much information as he can, even though he could never gather all the information possible. This also means that the investigator cannot uncover or discover the one truth. Finally, the investigator should be sure that any conclusion can be revised at any time, if it may become clear that it has flaws.