I don't know which is more impressive - your profound knowledge in this subject or the fact that you can say "proceduralisation" without effort. Well done!
Really helpful look at the key differences between the thinking of James Reason and the ideas of complexity - I've previously followed Sidney Dekker's work and so I was familiar with the different schools of thought but the classification of these two schools of thought have helped coalesce the key differences.
The fundamental issue is maybe people assume things go right or wrong due to different causes and do not necessarily understand the relationship between “human error” and “performance variability”. If resilience is defined as the ability to sustain required operations in both expected and unexpected conditions (Hollnagel, 2012; Hollnagel et al., 2010; Rankin et al., 2014), or about “having the generic ability to cope with unforeseen challenges, and having adaptable reserves and flexibility to accommodate those challenges” (Nemeth, 2008, pp. 5-6), then performance variability is a vital factor to ensure resilience of a system (Dekker, 2004; Hollnagel, 2009a). Human adaptations actually manifest themselves in error mechanisms: errors that are associated with adaptation is in nature a behavioural process of safety boundary seeking, thus adaptations or human errors cannot be eliminated per se (Foord & Gulland, 2006; Rasmussen et al., 1994).
I would guess that most systems tend towards increasing complexity so that there is a transition from early system errors being mainly human error and late system errors being complexity driven. If you look beyond the root cause analysis to the fix, I often see complexity and human choice pushed upstream orr downstream to other groups resulting in no real fix but a reduction in local error rate. Errors in error analysis, or errors in selecting the best framework for a given process pop up everywhere. Who is to blame when a process is well designed (precise) but built on faulty beliefs (accuracy)? I have seen a hundred companies design pay structures that everyone agrees are effective but consistently produce exactly the wrong behaviors. The humans are following the process well. The process is mature, stable and approved. The results stink. 99% of the time they blame the humans for producing the wrong results.
