The AIHS Bok on Causation commences with the following:
‘If safety management is effective, then there should be an absence of accidents. Conversely, if accidents are occurring then effective safety management must be absent’ (p.1).
The assumptions in these statements are of course not declared in the BoK but no wonder the industry is the industry of counting.
Could it possibly be that the presence of accidents is simply evidence that people are fallible and the world is random? Perhaps this is why 5 lines later the BoK itself declares that accident causation is a ‘mystery’? Well, if it’s a ‘mystery’ why commence with the double affirmations at the start of the chapter?
Of course, the focus of the BoK on Causation is on objects (hazards). This is consistent with the common Safety worldview.
The Chapter discusses three approaches to causation models:
· Simple linear models
· Complex linear models
· Complex non-linear models
Again, there are other models regarding causation but these are not presented in the BoK.
One of the models discussed is the Energy Causation Model, still espoused by many. The model appears as below:
What is amusing about this model is the notion that objects/hazards contain an ‘energy’ stimulated by a ‘space transfer mechanism’ and then a recipient. What a fascinating way of understanding an accident or life. Only an engineer could concoct such a model. The mechanistic assumptions of the model are accepted as a given and this weird thing called a ‘space transfer mechanism’??? Some interesting assumptions here:
· In what way does a hazard/object think?
· In what way does a hazard/object make decisions?
· In what way does a hazard/object enact?
This model, similar to the Bowtie model leaves the mysterious nature of the fallible human out of the picture (stuck in the middle). Rather than discussing humans the model discusses ‘supervisory systems’. Rather than discussing the object-subject dialectic, the model puts forward the idea of ‘opportunity for ‘active control’ of the occurrence’. How interesting that the Damaging Time Model excludes any issues to do with social psychology, culture or a range of non-engineering influences. On the other hand if the worldview included an understanding of wickedity, fallibility and ecological dynamics then the idea of ‘damaging energies’ doesn’t make sense.
It is an positive feature of the BoK on Causation that it makes one aware of the models as metaphors but at no time discusses the mechanism of metaphor itself. Such a discussion is essential for understanding how worldviews construct models and how such models limit perception. (The Rule of Metaphor or Metaphors We Live By are a useful read in this regard). The ‘generic epidemiological model’ discussed in the BoK is an example of this.
The idea that elements of causation could be likened to ‘pathogens’ in a body is a strange metaphor indeed. A useful construction for Reason (1987) who wanted to see accidents as caused by ‘latent conditions’, ‘unsafe acts’ and ‘latent errors’. Of course, there are types of pathogens and they need not disable the body nor result in illness. By jumping to the epidemiological model this approach to causation creates many more problems that it solves.
Whatever accident causation model one holds is underpinned by a philosophical-anthropological-ontological position. Each school of thought in safety (https://safetyrisk.net/a-great-comparison-of-risk-and-safety-schools-of-thought/) constructs it own model to suit its anthropological assumptions.
We need to know much more about the assumptions of each model of causation than getting caught up in the technique (Ellul) of each model itself. One thing is for sure, behaviourist, positivist and rationalist assumptions deliver mechanistic models of accident causation. When humans are understood as ‘factors’ in a system then causation becomes systemic and humans sub-set of systems.
Finally the BoK discusses Hollnagel’s Functional Resonance Accident Model (FRAM). Again, this is a model that views humans through systems and humans as variables/functions within systems. The functions are categorised as inputs, outputs, preconditions, resources, time and control. Read the language in all of these models, it is your indicator to the philosophical-anthropological-ontological position.
Whilst the idea of resonance introduced in this model is helpful, it’s not the way a Social Psychological approach understands ‘resonance’. Similarly, the idea that resilience can be ‘engineered’ seem a weird metaphor for understanding human ecologies. If one understands risk as a wicked problem, one gets an entirely different understanding of systems. Rather than looking at complex adaptive systems one needs to explore collective coherences and transdisciplinarity to better understand causation. Dekker is partly right as stated in the BoK:
‘… the search for the “broken and part or person” that underpins linear models where risk is considered in terms of energy-to-be-contained, barriers and layers of defence, or cause and effect are misleading because they assume rational decision-making (p.19)’.
Interestingly the popular iCAM model of incident causation seems to be the ‘go-to’ for no other reason that it has been accepted mysteriously as some kind of standard. iCAM is perhaps one of the worst models of causation on the market.
Safety doesn’t seem to care much about the assumptions of the checklist, just as long as they have one.
Could it be that one’s own worldview combined with the worldview of the iCAM simply presents the world as you want to see it? Maybe Safety might be better off with no model, no checklist? Ah!!! Shudder!
Moreso, if risk and fallibility are a ‘wicked problem’ maybe disciplines other than Safety might offer something of value to the industry in the issue of causation? Then again, does Safety have the capability to listen and perceive the value in transdisciplinarity itself? Or must it always seek approaches to causation from the comfort of its own camp? The BoK certainly endorses such a view.