(c) Erik Hollnagel, 2017

SafetySynthesis is registered as a company in Denmark (Cvr-nr.: 35126910)

What is SafetySynthesis?

The meaning of the noun synthesis (from the ancient Greek σύνθεσις) is a combination of two or more entities that together results in something new; alternately, it may mean the activity of creating something out of something else that already exists. The meaning of safety synthesis is therefore the system quality that ensures that a system is able to succeed under varying conditions, so that the number of intended and acceptable outcomes is as high as possible. (This is obviously a paraphrase of the definition of resilience in Resilience Engineering.) This is however not a natural and stable condition, but an artificial and potentially unstable one.1 Safety is, with a paraphrase of Weick’s terminology, a ‘dynamic event’, hence something that must be created constantly and continuously. The basis is that which makes up everyday work and everyday existence. The synthesis of that, the bringing together of what individuals and organisations do on all levels and over time, is what creates safety – at least until a better term has been adopted.2 This synthesis has two different forms, a synthesis across organisational levels and a synthesis over time.

The synthesis across levels is relatively easy to explain. It means that one must understand the dependencies or couplings between everything that happens in an organisation or when carrying out an activity, no matter which levels of the organisation or types of work are involved. And it is of course necessary that the people who do the work understand that themselves, i.e., that the synthesis is part of what everyone does or at least that it is recognised by them.

The synthesis over time is more difficult to explain, but no less essential. In many kinds of activities – probably in all – synchronisation is important. This certainly goes for industries where safety (whether as Safety-I or as Safety-II) is a concern; it goes for services; for communication; for production – not least if it is lean - and so on. Synchronisation is achieved by organising the various productive processes to avoid delays (outputs arriving too early or too late), to ensure a better use of the resources (for instance, doing things in parallel so that the same preconditions do not have to be established twice), to coordinate transportation of matter and energy between processes and sites, and so on.

But synchronisation is not the same as synthesis. A synthesis is first achieved when we understand how things really fit together, when we understand the variability of everyday performance (the approximate adjustments), and how this variability may wax and wane, leading to outcomes that sometimes are detrimental and sometimes beneficial. A temporal synthesis cannot be achieved by exploring pairwise combinations of functions, even if it is done exhaustively. Neither can it be achieved by a bottom-up approach. Instead it requires a genuine top-down, synthetic perspective in the sense of being able to see (perhaps by serendipity rather than by combinatorics) when something new and useful has been achieved. Safety synthesis is the constant creating and maintenance of conditions that allow work to succeed on all criteria taken together.


  1. It may seem so in nature or for natural systems. But this is because the changes in the environment happen so slowly, that the system's adjustments are subliminal or imperceptible. For human-made systems, the socio-technical habitat, adjustments are needed so often and have so large a magnitude, that they are easier to see. It is therefore also easier to understand that the dynamic equiibrium that the system may be able (temporarily) to maintain represents a potentially unstable condition.
  2. The outcome of the synthesis is clearly not safety as such, neither in the Safety-I or the Safety-II interpretation. The outcome is rather system performance as a whole, corresponding to a holistic view, where safety is one aspect, but not the only one. Other aspects are quality, productivity, efficacy, etc. We have conventionally looked at system performance from different perspectives without acknowledging that they are different perceptions of the same behaviour, the 'blooming, buzzing confusion' (to borrow a phrase from William James) that we always break asunder but rarely reunite.