*In consequence of this it is necessary to abandon the tradition of basing risk assessments on the notion of failures, at least when the assessment includes human and organisational functions. Instead we must acknowledge that there can be situations where failures as such cannot be found, but where the outcome nevertheless is unacceptable. It is also necessary to refrain from relying on the principle of linear combinations of effects. This follows from a simple apagogical (reductio ad absurdum) argument: if adverse outcomes were due to a linear combination of effects, then there would have to be a failure or a malfunction for the chain of events to start. But since the latter is no longer a necessary condition (given that the ETTO principle is correct, of course), then there is no need to assume that effects combine linearly. (This argument applies regardless of whether the causal chain depicts 1:1, 1:n, n:1, or n:n relations.)*

The ETTO Principle, Erik Hollnagel