Blind Managers, Systems Complexity and Weak Signals

International audienceIn our rapid and global society, the complexity of dimensions entering in accidental sequences (new and emerging Technology, increasing interdependencies between systems, external constraints, and societal concerns) imposes to consider industrial accident through holistic and interdisciplinary approaches. Several scientific studies have validated this approach with models describing these systems as socio-technical and complex (Reason, 1990, Rasmussen, 2000, Leveson, 2004). In this context, classical management tools adopting systems decomposition (risk analysis, reliability, technical and organizational barriers...) strongly reduce the probability of major accident. However, they meet several limitations, and major accidents still occurring the most cases, study of accidents shows us that systems degradation is often associated with managerial failures showing up through several forms (lack of control, communication problems, deficiencies in the management of change...). Consequently, technical barriers are gradually losing their efficiency and systems safety performances are decreasing. In this process, tension is extreme in the resilience of systems, causing an intermediate phase of resistance, announcing the rupture. In these conditions, some particular operating phases (temporary period, maintenance shutdown, restarting) can no longer be contained until the major accident. This is during the period of crucial resistance that the system sends the most representative symptoms. This kind of precursor signals, known as weak signals, is broadly described in the literature. In this context, how to deal with the complexity of systems, and gather weak signals shown by organizations, to anticipate major accident? Our research question is the following. The literature demonstrates the existence of weak signals, and their relevance as accident precursors. Why these signals have not been taken into account? Our work should help to analyze the rupture between resilience and resistance and find the weak signals in organizations. Our study is based on the methodology of case study.. As result, our study enabled us to identify several filters that have hidden weak signals in organizations. Development of automatization, management focused on standards, involvement of stakeholders increasingly numerous, and economic issues freeze the intuitive perception of the system by the manager.. It is now in management systems that we must seek the weak signs of an announced catastrophe